From 905add56a803651d692fc700abffcd799d614fb3 Mon Sep 17 00:00:00 2001 From: "dependabot[bot]" <49699333+dependabot[bot]@users.noreply.github.com> Date: Mon, 13 Jan 2025 11:53:09 -0800 Subject: [PATCH 01/20] Bump cryptography from 42.0.4 to 43.0.1 in /utils/git (#6894) MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Bumps [cryptography](https://github.com/pyca/cryptography) from 42.0.4 to 43.0.1.
Changelog

Sourced from cryptography's changelog.

43.0.1 - 2024-09-03


* Updated Windows, macOS, and Linux wheels to be compiled with OpenSSL
3.3.2.

.. _v43-0-0:


43.0.0 - 2024-07-20

... (truncated)

Commits

[![Dependabot compatibility score](https://dependabot-badges.githubapp.com/badges/compatibility_score?dependency-name=cryptography&package-manager=pip&previous-version=42.0.4&new-version=43.0.1)](https://docs.github.com/en/github/managing-security-vulnerabilities/about-dependabot-security-updates#about-compatibility-scores) Dependabot will resolve any conflicts with this PR as long as you don't alter it yourself. You can also trigger a rebase manually by commenting `@dependabot rebase`. [//]: # (dependabot-automerge-start) [//]: # (dependabot-automerge-end) ---
Dependabot commands and options
You can trigger Dependabot actions by commenting on this PR: - `@dependabot rebase` will rebase this PR - `@dependabot recreate` will recreate this PR, overwriting any edits that have been made to it - `@dependabot merge` will merge this PR after your CI passes on it - `@dependabot squash and merge` will squash and merge this PR after your CI passes on it - `@dependabot cancel merge` will cancel a previously requested merge and block automerging - `@dependabot reopen` will reopen this PR if it is closed - `@dependabot close` will close this PR and stop Dependabot recreating it. You can achieve the same result by closing it manually - `@dependabot show ignore conditions` will show all of the ignore conditions of the specified dependency - `@dependabot ignore this major version` will close this PR and stop Dependabot creating any more for this major version (unless you reopen the PR or upgrade to it yourself) - `@dependabot ignore this minor version` will close this PR and stop Dependabot creating any more for this minor version (unless you reopen the PR or upgrade to it yourself) - `@dependabot ignore this dependency` will close this PR and stop Dependabot creating any more for this dependency (unless you reopen the PR or upgrade to it yourself) You can disable automated security fix PRs for this repo from the [Security Alerts page](https://github.com/microsoft/DirectXShaderCompiler/network/alerts).
Signed-off-by: dependabot[bot] Co-authored-by: dependabot[bot] <49699333+dependabot[bot]@users.noreply.github.com> --- utils/git/requirements_formatting.txt | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/utils/git/requirements_formatting.txt b/utils/git/requirements_formatting.txt index 546662aa33..06db8176c9 100644 --- a/utils/git/requirements_formatting.txt +++ b/utils/git/requirements_formatting.txt @@ -18,7 +18,7 @@ charset-normalizer==3.2.0 # via requests click==8.1.7 # via black -cryptography==42.0.4 +cryptography==43.0.1 # via pyjwt darker==1.7.2 # via -r llvm/utils/git/requirements_formatting.txt.in From 5bf6f77ea8944bf05074e21937674820eb56aa1d Mon Sep 17 00:00:00 2001 From: Joshua Batista Date: Mon, 13 Jan 2025 17:49:14 -0800 Subject: [PATCH 02/20] [DXIL] Avoid infinite loop / stack overflow on invalid instruction (#7048) The simplify instruction pass has a function called `findScalarElement`. This function will possibly recurse infinitely into the first operand of the instruction that it is passed. If this instruction, for example, is `%2 = insertelement <2 x i32> %2, i32 %sub, i32 0` Then the pass will recurse into the first operand of the insertelement instruction, which is the same insertelement instruction, and so on. This PR prevents infinite recursion and stack overflow by bailing out of the function early if there is a detection of the same value in the first operand. Fixes https://github.com/microsoft/DirectXShaderCompiler/issues/7034 --------- Co-authored-by: github-actions[bot] --- lib/Analysis/VectorUtils2.cpp | 4 ++ tools/clang/test/CodeGenHLSL/minimal2.ll | 49 ++++++++++++++++++++++++ 2 files changed, 53 insertions(+) create mode 100644 tools/clang/test/CodeGenHLSL/minimal2.ll diff --git a/lib/Analysis/VectorUtils2.cpp b/lib/Analysis/VectorUtils2.cpp index b8cac7a9a6..97bae15abe 100644 --- a/lib/Analysis/VectorUtils2.cpp +++ b/lib/Analysis/VectorUtils2.cpp @@ -42,6 +42,10 @@ llvm::Value *llvm::findScalarElement(llvm::Value *V, unsigned EltNo) { if (EltNo == IIElt) return III->getOperand(1); + // Guard against infinite loop on malformed, unreachable IR. + if (III == III->getOperand(0)) + return nullptr; + // Otherwise, the insertelement doesn't modify the value, recurse on its // vector input. return findScalarElement(III->getOperand(0), EltNo); diff --git a/tools/clang/test/CodeGenHLSL/minimal2.ll b/tools/clang/test/CodeGenHLSL/minimal2.ll new file mode 100644 index 0000000000..846ce51f06 --- /dev/null +++ b/tools/clang/test/CodeGenHLSL/minimal2.ll @@ -0,0 +1,49 @@ +; RUN: opt -simplify-inst -S < %s | FileCheck %s + +%"class.Texture3D" = type { i32, %"class.Texture3D::mips_type" } +%"class.Texture3D::mips_type" = type { i32 } +%dx.types.Handle = type { i8* } +%dx.types.ResourceProperties = type { i32, i32 } + +@"\01?t1@@3V?$Texture3D@I@@A" = external global %"class.Texture3D", align 4 + +; Function Attrs: nounwind +define void @main(<2 x i32> %dtid, i32 %laneID) #0 { +"\01?f1@@s1@@U1@@Z.exit": + br label %if.end.13 + +while.cond.6.preheader: ; No predecessors! + ; CHECK: %0 = extractelement <2 x i32> %1, i32 1 + %0 = extractelement <2 x i32> %2, i32 1 + %cmp7.14 = icmp ne i32 %0, 0 + br i1 %cmp7.14, label %while.body, label %if.end.13 + +while.body: ; preds = %while.cond.6.preheader + %1 = extractelement <2 x i32> %2, i32 0 + ; CHECK: %sub = sub i32 %sub, 1 + ; CHECK: %1 = insertelement <2 x i32> %1, i32 %sub, i32 0 + %sub = sub i32 %1, 1 + %2 = insertelement <2 x i32> %2, i32 %sub, i32 0 + br label %if.end.13 + +if.end.13: ; preds = %while.body, %while.cond.6.preheader, %"\01?f1@@s1@@U1@@Z.exit" + ret void +} + +; Function Attrs: nounwind +declare void @llvm.lifetime.start(i64, i8* nocapture) #0 + +; Function Attrs: nounwind +declare void @llvm.lifetime.end(i64, i8* nocapture) #0 + +; Function Attrs: nounwind readnone +declare i1 @"dx.hl.op.rn.i1 (i32, i1)"(i32, i1) #1 + +; Function Attrs: nounwind readonly +declare i32 @"dx.hl.op.ro.i32 (i32, %dx.types.Handle, <4 x i32>)"(i32, %dx.types.Handle, <4 x i32>) #2 + +; Function Attrs: nounwind readnone +declare %dx.types.Handle @"dx.hl.createhandle..%dx.types.Handle (i32, %\22class.Texture3D\22)"(i32, %"class.Texture3D") #1 + +; Function Attrs: nounwind readnone +declare %dx.types.Handle @"dx.hl.annotatehandle..%dx.types.Handle (i32, %dx.types.Handle, %dx.types.ResourceProperties, %\22class.Texture3D\22)"(i32, %dx.types.Handle, %dx.types.ResourceProperties, %"class.Texture3D") #1 From e52b6bc1d9bc524959299fe9d247490e1e778877 Mon Sep 17 00:00:00 2001 From: Joshua Batista Date: Tue, 14 Jan 2025 19:50:45 -0800 Subject: [PATCH 03/20] Update version to 1.8.2502 (#7065) Update version to 1.8.2502 --- docs/ReleaseNotes.md | 4 +--- utils/version/latest-release.json | 10 +++++----- utils/version/version.inc | 6 +++--- 3 files changed, 9 insertions(+), 11 deletions(-) diff --git a/docs/ReleaseNotes.md b/docs/ReleaseNotes.md index cd7c7b874a..b47f17b600 100644 --- a/docs/ReleaseNotes.md +++ b/docs/ReleaseNotes.md @@ -17,9 +17,7 @@ The included licenses apply to the following files: ## Changelog -### Upcoming Release - -Place release notes for the upcoming release below this line and remove this line upon naming this release. +### Version 1.8.2502 - The incomplete WaveMatrix implementation has been removed. - DXIL Validator Hash is open sourced. diff --git a/utils/version/latest-release.json b/utils/version/latest-release.json index 04a3881343..3138ccd2b1 100644 --- a/utils/version/latest-release.json +++ b/utils/version/latest-release.json @@ -1,8 +1,8 @@ { - "version" : { - "major" : "1", - "minor" : "8", - "rev" : "2407" + "version": { + "major": "1", + "minor": "8", + "rev": "2502" }, - "sha" : "737a12a663f1697d3755a522d8fbf30481ecd2f6" + "sha": "070d0d5a2beacef9eeb51037a9b04665716fd6f3" } diff --git a/utils/version/version.inc b/utils/version/version.inc index 4e5d9f7313..2577daa529 100644 --- a/utils/version/version.inc +++ b/utils/version/version.inc @@ -18,7 +18,7 @@ #ifdef RC_VERSION_FIELD_3 #undef RC_VERSION_FIELD_3 #endif -#define RC_VERSION_FIELD_3 2407 +#define RC_VERSION_FIELD_3 2502 #ifdef RC_VERSION_FIELD_4 #undef RC_VERSION_FIELD_4 @@ -28,7 +28,7 @@ #ifdef RC_FILE_VERSION #undef RC_FILE_VERSION #endif -#define RC_FILE_VERSION "1.8.2407.0" +#define RC_FILE_VERSION "1.8.2502.0" #ifdef RC_FILE_DESCRIPTION #undef RC_FILE_DESCRIPTION @@ -49,7 +49,7 @@ #ifdef RC_PRODUCT_VERSION #undef RC_PRODUCT_VERSION #endif -#define RC_PRODUCT_VERSION "1.8.2407.0" +#define RC_PRODUCT_VERSION "1.8.2502.0" #ifdef HLSL_TOOL_NAME #undef HLSL_TOOL_NAME From e4636f06383f3ccb8aefc49aa9b83a0f63676036 Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Nathan=20Gau=C3=ABr?= Date: Fri, 17 Jan 2025 11:58:02 +0100 Subject: [PATCH 04/20] [SPIR-V] Avoid emitting Int64 when loading Float64 (#7073) MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit When loading a Float64 from a raw buffer, we used an Int64, which required an additional capability, even if the code wasn't using any Int64. In practice, it seems most devices supporting Float64 do also support Int64, but this it doesn't have to. By changing the codegen a bit, we can avoid the Int64 value. Tested the word-order using a vulkan compute shader, and checking the returned value on the API side. ```hlsl double tmp = buffer.Load(0); if (tmp == 12.0) buffer.Store(0, 13.0); ``` Fixes #7038 --------- Signed-off-by: Nathan Gauër --- tools/clang/lib/SPIRV/RawBufferMethods.cpp | 82 ++++------- ...address-buffer.load.double.capability.hlsl | 39 +++++ ...-address-buffer.templated-load.matrix.hlsl | 79 +++++------ ...-address-buffer.templated-load.scalar.hlsl | 133 ++++++++---------- ...-address-buffer.templated-load.vector.hlsl | 43 +++--- ...address-buffer.templated-store.struct.hlsl | 88 ++++++------ 6 files changed, 229 insertions(+), 235 deletions(-) create mode 100644 tools/clang/test/CodeGenSPIRV/method.byte-address-buffer.load.double.capability.hlsl diff --git a/tools/clang/lib/SPIRV/RawBufferMethods.cpp b/tools/clang/lib/SPIRV/RawBufferMethods.cpp index 537c012ace..87409e7ccc 100644 --- a/tools/clang/lib/SPIRV/RawBufferMethods.cpp +++ b/tools/clang/lib/SPIRV/RawBufferMethods.cpp @@ -117,48 +117,32 @@ SpirvInstruction *RawBufferHandler::load64Bits(SpirvInstruction *buffer, SpirvInstruction *ptr = nullptr; auto *constUint0 = spvBuilder.getConstantInt(astContext.UnsignedIntTy, llvm::APInt(32, 0)); - auto *constUint32 = - spvBuilder.getConstantInt(astContext.UnsignedIntTy, llvm::APInt(32, 32)); + // Load the first word and increment index. auto *index = address.getWordIndex(loc, range); - - // Need to perform two 32-bit uint loads and construct a 64-bit value. - - // Load the first 32-bit uint (word0). ptr = spvBuilder.createAccessChain(astContext.UnsignedIntTy, buffer, {constUint0, index}, loc, range); SpirvInstruction *word0 = spvBuilder.createLoad(astContext.UnsignedIntTy, ptr, loc, range); - // Increment the base index address.incrementWordIndex(loc, range); + + // Load the second word and increment index. index = address.getWordIndex(loc, range); - // Load the second 32-bit uint (word1). ptr = spvBuilder.createAccessChain(astContext.UnsignedIntTy, buffer, {constUint0, index}, loc, range); SpirvInstruction *word1 = spvBuilder.createLoad(astContext.UnsignedIntTy, ptr, loc, range); - - // Convert both word0 and word1 to 64-bit uints. - word0 = spvBuilder.createUnaryOp( - spv::Op::OpUConvert, astContext.UnsignedLongLongTy, word0, loc, range); - word1 = spvBuilder.createUnaryOp( - spv::Op::OpUConvert, astContext.UnsignedLongLongTy, word1, loc, range); - - // Shift word1 to the left by 32 bits. - word1 = spvBuilder.createBinaryOp(spv::Op::OpShiftLeftLogical, - astContext.UnsignedLongLongTy, word1, - constUint32, loc, range); - - // BitwiseOr word0 and word1. - result = spvBuilder.createBinaryOp(spv::Op::OpBitwiseOr, - astContext.UnsignedLongLongTy, word0, - word1, loc, range); - result = bitCastToNumericalOrBool(result, astContext.UnsignedLongLongTy, - target64BitType, loc, range); - result->setRValue(); - address.incrementWordIndex(loc, range); + // Combine the 2 words into a composite, and bitcast into the destination + // type. + const auto uintVec2Type = + astContext.getExtVectorType(astContext.UnsignedIntTy, 2); + auto *operand = spvBuilder.createCompositeConstruct( + uintVec2Type, {word0, word1}, loc, range); + result = spvBuilder.createUnaryOp(spv::Op::OpBitcast, target64BitType, + operand, loc, range); + result->setRValue(); return result; } @@ -441,39 +425,31 @@ void RawBufferHandler::store64Bits(SpirvInstruction *value, const auto loc = buffer->getSourceLocation(); auto *constUint0 = spvBuilder.getConstantInt(astContext.UnsignedIntTy, llvm::APInt(32, 0)); - auto *constUint32 = - spvBuilder.getConstantInt(astContext.UnsignedIntTy, llvm::APInt(32, 32)); - auto *index = address.getWordIndex(loc, range); + // Bitcast the source into a 32-bit words composite. + const auto uintVec2Type = + astContext.getExtVectorType(astContext.UnsignedIntTy, 2); + auto *tmp = spvBuilder.createUnaryOp(spv::Op::OpBitcast, uintVec2Type, value, + loc, range); - // The underlying element type of the ByteAddressBuffer is uint. So we - // need to store two 32-bit values. + // Extract the low and high word (careful! word order). + auto *A = spvBuilder.createCompositeExtract(astContext.UnsignedIntTy, tmp, + {0}, loc, range); + auto *B = spvBuilder.createCompositeExtract(astContext.UnsignedIntTy, tmp, + {1}, loc, range); + + // Store the first word, and increment counter. + auto *index = address.getWordIndex(loc, range); auto *ptr = spvBuilder.createAccessChain(astContext.UnsignedIntTy, buffer, {constUint0, index}, loc, range); - // First convert the 64-bit value to uint64_t. Then extract two 32-bit words - // from it. - value = bitCastToNumericalOrBool(value, valueType, - astContext.UnsignedLongLongTy, loc, range); - - // Use OpUConvert to perform truncation (produces the least significant bits). - SpirvInstruction *lsb = spvBuilder.createUnaryOp( - spv::Op::OpUConvert, astContext.UnsignedIntTy, value, loc, range); - - // Shift uint64_t to the right by 32 bits and truncate to get the most - // significant bits. - SpirvInstruction *msb = spvBuilder.createUnaryOp( - spv::Op::OpUConvert, astContext.UnsignedIntTy, - spvBuilder.createBinaryOp(spv::Op::OpShiftRightLogical, - astContext.UnsignedLongLongTy, value, - constUint32, loc, range), - loc, range); - - spvBuilder.createStore(ptr, lsb, loc, range); + spvBuilder.createStore(ptr, A, loc, range); address.incrementWordIndex(loc, range); + + // Store the second word, and increment counter. index = address.getWordIndex(loc, range); ptr = spvBuilder.createAccessChain(astContext.UnsignedIntTy, buffer, {constUint0, index}, loc, range); - spvBuilder.createStore(ptr, msb, loc, range); + spvBuilder.createStore(ptr, B, loc, range); address.incrementWordIndex(loc, range); } diff --git a/tools/clang/test/CodeGenSPIRV/method.byte-address-buffer.load.double.capability.hlsl b/tools/clang/test/CodeGenSPIRV/method.byte-address-buffer.load.double.capability.hlsl new file mode 100644 index 0000000000..535bbecfe6 --- /dev/null +++ b/tools/clang/test/CodeGenSPIRV/method.byte-address-buffer.load.double.capability.hlsl @@ -0,0 +1,39 @@ +// RUN: %dxc -T cs_6_0 -E main -O0 %s -spirv | FileCheck %s + +// CHECK-NOT: OpCapability Int64 +// CHECK-DAG: OpCapability Float64 +// CHECK-NOT: OpCapability Int64 + +RWByteAddressBuffer buffer; + +[numthreads(1, 1, 1)] +void main() { + double tmp; + +// CHECK: [[addr1:%[0-9]+]] = OpShiftRightLogical %uint %uint_0 %uint_2 +// CHECK: [[ptr:%[0-9]+]] = OpAccessChain %_ptr_Uniform_uint %buffer %uint_0 [[addr1]] +// CHECK: [[word0:%[0-9]+]] = OpLoad %uint [[ptr]] +// CHECK: [[addr2:%[0-9]+]] = OpIAdd %uint [[addr1]] %uint_1 +// CHECK: [[ptr:%[0-9]+]] = OpAccessChain %_ptr_Uniform_uint %buffer %uint_0 [[addr2]] +// CHECK: [[word1:%[0-9]+]] = OpLoad %uint [[ptr]] +// CHECK: [[addr3:%[0-9]+]] = OpIAdd %uint [[addr2]] %uint_1 +// CHECK: [[merge:%[0-9]+]] = OpCompositeConstruct %v2uint [[word0]] [[word1]] +// CHECK: [[value:%[0-9]+]] = OpBitcast %double [[merge]] +// CHECK: OpStore %tmp [[value]] + tmp = buffer.Load(0); + +// CHECK: [[value:%[0-9]+]] = OpLoad %double %tmp +// CHECK: [[merge:%[0-9]+]] = OpBitcast %v2uint [[value]] +// CHECK: [[word0:%[0-9]+]] = OpCompositeExtract %uint [[merge]] 0 +// CHECK: [[word1:%[0-9]+]] = OpCompositeExtract %uint [[merge]] 1 + +// CHECK: [[addr1:%[0-9]+]] = OpShiftRightLogical %uint %uint_0 %uint_2 +// CHECK: [[ptr:%[0-9]+]] = OpAccessChain %_ptr_Uniform_uint %buffer %uint_0 [[addr1]] +// CHECK: OpStore [[ptr]] [[word0]] +// CHECK: [[addr2:%[0-9]+]] = OpIAdd %uint [[addr1]] %uint_1 +// CHECK: [[ptr:%[0-9]+]] = OpAccessChain %_ptr_Uniform_uint %buffer %uint_0 [[addr2]] +// CHECK: OpStore [[ptr]] [[word1]] +// CHECK: [[addr3:%[0-9]+]] = OpIAdd %uint [[addr2]] %uint_1 + buffer.Store(0, tmp); +} + diff --git a/tools/clang/test/CodeGenSPIRV/method.byte-address-buffer.templated-load.matrix.hlsl b/tools/clang/test/CodeGenSPIRV/method.byte-address-buffer.templated-load.matrix.hlsl index c4ac7bca5a..7a4c968f42 100644 --- a/tools/clang/test/CodeGenSPIRV/method.byte-address-buffer.templated-load.matrix.hlsl +++ b/tools/clang/test/CodeGenSPIRV/method.byte-address-buffer.templated-load.matrix.hlsl @@ -98,53 +98,46 @@ void main(uint3 tid : SV_DispatchThreadId) // ********* 64-bit matrix ******************** // CHECK: [[index_1:%[0-9]+]] = OpShiftRightLogical %uint [[addr0_1:%[0-9]+]] %uint_2 -// CHECK: [[ptr_11:%[0-9]+]] = OpAccessChain %_ptr_Uniform_uint %buf %uint_0 [[index_1]] -// CHECK: [[word0_2:%[0-9]+]] = OpLoad %uint [[ptr_11]] -// CHECK: [[index_1_2:%[0-9]+]] = OpIAdd %uint [[index_1]] %uint_1 -// CHECK: [[ptr_12:%[0-9]+]] = OpAccessChain %_ptr_Uniform_uint %buf %uint_0 [[index_1_2]] -// CHECK: [[word1_3:%[0-9]+]] = OpLoad %uint [[ptr_12]] -// CHECK: [[word0_ulong:%[0-9]+]] = OpUConvert %ulong [[word0_2]] -// CHECK: [[word1_ulong:%[0-9]+]] = OpUConvert %ulong [[word1_3]] -// CHECK: [[word1_ulong_shifted:%[0-9]+]] = OpShiftLeftLogical %ulong [[word1_ulong]] %uint_32 -// CHECK: [[val0_ulong:%[0-9]+]] = OpBitwiseOr %ulong [[word0_ulong]] [[word1_ulong_shifted]] -// CHECK: [[val0_1:%[0-9]+]] = OpBitcast %double [[val0_ulong]] -// CHECK: [[index_2_2:%[0-9]+]] = OpIAdd %uint [[index_1_2]] %uint_1 -// CHECK: [[ptr_13:%[0-9]+]] = OpAccessChain %_ptr_Uniform_uint %buf %uint_0 [[index_2_2]] -// CHECK: [[word2_2:%[0-9]+]] = OpLoad %uint [[ptr_13]] -// CHECK: [[index_3_0:%[0-9]+]] = OpIAdd %uint [[index_2_2]] %uint_1 -// CHECK: [[ptr_14:%[0-9]+]] = OpAccessChain %_ptr_Uniform_uint %buf %uint_0 [[index_3_0]] -// CHECK: [[word3_0:%[0-9]+]] = OpLoad %uint [[ptr_14]] -// CHECK: [[word2_ulong:%[0-9]+]] = OpUConvert %ulong [[word2_2]] -// CHECK: [[word3_ulong:%[0-9]+]] = OpUConvert %ulong [[word3_0]] -// CHECK: [[word3_ulong_shifted:%[0-9]+]] = OpShiftLeftLogical %ulong [[word3_ulong]] %uint_32 -// CHECK: [[val1_ulong:%[0-9]+]] = OpBitwiseOr %ulong [[word2_ulong]] [[word3_ulong_shifted]] -// CHECK: [[val1_1:%[0-9]+]] = OpBitcast %double [[val1_ulong]] -// CHECK: [[index_4_0:%[0-9]+]] = OpIAdd %uint [[index_3_0]] %uint_1 -// CHECK: [[ptr_15:%[0-9]+]] = OpAccessChain %_ptr_Uniform_uint %buf %uint_0 [[index_4_0]] -// CHECK: [[word4_0:%[0-9]+]] = OpLoad %uint [[ptr_15]] -// CHECK: [[index_5_0:%[0-9]+]] = OpIAdd %uint [[index_4_0]] %uint_1 -// CHECK: [[ptr_16:%[0-9]+]] = OpAccessChain %_ptr_Uniform_uint %buf %uint_0 [[index_5_0]] -// CHECK: [[word5_0:%[0-9]+]] = OpLoad %uint [[ptr_16]] -// CHECK: [[word4_ulong:%[0-9]+]] = OpUConvert %ulong [[word4_0]] -// CHECK: [[word5_ulong:%[0-9]+]] = OpUConvert %ulong [[word5_0]] -// CHECK: [[word5_ulong_shifted:%[0-9]+]] = OpShiftLeftLogical %ulong [[word5_ulong]] %uint_32 -// CHECK: [[val2_ulong:%[0-9]+]] = OpBitwiseOr %ulong [[word4_ulong]] [[word5_ulong_shifted]] -// CHECK: [[val2_1:%[0-9]+]] = OpBitcast %double [[val2_ulong]] +// CHECK: [[ptr_11:%[0-9]+]] = OpAccessChain %_ptr_Uniform_uint %buf %uint_0 [[index_1]] +// CHECK: [[word0_2:%[0-9]+]] = OpLoad %uint [[ptr_11]] +// CHECK: [[index_1_2:%[0-9]+]] = OpIAdd %uint [[index_1]] %uint_1 +// CHECK: [[ptr_12:%[0-9]+]] = OpAccessChain %_ptr_Uniform_uint %buf %uint_0 [[index_1_2]] +// CHECK: [[word1_3:%[0-9]+]] = OpLoad %uint [[ptr_12]] +// CHECK: [[index_2_2:%[0-9]+]] = OpIAdd %uint [[index_1_2]] %uint_1 +// CHECK: [[merge:%[0-9]+]] = OpCompositeConstruct %v2uint [[word0_2]] [[word1_3]] +// CHECK: [[val0_1:%[0-9]+]] = OpBitcast %double [[merge]] + +// CHECK: [[ptr_13:%[0-9]+]] = OpAccessChain %_ptr_Uniform_uint %buf %uint_0 [[index_2_2]] +// CHECK: [[word2_2:%[0-9]+]] = OpLoad %uint [[ptr_13]] +// CHECK: [[index_3_0:%[0-9]+]] = OpIAdd %uint [[index_2_2]] %uint_1 +// CHECK: [[ptr_14:%[0-9]+]] = OpAccessChain %_ptr_Uniform_uint %buf %uint_0 [[index_3_0]] +// CHECK: [[word3_0:%[0-9]+]] = OpLoad %uint [[ptr_14]] +// CHECK: [[index_4_0:%[0-9]+]] = OpIAdd %uint [[index_3_0]] %uint_1 +// CHECK: [[merge:%[0-9]+]] = OpCompositeConstruct %v2uint [[word2_2]] [[word3_0]] +// CHECK: [[val1_1:%[0-9]+]] = OpBitcast %double [[merge]] + +// CHECK: [[ptr_15:%[0-9]+]] = OpAccessChain %_ptr_Uniform_uint %buf %uint_0 [[index_4_0]] +// CHECK: [[word4_0:%[0-9]+]] = OpLoad %uint [[ptr_15]] +// CHECK: [[index_5_0:%[0-9]+]] = OpIAdd %uint [[index_4_0]] %uint_1 +// CHECK: [[ptr_16:%[0-9]+]] = OpAccessChain %_ptr_Uniform_uint %buf %uint_0 [[index_5_0]] +// CHECK: [[word5_0:%[0-9]+]] = OpLoad %uint [[ptr_16]] // CHECK: [[index_6:%[0-9]+]] = OpIAdd %uint [[index_5_0]] %uint_1 -// CHECK: [[ptr_17:%[0-9]+]] = OpAccessChain %_ptr_Uniform_uint %buf %uint_0 [[index_6]] +// CHECK: [[merge:%[0-9]+]] = OpCompositeConstruct %v2uint [[word4_0]] [[word5_0]] +// CHECK: [[val2_1:%[0-9]+]] = OpBitcast %double [[merge]] + +// CHECK: [[ptr_17:%[0-9]+]] = OpAccessChain %_ptr_Uniform_uint %buf %uint_0 [[index_6]] // CHECK: [[word6:%[0-9]+]] = OpLoad %uint [[ptr_17]] // CHECK: [[index_7:%[0-9]+]] = OpIAdd %uint [[index_6]] %uint_1 -// CHECK: [[ptr_18:%[0-9]+]] = OpAccessChain %_ptr_Uniform_uint %buf %uint_0 [[index_7]] +// CHECK: [[ptr_18:%[0-9]+]] = OpAccessChain %_ptr_Uniform_uint %buf %uint_0 [[index_7]] // CHECK: [[word7:%[0-9]+]] = OpLoad %uint [[ptr_18]] -// CHECK: [[word6_ulong:%[0-9]+]] = OpUConvert %ulong [[word6]] -// CHECK: [[word7_ulong:%[0-9]+]] = OpUConvert %ulong [[word7]] -// CHECK: [[word7_ulong_shifted:%[0-9]+]] = OpShiftLeftLogical %ulong [[word7_ulong]] %uint_32 -// CHECK: [[val3_ulong:%[0-9]+]] = OpBitwiseOr %ulong [[word6_ulong]] [[word7_ulong_shifted]] -// CHECK: [[val3_1:%[0-9]+]] = OpBitcast %double [[val3_ulong]] -// CHECK: [[row0_1:%[0-9]+]] = OpCompositeConstruct %v2double [[val0_1]] [[val2_1]] -// CHECK: [[row1_1:%[0-9]+]] = OpCompositeConstruct %v2double [[val1_1]] [[val3_1]] -// CHECK: [[matrix_1:%[0-9]+]] = OpCompositeConstruct %mat2v2double [[row0_1]] [[row1_1]] -// CHECK: OpStore %f64 [[matrix_1]] +// CHECK: [[index_8:%[0-9]+]] = OpIAdd %uint [[index_7]] %uint_1 +// CHECK: [[merge:%[0-9]+]] = OpCompositeConstruct %v2uint [[word6]] [[word7]] +// CHECK: [[val3_1:%[0-9]+]] = OpBitcast %double [[merge]] + +// CHECK: [[row0_1:%[0-9]+]] = OpCompositeConstruct %v2double [[val0_1]] [[val2_1]] +// CHECK: [[row1_1:%[0-9]+]] = OpCompositeConstruct %v2double [[val1_1]] [[val3_1]] +// CHECK: [[matrix_1:%[0-9]+]] = OpCompositeConstruct %mat2v2double [[row0_1]] [[row1_1]] +// CHECK: OpStore %f64 [[matrix_1]] float64_t2x2 f64 = buf.Load(tid.x); // ********* array of matrices ******************** diff --git a/tools/clang/test/CodeGenSPIRV/method.byte-address-buffer.templated-load.scalar.hlsl b/tools/clang/test/CodeGenSPIRV/method.byte-address-buffer.templated-load.scalar.hlsl index a1a3e7694a..96b20034b3 100644 --- a/tools/clang/test/CodeGenSPIRV/method.byte-address-buffer.templated-load.scalar.hlsl +++ b/tools/clang/test/CodeGenSPIRV/method.byte-address-buffer.templated-load.scalar.hlsl @@ -80,42 +80,34 @@ ByteAddressBuffer buf; // ********* 64-bit scalar ******************** -// CHECK: [[ptr_9:%[0-9]+]] = OpAccessChain %_ptr_Uniform_uint %buf %uint_0 [[addr:%[0-9]+]] -// CHECK: [[word0:%[0-9]+]] = OpLoad %uint [[ptr_9]] -// CHECK: [[newAddr:%[0-9]+]] = OpIAdd %uint [[addr]] %uint_1 -// CHECK: [[ptr_10:%[0-9]+]] = OpAccessChain %_ptr_Uniform_uint %buf %uint_0 [[newAddr]] -// CHECK: [[word1:%[0-9]+]] = OpLoad %uint [[ptr_10]] -// CHECK: [[word0ULong:%[0-9]+]] = OpUConvert %ulong [[word0]] -// CHECK: [[word1ULong:%[0-9]+]] = OpUConvert %ulong [[word1]] -// CHECK:[[shiftedWord1ULong:%[0-9]+]] = OpShiftLeftLogical %ulong [[word1ULong]] %uint_32 -// CHECK: [[val:%[0-9]+]] = OpBitwiseOr %ulong [[word0ULong]] [[shiftedWord1ULong]] -// CHECK: OpStore %u64 [[val]] +// CHECK: [[ptr_9:%[0-9]+]] = OpAccessChain %_ptr_Uniform_uint %buf %uint_0 [[addr:%[0-9]+]] +// CHECK: [[word0:%[0-9]+]] = OpLoad %uint [[ptr_9]] +// CHECK: [[newAddr:%[0-9]+]] = OpIAdd %uint [[addr]] %uint_1 +// CHECK: [[ptr_10:%[0-9]+]] = OpAccessChain %_ptr_Uniform_uint %buf %uint_0 [[newAddr]] +// CHECK: [[word1:%[0-9]+]] = OpLoad %uint [[ptr_10]] +// CHECK: [[merge:%[0-9]+]] = OpCompositeConstruct %v2uint [[word0]] [[word1]] +// CHECK: [[val:%[0-9]+]] = OpBitcast %ulong [[merge]] +// CHECK: OpStore %u64 [[val]] uint64_t u64 = buf.Load(tid.x); -// CHECK: [[ptr_11:%[0-9]+]] = OpAccessChain %_ptr_Uniform_uint %buf %uint_0 [[addr_0:%[0-9]+]] -// CHECK: [[word0_0:%[0-9]+]] = OpLoad %uint [[ptr_11]] -// CHECK: [[newAddr_0:%[0-9]+]] = OpIAdd %uint [[addr_0]] %uint_1 -// CHECK: [[ptr_12:%[0-9]+]] = OpAccessChain %_ptr_Uniform_uint %buf %uint_0 [[newAddr_0]] -// CHECK: [[word1_0:%[0-9]+]] = OpLoad %uint [[ptr_12]] -// CHECK: [[word0Long:%[0-9]+]] = OpUConvert %ulong [[word0_0]] -// CHECK: [[word1Long:%[0-9]+]] = OpUConvert %ulong [[word1_0]] -// CHECK: [[shiftedWord1Long:%[0-9]+]] = OpShiftLeftLogical %ulong [[word1Long]] %uint_32 -// CHECK: [[val_ulong:%[0-9]+]] = OpBitwiseOr %ulong [[word0Long]] [[shiftedWord1Long]] -// CHECK: [[val_long:%[0-9]+]] = OpBitcast %long [[val_ulong]] -// CHECK: OpStore %i64 [[val_long]] +// CHECK: [[ptr_11:%[0-9]+]] = OpAccessChain %_ptr_Uniform_uint %buf %uint_0 [[addr_0:%[0-9]+]] +// CHECK: [[word0_0:%[0-9]+]] = OpLoad %uint [[ptr_11]] +// CHECK: [[newAddr_0:%[0-9]+]] = OpIAdd %uint [[addr_0]] %uint_1 +// CHECK: [[ptr_12:%[0-9]+]] = OpAccessChain %_ptr_Uniform_uint %buf %uint_0 [[newAddr_0]] +// CHECK: [[word1_0:%[0-9]+]] = OpLoad %uint [[ptr_12]] +// CHECK: [[merge:%[0-9]+]] = OpCompositeConstruct %v2uint [[word0_0]] [[word1_0]] +// CHECK: [[val_long:%[0-9]+]] = OpBitcast %long [[merge]] +// CHECK: OpStore %i64 [[val_long]] int64_t i64 = buf.Load(tid.x); -// CHECK: [[ptr_13:%[0-9]+]] = OpAccessChain %_ptr_Uniform_uint %buf %uint_0 [[addr_1:%[0-9]+]] -// CHECK: [[word0_1:%[0-9]+]] = OpLoad %uint [[ptr_13]] -// CHECK: [[newAddr_1:%[0-9]+]] = OpIAdd %uint [[addr_1]] %uint_1 -// CHECK: [[ptr_14:%[0-9]+]] = OpAccessChain %_ptr_Uniform_uint %buf %uint_0 [[newAddr_1]] -// CHECK: [[word1_1:%[0-9]+]] = OpLoad %uint [[ptr_14]] -// CHECK: [[word0Long_0:%[0-9]+]] = OpUConvert %ulong [[word0_1]] -// CHECK: [[word1Long_0:%[0-9]+]] = OpUConvert %ulong [[word1_1]] -// CHECK: [[shiftedWord1Long_0:%[0-9]+]] = OpShiftLeftLogical %ulong [[word1Long_0]] %uint_32 -// CHECK: [[val_ulong_0:%[0-9]+]] = OpBitwiseOr %ulong [[word0Long_0]] [[shiftedWord1Long_0]] -// CHECK: [[val_double:%[0-9]+]] = OpBitcast %double [[val_ulong_0]] -// CHECK: OpStore %f64 [[val_double]] +// CHECK: [[ptr_13:%[0-9]+]] = OpAccessChain %_ptr_Uniform_uint %buf %uint_0 [[addr_1:%[0-9]+]] +// CHECK: [[word0_1:%[0-9]+]] = OpLoad %uint [[ptr_13]] +// CHECK: [[newAddr_1:%[0-9]+]] = OpIAdd %uint [[addr_1]] %uint_1 +// CHECK: [[ptr_14:%[0-9]+]] = OpAccessChain %_ptr_Uniform_uint %buf %uint_0 [[newAddr_1]] +// CHECK: [[word1_1:%[0-9]+]] = OpLoad %uint [[ptr_14]] +// CHECK: [[merge:%[0-9]+]] = OpCompositeConstruct %v2uint [[word0_1]] [[word1_1]] +// CHECK: [[val_double:%[0-9]+]] = OpBitcast %double [[merge]] +// CHECK: OpStore %f64 [[val_double]] double f64 = buf.Load(tid.x); // ********* array of scalars ***************** @@ -124,68 +116,63 @@ ByteAddressBuffer buf; // CHECK: [[index0:%[0-9]+]] = OpShiftRightLogical %uint [[addr0:%[0-9]+]] %uint_2 // CHECK: [[byteOff0:%[0-9]+]] = OpUMod %uint [[addr0]] %uint_4 // CHECK: [[bitOff0:%[0-9]+]] = OpShiftLeftLogical %uint [[byteOff0]] %uint_3 -// CHECK: [[ptr_15:%[0-9]+]] = OpAccessChain %_ptr_Uniform_uint %buf %uint_0 [[index0]] -// CHECK: [[word0_2:%[0-9]+]] = OpLoad %uint [[ptr_15]] -// CHECK: [[shift_4:%[0-9]+]] = OpShiftRightLogical %uint [[word0_2]] [[bitOff0]] +// CHECK: [[ptr_15:%[0-9]+]] = OpAccessChain %_ptr_Uniform_uint %buf %uint_0 [[index0]] +// CHECK: [[word0_2:%[0-9]+]] = OpLoad %uint [[ptr_15]] +// CHECK: [[shift_4:%[0-9]+]] = OpShiftRightLogical %uint [[word0_2]] [[bitOff0]] // CHECK: [[val0:%[0-9]+]] = OpUConvert %ushort [[shift_4]] // CHECK: [[addr1:%[0-9]+]] = OpIAdd %uint [[addr0]] %uint_2 // CHECK: [[index1:%[0-9]+]] = OpShiftRightLogical %uint [[addr1]] %uint_2 // CHECK: [[byteOff1:%[0-9]+]] = OpUMod %uint [[addr1]] %uint_4 // CHECK: [[bitOff1:%[0-9]+]] = OpShiftLeftLogical %uint [[byteOff1]] %uint_3 -// CHECK: [[ptr_16:%[0-9]+]] = OpAccessChain %_ptr_Uniform_uint %buf %uint_0 [[index1]] -// CHECK: [[word0_3:%[0-9]+]] = OpLoad %uint [[ptr_16]] +// CHECK: [[ptr_16:%[0-9]+]] = OpAccessChain %_ptr_Uniform_uint %buf %uint_0 [[index1]] +// CHECK: [[word0_3:%[0-9]+]] = OpLoad %uint [[ptr_16]] // CHECK: [[val1uint:%[0-9]+]] = OpShiftRightLogical %uint [[word0_3]] [[bitOff1]] // CHECK: [[val1:%[0-9]+]] = OpUConvert %ushort [[val1uint]] // CHECK: [[addr2:%[0-9]+]] = OpIAdd %uint [[addr1]] %uint_2 // CHECK: [[index2:%[0-9]+]] = OpShiftRightLogical %uint [[addr2]] %uint_2 // CHECK: [[byteOff2:%[0-9]+]] = OpUMod %uint [[addr2]] %uint_4 // CHECK: [[bitOff2:%[0-9]+]] = OpShiftLeftLogical %uint [[byteOff2]] %uint_3 -// CHECK: [[ptr_17:%[0-9]+]] = OpAccessChain %_ptr_Uniform_uint %buf %uint_0 [[index2]] -// CHECK: [[word1_2:%[0-9]+]] = OpLoad %uint [[ptr_17]] -// CHECK: [[shift_5:%[0-9]+]] = OpShiftRightLogical %uint [[word1_2]] [[bitOff2]] +// CHECK: [[ptr_17:%[0-9]+]] = OpAccessChain %_ptr_Uniform_uint %buf %uint_0 [[index2]] +// CHECK: [[word1_2:%[0-9]+]] = OpLoad %uint [[ptr_17]] +// CHECK: [[shift_5:%[0-9]+]] = OpShiftRightLogical %uint [[word1_2]] [[bitOff2]] // CHECK: [[val2:%[0-9]+]] = OpUConvert %ushort [[shift_5]] // CHECK: [[uArr:%[0-9]+]] = OpCompositeConstruct %_arr_ushort_uint_3 [[val0]] [[val1]] [[val2]] -// CHECK: OpStore %uArr [[uArr]] +// CHECK: OpStore %uArr [[uArr]] uint16_t uArr[3] = buf.Load(tid.x); // CHECK: [[index_1:%[0-9]+]] = OpShiftRightLogical %uint [[addr_2:%[0-9]+]] %uint_2 -// CHECK: [[ptr_18:%[0-9]+]] = OpAccessChain %_ptr_Uniform_uint %buf %uint_0 [[index_1]] -// CHECK: [[val0_uint:%[0-9]+]] = OpLoad %uint [[ptr_18]] +// CHECK: [[ptr_18:%[0-9]+]] = OpAccessChain %_ptr_Uniform_uint %buf %uint_0 [[index_1]] +// CHECK: [[val0_uint:%[0-9]+]] = OpLoad %uint [[ptr_18]] // CHECK: [[val0_0:%[0-9]+]] = OpBitcast %int [[val0_uint]] -// CHECK: [[newIndex:%[0-9]+]] = OpIAdd %uint [[index_1]] %uint_1 -// CHECK: [[ptr_19:%[0-9]+]] = OpAccessChain %_ptr_Uniform_uint %buf %uint_0 [[newIndex]] -// CHECK: [[val1_uint:%[0-9]+]] = OpLoad %uint [[ptr_19]] +// CHECK: [[newIndex:%[0-9]+]] = OpIAdd %uint [[index_1]] %uint_1 +// CHECK: [[ptr_19:%[0-9]+]] = OpAccessChain %_ptr_Uniform_uint %buf %uint_0 [[newIndex]] +// CHECK: [[val1_uint:%[0-9]+]] = OpLoad %uint [[ptr_19]] // CHECK: [[val1_0:%[0-9]+]] = OpBitcast %int [[val1_uint]] -// CHECK: [[iArr:%[0-9]+]] = OpCompositeConstruct %_arr_int_uint_2 [[val0_0]] [[val1_0]] -// CHECK: OpStore %iArr [[iArr]] +// CHECK: [[iArr:%[0-9]+]] = OpCompositeConstruct %_arr_int_uint_2 [[val0_0]] [[val1_0]] +// CHECK: OpStore %iArr [[iArr]] int iArr[2] = buf.Load(tid.x); -// CHECK: [[index_0:%[0-9]+]] = OpShiftRightLogical %uint [[addr_0:%[0-9]+]] %uint_2 -// CHECK: [[ptr_20:%[0-9]+]] = OpAccessChain %_ptr_Uniform_uint %buf %uint_0 [[index_0]] -// CHECK: [[val0_word0_uint:%[0-9]+]] = OpLoad %uint [[ptr_20]] -// CHECK: [[index_1:%[0-9]+]] = OpIAdd %uint [[index_0]] %uint_1 -// CHECK: [[ptr_21:%[0-9]+]] = OpAccessChain %_ptr_Uniform_uint %buf %uint_0 [[index_1]] -// CHECK: [[val0_word1_uint:%[0-9]+]] = OpLoad %uint [[ptr_21]] -// CHECK: [[val0_word0_ulong:%[0-9]+]] = OpUConvert %ulong [[val0_word0_uint]] -// CHECK: [[val0_word1_ulong:%[0-9]+]] = OpUConvert %ulong [[val0_word1_uint]] -// CHECK: [[shifted_val0_word1_ulong:%[0-9]+]] = OpShiftLeftLogical %ulong [[val0_word1_ulong]] %uint_32 -// CHECK: [[val0_ulong:%[0-9]+]] = OpBitwiseOr %ulong [[val0_word0_ulong]] [[shifted_val0_word1_ulong]] -// CHECK: [[val0_double:%[0-9]+]] = OpBitcast %double [[val0_ulong]] +// CHECK: [[index_0:%[0-9]+]] = OpShiftRightLogical %uint [[addr_0:%[0-9]+]] %uint_2 +// CHECK: [[ptr_20:%[0-9]+]] = OpAccessChain %_ptr_Uniform_uint %buf %uint_0 [[index_0]] +// CHECK: [[val0_word0_uint:%[0-9]+]] = OpLoad %uint [[ptr_20]] +// CHECK: [[index_1:%[0-9]+]] = OpIAdd %uint [[index_0]] %uint_1 +// CHECK: [[ptr_21:%[0-9]+]] = OpAccessChain %_ptr_Uniform_uint %buf %uint_0 [[index_1]] +// CHECK: [[val0_word1_uint:%[0-9]+]] = OpLoad %uint [[ptr_21]] +// CHECK: [[index_2:%[0-9]+]] = OpIAdd %uint [[index_1]] %uint_1 +// CHECK: [[merge:%[0-9]+]] = OpCompositeConstruct %v2uint [[val0_word0_uint]] [[val0_word1_uint]] +// CHECK: [[val0_double:%[0-9]+]] = OpBitcast %double [[merge]] + +// CHECK: [[ptr_22:%[0-9]+]] = OpAccessChain %_ptr_Uniform_uint %buf %uint_0 [[index_2]] +// CHECK: [[val1_word0_uint:%[0-9]+]] = OpLoad %uint [[ptr_22]] +// CHECK: [[index_3:%[0-9]+]] = OpIAdd %uint [[index_2]] %uint_1 +// CHECK: [[ptr_23:%[0-9]+]] = OpAccessChain %_ptr_Uniform_uint %buf %uint_0 [[index_3]] +// CHECK: [[val1_word1_uint:%[0-9]+]] = OpLoad %uint [[ptr_23]] +// CHECK: [[index_4:%[0-9]+]] = OpIAdd %uint [[index_3]] %uint_1 +// CHECK: [[merge:%[0-9]+]] = OpCompositeConstruct %v2uint [[val1_word0_uint]] [[val1_word1_uint]] +// CHECK: [[val1_double:%[0-9]+]] = OpBitcast %double [[merge]] // -// CHECK: [[index_2:%[0-9]+]] = OpIAdd %uint [[index_1]] %uint_1 -// CHECK: [[ptr_22:%[0-9]+]] = OpAccessChain %_ptr_Uniform_uint %buf %uint_0 [[index_2]] -// CHECK: [[val1_word0_uint:%[0-9]+]] = OpLoad %uint [[ptr_22]] -// CHECK: [[index_3:%[0-9]+]] = OpIAdd %uint [[index_2]] %uint_1 -// CHECK: [[ptr_23:%[0-9]+]] = OpAccessChain %_ptr_Uniform_uint %buf %uint_0 [[index_3]] -// CHECK: [[val1_word1_uint:%[0-9]+]] = OpLoad %uint [[ptr_23]] -// CHECK: [[val1_word0_ulong:%[0-9]+]] = OpUConvert %ulong [[val1_word0_uint]] -// CHECK: [[val1_word1_ulong:%[0-9]+]] = OpUConvert %ulong [[val1_word1_uint]] -// CHECK: [[shifted_val1_word1_ulong:%[0-9]+]] = OpShiftLeftLogical %ulong [[val1_word1_ulong]] %uint_32 -// CHECK: [[val1_ulong:%[0-9]+]] = OpBitwiseOr %ulong [[val1_word0_ulong]] [[shifted_val1_word1_ulong]] -// CHECK: [[val1_double:%[0-9]+]] = OpBitcast %double [[val1_ulong]] -// -// CHECK: [[fArr:%[0-9]+]] = OpCompositeConstruct %_arr_double_uint_2 [[val0_double]] [[val1_double]] -// CHECK: OpStore %fArr [[fArr]] +// CHECK: [[fArr:%[0-9]+]] = OpCompositeConstruct %_arr_double_uint_2 [[val0_double]] [[val1_double]] +// CHECK: OpStore %fArr [[fArr]] double fArr[2] = buf.Load(tid.x); } diff --git a/tools/clang/test/CodeGenSPIRV/method.byte-address-buffer.templated-load.vector.hlsl b/tools/clang/test/CodeGenSPIRV/method.byte-address-buffer.templated-load.vector.hlsl index 07f76aad6a..16702c0e37 100644 --- a/tools/clang/test/CodeGenSPIRV/method.byte-address-buffer.templated-load.vector.hlsl +++ b/tools/clang/test/CodeGenSPIRV/method.byte-address-buffer.templated-load.vector.hlsl @@ -65,30 +65,25 @@ void main(uint3 tid : SV_DispatchThreadId) // ********* 64-bit vector ******************** -// CHECK: [[index_3:%[0-9]+]] = OpShiftRightLogical %uint [[addr0_2:%[0-9]+]] %uint_2 -// CHECK: [[ptr_6:%[0-9]+]] = OpAccessChain %_ptr_Uniform_uint %buf %uint_0 [[index_3]] -// CHECK: [[word0_3:%[0-9]+]] = OpLoad %uint [[ptr_6]] -// CHECK: [[index_1_2:%[0-9]+]] = OpIAdd %uint [[index_3]] %uint_1 -// CHECK: [[ptr_7:%[0-9]+]] = OpAccessChain %_ptr_Uniform_uint %buf %uint_0 [[index_1_2]] -// CHECK: [[word1_2:%[0-9]+]] = OpLoad %uint [[ptr_7]] -// CHECK: [[word0_ulong:%[0-9]+]] = OpUConvert %ulong [[word0_3]] -// CHECK: [[word1_ulong:%[0-9]+]] = OpUConvert %ulong [[word1_2]] -// CHECK: [[shifted_word1_ulong:%[0-9]+]] = OpShiftLeftLogical %ulong [[word1_ulong]] %uint_32 -// CHECK: [[val0_ulong:%[0-9]+]] = OpBitwiseOr %ulong [[word0_ulong]] [[shifted_word1_ulong]] -// CHECK: [[val0_2:%[0-9]+]] = OpBitcast %double [[val0_ulong]] -// CHECK: [[index_2_0:%[0-9]+]] = OpIAdd %uint [[index_1_2]] %uint_1 -// CHECK: [[ptr_8:%[0-9]+]] = OpAccessChain %_ptr_Uniform_uint %buf %uint_0 [[index_2_0]] -// CHECK: [[word0_4:%[0-9]+]] = OpLoad %uint [[ptr_8]] -// CHECK: [[index_3:%[0-9]+]] = OpIAdd %uint [[index_2_0]] %uint_1 -// CHECK: [[ptr_9:%[0-9]+]] = OpAccessChain %_ptr_Uniform_uint %buf %uint_0 [[index_3]] -// CHECK: [[word1_3:%[0-9]+]] = OpLoad %uint [[ptr_9]] -// CHECK: [[word0_ulong_0:%[0-9]+]] = OpUConvert %ulong [[word0_4]] -// CHECK: [[word1_ulong_0:%[0-9]+]] = OpUConvert %ulong [[word1_3]] -// CHECK: [[shifted_word1_ulong_0:%[0-9]+]] = OpShiftLeftLogical %ulong [[word1_ulong_0]] %uint_32 -// CHECK: [[val1_ulong:%[0-9]+]] = OpBitwiseOr %ulong [[word0_ulong_0]] [[shifted_word1_ulong_0]] -// CHECK: [[val1_2:%[0-9]+]] = OpBitcast %double [[val1_ulong]] -// CHECK: [[fVec:%[0-9]+]] = OpCompositeConstruct %v2double [[val0_2]] [[val1_2]] -// CHECK: OpStore %f64 [[fVec]] +// CHECK: [[index_3:%[0-9]+]] = OpShiftRightLogical %uint [[addr0_2:%[0-9]+]] %uint_2 +// CHECK: [[ptr_6:%[0-9]+]] = OpAccessChain %_ptr_Uniform_uint %buf %uint_0 [[index_3]] +// CHECK: [[word0_3:%[0-9]+]] = OpLoad %uint [[ptr_6]] +// CHECK: [[index_1_2:%[0-9]+]] = OpIAdd %uint [[index_3]] %uint_1 +// CHECK: [[ptr_7:%[0-9]+]] = OpAccessChain %_ptr_Uniform_uint %buf %uint_0 [[index_1_2]] +// CHECK: [[word1_2:%[0-9]+]] = OpLoad %uint [[ptr_7]] +// CHECK: [[index_2_0:%[0-9]+]] = OpIAdd %uint [[index_1_2]] %uint_1 +// CHECK: [[merge:%[0-9]+]] = OpCompositeConstruct %v2uint [[word0_3]] [[word1_2]] +// CHECK: [[val0_2:%[0-9]+]] = OpBitcast %double [[merge]] +// CHECK: [[ptr_8:%[0-9]+]] = OpAccessChain %_ptr_Uniform_uint %buf %uint_0 [[index_2_0]] +// CHECK: [[word0_4:%[0-9]+]] = OpLoad %uint [[ptr_8]] +// CHECK: [[index_3:%[0-9]+]] = OpIAdd %uint [[index_2_0]] %uint_1 +// CHECK: [[ptr_9:%[0-9]+]] = OpAccessChain %_ptr_Uniform_uint %buf %uint_0 [[index_3]] +// CHECK: [[word1_3:%[0-9]+]] = OpLoad %uint [[ptr_9]] +// CHECK: [[index_4:%[0-9]+]] = OpIAdd %uint [[index_3]] %uint_1 +// CHECK: [[merge:%[0-9]+]] = OpCompositeConstruct %v2uint [[word0_4]] [[word1_3]] +// CHECK: [[val1_2:%[0-9]+]] = OpBitcast %double [[merge]] +// CHECK: [[fVec:%[0-9]+]] = OpCompositeConstruct %v2double [[val0_2]] [[val1_2]] +// CHECK: OpStore %f64 [[fVec]] float64_t2 f64 = buf.Load(tid.x); // ********* array of vectors ******************** diff --git a/tools/clang/test/CodeGenSPIRV/method.byte-address-buffer.templated-store.struct.hlsl b/tools/clang/test/CodeGenSPIRV/method.byte-address-buffer.templated-store.struct.hlsl index 3bf947afe2..10c978e44d 100644 --- a/tools/clang/test/CodeGenSPIRV/method.byte-address-buffer.templated-store.struct.hlsl +++ b/tools/clang/test/CodeGenSPIRV/method.byte-address-buffer.templated-store.struct.hlsl @@ -226,18 +226,20 @@ void main(uint3 tid : SV_DispatchThreadId) { // The second member of S starts at byte offset 24 (6 words) // // CHECK: [[c_addr:%[0-9]+]] = OpIAdd %uint [[base_addr]] %uint_24 -// + // CHECK: [[c:%[0-9]+]] = OpCompositeExtract %double [[s0]] 1 +// CHECK: [[merge:%[0-9]+]] = OpBitcast %v2uint [[c]] +// CHECK: [[c_word0:%[0-9]+]] = OpCompositeExtract %uint [[merge]] 0 +// CHECK: [[c_word1:%[0-9]+]] = OpCompositeExtract %uint [[merge]] 1 + // CHECK: [[c_index:%[0-9]+]] = OpShiftRightLogical %uint [[c_addr]] %uint_2 -// CHECK: [[ptr_4:%[0-9]+]] = OpAccessChain %_ptr_Uniform_uint %buf2 %uint_0 [[c_index]] -// CHECK: [[c_u64:%[0-9]+]] = OpBitcast %ulong [[c]] -// CHECK: [[c_word0:%[0-9]+]] = OpUConvert %uint [[c_u64]] -// CHECK: [[c_u64_shifted:%[0-9]+]] = OpShiftRightLogical %ulong [[c_u64]] %uint_32 -// CHECK: [[c_word1:%[0-9]+]] = OpUConvert %uint [[c_u64_shifted]] -// CHECK: OpStore [[ptr_4]] [[c_word0]] +// CHECK: [[ptr_4:%[0-9]+]] = OpAccessChain %_ptr_Uniform_uint %buf2 %uint_0 [[c_index]] +// CHECK: OpStore [[ptr_4]] [[c_word0]] // CHECK: [[c_msb_index:%[0-9]+]] = OpIAdd %uint [[c_index]] %uint_1 -// CHECK: [[ptr_5:%[0-9]+]] = OpAccessChain %_ptr_Uniform_uint %buf2 %uint_0 [[c_msb_index]] -// CHECK: OpStore [[ptr_5]] [[c_word1]] + +// CHECK: [[ptr_5:%[0-9]+]] = OpAccessChain %_ptr_Uniform_uint %buf2 %uint_0 [[c_msb_index]] +// CHECK: OpStore [[ptr_5]] [[c_word1]] +// CHECK: [[next_index:%[0-9]+]] = OpIAdd %uint [[c_msb_index]] %uint_1 // // The third member of S starts at byte offset 32 (8 words) @@ -305,16 +307,17 @@ void main(uint3 tid : SV_DispatchThreadId) { // CHECK: [[b_addr:%[0-9]+]] = OpIAdd %uint [[base_addr]] %uint_48 // // CHECK: [[b:%[0-9]+]] = OpCompositeExtract %double [[s0]] 3 +// CHECK: [[merge:%[0-9]+]] = OpBitcast %v2uint [[b]] +// CHECK: [[b_word0:%[0-9]+]] = OpCompositeExtract %uint [[merge]] 0 +// CHECK: [[b_word1:%[0-9]+]] = OpCompositeExtract %uint [[merge]] 1 + // CHECK: [[b_index:%[0-9]+]] = OpShiftRightLogical %uint [[b_addr]] %uint_2 -// CHECK: [[ptr_9:%[0-9]+]] = OpAccessChain %_ptr_Uniform_uint %buf2 %uint_0 [[b_index]] -// CHECK: [[b_u64:%[0-9]+]] = OpBitcast %ulong [[b]] -// CHECK: [[b_word0:%[0-9]+]] = OpUConvert %uint [[b_u64]] -// CHECK: [[b_u64_shifted:%[0-9]+]] = OpShiftRightLogical %ulong [[b_u64]] %uint_32 -// CHECK: [[b_word1:%[0-9]+]] = OpUConvert %uint [[b_u64_shifted]] -// CHECK: OpStore [[ptr_9]] [[b_word0]] +// CHECK: [[ptr_9:%[0-9]+]] = OpAccessChain %_ptr_Uniform_uint %buf2 %uint_0 [[b_index]] +// CHECK: OpStore [[ptr_9]] [[b_word0]] // CHECK: [[b_msb_index:%[0-9]+]] = OpIAdd %uint [[b_index]] %uint_1 -// CHECK: [[ptr_10:%[0-9]+]] = OpAccessChain %_ptr_Uniform_uint %buf2 %uint_0 [[b_msb_index]] -// CHECK: OpStore [[ptr_10]] [[b_word1]] +// CHECK: [[ptr_10:%[0-9]+]] = OpAccessChain %_ptr_Uniform_uint %buf2 %uint_0 [[b_msb_index]] +// CHECK: OpStore [[ptr_10]] [[b_word1]] +// CHECK: [[next_index:%[0-9]+]] = OpIAdd %uint [[b_msb_index]] %uint_1 // // The fifth member of S starts at byte offset 56 (14 words) @@ -651,19 +654,20 @@ void main(uint3 tid : SV_DispatchThreadId) { // // The second member of S starts at byte offset 24 (6 words) // -// CHECK: [[c_addr_0:%[0-9]+]] = OpIAdd %uint [[s1_addr]] %uint_24 +// CHECK: [[c_addr_0:%[0-9]+]] = OpIAdd %uint [[s1_addr]] %uint_24 // -// CHECK: [[c_0:%[0-9]+]] = OpCompositeExtract %double [[s1]] 1 -// CHECK: [[c_index_0:%[0-9]+]] = OpShiftRightLogical %uint [[c_addr_0]] %uint_2 -// CHECK: [[ptr_28:%[0-9]+]] = OpAccessChain %_ptr_Uniform_uint %buf2 %uint_0 [[c_index_0]] -// CHECK: [[c_u64_0:%[0-9]+]] = OpBitcast %ulong [[c_0]] -// CHECK: [[c_word0_0:%[0-9]+]] = OpUConvert %uint [[c_u64_0]] -// CHECK: [[c_u64_shifted_0:%[0-9]+]] = OpShiftRightLogical %ulong [[c_u64_0]] %uint_32 -// CHECK: [[c_word1_0:%[0-9]+]] = OpUConvert %uint [[c_u64_shifted_0]] -// CHECK: OpStore [[ptr_28]] [[c_word0_0]] -// CHECK: [[c_msb_index_0:%[0-9]+]] = OpIAdd %uint [[c_index_0]] %uint_1 -// CHECK: [[ptr_29:%[0-9]+]] = OpAccessChain %_ptr_Uniform_uint %buf2 %uint_0 [[c_msb_index_0]] -// CHECK: OpStore [[ptr_29]] [[c_word1_0]] +// CHECK: [[c_0:%[0-9]+]] = OpCompositeExtract %double [[s1]] 1 +// CHECK: [[merge:%[0-9]+]] = OpBitcast %v2uint [[c_0]] +// CHECK: [[c_word0_0:%[0-9]+]] = OpCompositeExtract %uint [[merge]] 0 +// CHECK: [[c_word1_0:%[0-9]+]] = OpCompositeExtract %uint [[merge]] 1 + +// CHECK: [[c_index_0:%[0-9]+]] = OpShiftRightLogical %uint [[c_addr_0]] %uint_2 +// CHECK: [[ptr_28:%[0-9]+]] = OpAccessChain %_ptr_Uniform_uint %buf2 %uint_0 [[c_index_0]] +// CHECK: OpStore [[ptr_28]] [[c_word0_0]] +// CHECK: [[c_msb_index_0:%[0-9]+]] = OpIAdd %uint [[c_index_0]] %uint_1 +// CHECK: [[ptr_29:%[0-9]+]] = OpAccessChain %_ptr_Uniform_uint %buf2 %uint_0 [[c_msb_index_0]] +// CHECK: OpStore [[ptr_29]] [[c_word1_0]] +// CHECK: [[index_next:%[0-9]+]] = OpIAdd %uint [[c_msb_index_0]] %uint_1 // // The third member of S starts at byte offset 32 (8 words) @@ -728,19 +732,19 @@ void main(uint3 tid : SV_DispatchThreadId) { // // The fourth member of S starts at byte offset 48 (12 words) // -// CHECK: [[b_addr_0:%[0-9]+]] = OpIAdd %uint [[s1_addr]] %uint_48 -// -// CHECK: [[b_0:%[0-9]+]] = OpCompositeExtract %double [[s1]] 3 -// CHECK: [[b_index_0:%[0-9]+]] = OpShiftRightLogical %uint [[b_addr_0]] %uint_2 -// CHECK: [[ptr_33:%[0-9]+]] = OpAccessChain %_ptr_Uniform_uint %buf2 %uint_0 [[b_index_0]] -// CHECK: [[b_u64_0:%[0-9]+]] = OpBitcast %ulong [[b_0]] -// CHECK: [[b_word0_0:%[0-9]+]] = OpUConvert %uint [[b_u64_0]] -// CHECK: [[b_u64_shifted_0:%[0-9]+]] = OpShiftRightLogical %ulong [[b_u64_0]] %uint_32 -// CHECK: [[b_word1_0:%[0-9]+]] = OpUConvert %uint [[b_u64_shifted_0]] -// CHECK: OpStore [[ptr_33]] [[b_word0_0]] -// CHECK: [[b_msb_index_0:%[0-9]+]] = OpIAdd %uint [[b_index_0]] %uint_1 -// CHECK: [[ptr_34:%[0-9]+]] = OpAccessChain %_ptr_Uniform_uint %buf2 %uint_0 [[b_msb_index_0]] -// CHECK: OpStore [[ptr_34]] [[b_word1_0]] +// CHECK: [[b_addr_0:%[0-9]+]] = OpIAdd %uint [[s1_addr]] %uint_48 +// +// CHECK: [[b_0:%[0-9]+]] = OpCompositeExtract %double [[s1]] 3 +// CHECK: [[merge:%[0-9]+]] = OpBitcast %v2uint [[b_0]] +// CHECK: [[b_word0_0:%[0-9]+]] = OpCompositeExtract %uint [[merge]] 0 +// CHECK: [[b_word1_0:%[0-9]+]] = OpCompositeExtract %uint [[merge]] 1 +// CHECK: [[b_index_0:%[0-9]+]] = OpShiftRightLogical %uint [[b_addr_0]] %uint_2 +// CHECK: [[ptr_33:%[0-9]+]] = OpAccessChain %_ptr_Uniform_uint %buf2 %uint_0 [[b_index_0]] +// CHECK: OpStore [[ptr_33]] [[b_word0_0]] +// CHECK: [[b_msb_index_0:%[0-9]+]] = OpIAdd %uint [[b_index_0]] %uint_1 +// CHECK: [[ptr_34:%[0-9]+]] = OpAccessChain %_ptr_Uniform_uint %buf2 %uint_0 [[b_msb_index_0]] +// CHECK: OpStore [[ptr_34]] [[b_word1_0]] +// CHECK: [[next_index:%[0-9]+]] = OpIAdd %uint [[b_msb_index_0]] %uint_1 // // The fifth member of S starts at byte offset 56 (14 words) From e3f0efceb3d603fff16ec5dd1a5af15fa4861c23 Mon Sep 17 00:00:00 2001 From: Steven Perron Date: Wed, 22 Jan 2025 10:26:02 -0500 Subject: [PATCH 05/20] [SPIRV] Move inline spirv test files in a directory (#7077) Move the inline spir-v tests into their own directory. This allows people to find examples that use inline spir-v more easily. --- .../CodeGenSPIRV/{ => inline-spirv}/ext_builtin_input.lib.hlsl | 0 .../{ => inline-spirv}/spv.inline.builtin.both.error.hlsl | 0 .../CodeGenSPIRV/{ => inline-spirv}/spv.inline.builtin.enum.hlsl | 0 .../{ => inline-spirv}/spv.inline.builtin.input.flat.hlsl | 0 .../CodeGenSPIRV/{ => inline-spirv}/spv.inline.builtin.input.hlsl | 0 .../{ => inline-spirv}/spv.inline.builtin.input.nonconst.hlsl | 0 .../spv.inline.builtin.input.nonstatic.error.hlsl | 0 .../{ => inline-spirv}/spv.inline.builtin.output.hlsl | 0 .../spv.inline.builtin.output.nonstatic.error.hlsl | 0 .../{ => inline-spirv}/spv.inline.builtin.redefine.error.hlsl | 0 .../CodeGenSPIRV/{ => inline-spirv}/spv.inline.capability.hlsl | 0 .../CodeGenSPIRV/{ => inline-spirv}/spv.inline.executionmode.hlsl | 0 .../{ => inline-spirv}/spv.inline.executionmode.undefined.hlsl | 0 .../CodeGenSPIRV/{ => inline-spirv}/spv.inline.extension.hlsl | 0 .../{ => inline-spirv}/spv.inline.extension.unused.hlsl | 0 .../{ => inline-spirv}/spv.inline.type.alignment.hlsl | 0 .../{ => inline-spirv}/spv.inline.type.enum-class.hlsl | 0 .../test/CodeGenSPIRV/{ => inline-spirv}/spv.inline.type.hlsl | 0 .../{ => inline-spirv}/spv.inline.type.literal.error.hlsl | 0 .../{ => inline-spirv}/spv.intrinsic.reference.error.hlsl | 0 .../CodeGenSPIRV/{ => inline-spirv}/spv.intrinsic.result_id.hlsl | 0 .../CodeGenSPIRV/{ => inline-spirv}/spv.intrinsicDecorate.hlsl | 0 .../{ => inline-spirv}/spv.intrinsicExecutionMode.hlsl | 0 .../{ => inline-spirv}/spv.intrinsicExecutionMode.template.hlsl | 0 .../{ => inline-spirv}/spv.intrinsicExecutionModeId.hlsl | 0 .../CodeGenSPIRV/{ => inline-spirv}/spv.intrinsicInstruction.hlsl | 0 .../CodeGenSPIRV/{ => inline-spirv}/spv.intrinsicLiteral.hlsl | 0 .../{ => inline-spirv}/spv.intrinsicLiteralVariable.error.hlsl | 0 .../{ => inline-spirv}/spv.intrinsicLiteralVariable.hlsl | 0 .../{ => inline-spirv}/spv.intrinsicStorageClass.hlsl | 0 .../CodeGenSPIRV/{ => inline-spirv}/spv.intrinsicTypeInteger.hlsl | 0 .../{ => inline-spirv}/workgroupspirvpointer.const.hlsl | 0 .../CodeGenSPIRV/{ => inline-spirv}/workgroupspirvpointer.hlsl | 0 .../{ => inline-spirv}/workgroupspirvpointer.varpointer.hlsl | 0 34 files changed, 0 insertions(+), 0 deletions(-) rename tools/clang/test/CodeGenSPIRV/{ => inline-spirv}/ext_builtin_input.lib.hlsl (100%) rename tools/clang/test/CodeGenSPIRV/{ => inline-spirv}/spv.inline.builtin.both.error.hlsl (100%) rename tools/clang/test/CodeGenSPIRV/{ => inline-spirv}/spv.inline.builtin.enum.hlsl (100%) rename tools/clang/test/CodeGenSPIRV/{ => inline-spirv}/spv.inline.builtin.input.flat.hlsl (100%) rename tools/clang/test/CodeGenSPIRV/{ => inline-spirv}/spv.inline.builtin.input.hlsl (100%) rename tools/clang/test/CodeGenSPIRV/{ => inline-spirv}/spv.inline.builtin.input.nonconst.hlsl (100%) rename tools/clang/test/CodeGenSPIRV/{ => inline-spirv}/spv.inline.builtin.input.nonstatic.error.hlsl (100%) rename tools/clang/test/CodeGenSPIRV/{ => inline-spirv}/spv.inline.builtin.output.hlsl (100%) rename tools/clang/test/CodeGenSPIRV/{ => inline-spirv}/spv.inline.builtin.output.nonstatic.error.hlsl (100%) rename tools/clang/test/CodeGenSPIRV/{ => inline-spirv}/spv.inline.builtin.redefine.error.hlsl (100%) rename tools/clang/test/CodeGenSPIRV/{ => inline-spirv}/spv.inline.capability.hlsl (100%) rename tools/clang/test/CodeGenSPIRV/{ => inline-spirv}/spv.inline.executionmode.hlsl (100%) rename tools/clang/test/CodeGenSPIRV/{ => inline-spirv}/spv.inline.executionmode.undefined.hlsl (100%) rename tools/clang/test/CodeGenSPIRV/{ => inline-spirv}/spv.inline.extension.hlsl (100%) rename tools/clang/test/CodeGenSPIRV/{ => inline-spirv}/spv.inline.extension.unused.hlsl (100%) rename tools/clang/test/CodeGenSPIRV/{ => inline-spirv}/spv.inline.type.alignment.hlsl (100%) rename tools/clang/test/CodeGenSPIRV/{ => inline-spirv}/spv.inline.type.enum-class.hlsl (100%) rename tools/clang/test/CodeGenSPIRV/{ => inline-spirv}/spv.inline.type.hlsl (100%) rename tools/clang/test/CodeGenSPIRV/{ => inline-spirv}/spv.inline.type.literal.error.hlsl (100%) rename tools/clang/test/CodeGenSPIRV/{ => inline-spirv}/spv.intrinsic.reference.error.hlsl (100%) rename tools/clang/test/CodeGenSPIRV/{ => inline-spirv}/spv.intrinsic.result_id.hlsl (100%) rename tools/clang/test/CodeGenSPIRV/{ => inline-spirv}/spv.intrinsicDecorate.hlsl (100%) rename tools/clang/test/CodeGenSPIRV/{ => inline-spirv}/spv.intrinsicExecutionMode.hlsl (100%) rename tools/clang/test/CodeGenSPIRV/{ => inline-spirv}/spv.intrinsicExecutionMode.template.hlsl (100%) rename tools/clang/test/CodeGenSPIRV/{ => inline-spirv}/spv.intrinsicExecutionModeId.hlsl (100%) rename tools/clang/test/CodeGenSPIRV/{ => inline-spirv}/spv.intrinsicInstruction.hlsl (100%) rename tools/clang/test/CodeGenSPIRV/{ => inline-spirv}/spv.intrinsicLiteral.hlsl (100%) rename tools/clang/test/CodeGenSPIRV/{ => inline-spirv}/spv.intrinsicLiteralVariable.error.hlsl (100%) rename tools/clang/test/CodeGenSPIRV/{ => inline-spirv}/spv.intrinsicLiteralVariable.hlsl (100%) rename tools/clang/test/CodeGenSPIRV/{ => inline-spirv}/spv.intrinsicStorageClass.hlsl (100%) rename tools/clang/test/CodeGenSPIRV/{ => inline-spirv}/spv.intrinsicTypeInteger.hlsl (100%) rename tools/clang/test/CodeGenSPIRV/{ => inline-spirv}/workgroupspirvpointer.const.hlsl (100%) rename tools/clang/test/CodeGenSPIRV/{ => inline-spirv}/workgroupspirvpointer.hlsl (100%) rename tools/clang/test/CodeGenSPIRV/{ => inline-spirv}/workgroupspirvpointer.varpointer.hlsl (100%) diff --git a/tools/clang/test/CodeGenSPIRV/ext_builtin_input.lib.hlsl b/tools/clang/test/CodeGenSPIRV/inline-spirv/ext_builtin_input.lib.hlsl similarity index 100% rename from tools/clang/test/CodeGenSPIRV/ext_builtin_input.lib.hlsl rename to tools/clang/test/CodeGenSPIRV/inline-spirv/ext_builtin_input.lib.hlsl diff --git a/tools/clang/test/CodeGenSPIRV/spv.inline.builtin.both.error.hlsl b/tools/clang/test/CodeGenSPIRV/inline-spirv/spv.inline.builtin.both.error.hlsl similarity index 100% rename from tools/clang/test/CodeGenSPIRV/spv.inline.builtin.both.error.hlsl rename to tools/clang/test/CodeGenSPIRV/inline-spirv/spv.inline.builtin.both.error.hlsl diff --git a/tools/clang/test/CodeGenSPIRV/spv.inline.builtin.enum.hlsl b/tools/clang/test/CodeGenSPIRV/inline-spirv/spv.inline.builtin.enum.hlsl similarity index 100% rename from tools/clang/test/CodeGenSPIRV/spv.inline.builtin.enum.hlsl rename to tools/clang/test/CodeGenSPIRV/inline-spirv/spv.inline.builtin.enum.hlsl diff --git a/tools/clang/test/CodeGenSPIRV/spv.inline.builtin.input.flat.hlsl b/tools/clang/test/CodeGenSPIRV/inline-spirv/spv.inline.builtin.input.flat.hlsl similarity index 100% rename from tools/clang/test/CodeGenSPIRV/spv.inline.builtin.input.flat.hlsl rename to tools/clang/test/CodeGenSPIRV/inline-spirv/spv.inline.builtin.input.flat.hlsl diff --git a/tools/clang/test/CodeGenSPIRV/spv.inline.builtin.input.hlsl b/tools/clang/test/CodeGenSPIRV/inline-spirv/spv.inline.builtin.input.hlsl similarity index 100% rename from tools/clang/test/CodeGenSPIRV/spv.inline.builtin.input.hlsl rename to tools/clang/test/CodeGenSPIRV/inline-spirv/spv.inline.builtin.input.hlsl diff --git a/tools/clang/test/CodeGenSPIRV/spv.inline.builtin.input.nonconst.hlsl b/tools/clang/test/CodeGenSPIRV/inline-spirv/spv.inline.builtin.input.nonconst.hlsl similarity index 100% rename from tools/clang/test/CodeGenSPIRV/spv.inline.builtin.input.nonconst.hlsl rename to tools/clang/test/CodeGenSPIRV/inline-spirv/spv.inline.builtin.input.nonconst.hlsl diff --git a/tools/clang/test/CodeGenSPIRV/spv.inline.builtin.input.nonstatic.error.hlsl b/tools/clang/test/CodeGenSPIRV/inline-spirv/spv.inline.builtin.input.nonstatic.error.hlsl similarity index 100% rename from tools/clang/test/CodeGenSPIRV/spv.inline.builtin.input.nonstatic.error.hlsl rename to tools/clang/test/CodeGenSPIRV/inline-spirv/spv.inline.builtin.input.nonstatic.error.hlsl diff --git a/tools/clang/test/CodeGenSPIRV/spv.inline.builtin.output.hlsl b/tools/clang/test/CodeGenSPIRV/inline-spirv/spv.inline.builtin.output.hlsl similarity index 100% rename from tools/clang/test/CodeGenSPIRV/spv.inline.builtin.output.hlsl rename to tools/clang/test/CodeGenSPIRV/inline-spirv/spv.inline.builtin.output.hlsl diff --git a/tools/clang/test/CodeGenSPIRV/spv.inline.builtin.output.nonstatic.error.hlsl b/tools/clang/test/CodeGenSPIRV/inline-spirv/spv.inline.builtin.output.nonstatic.error.hlsl similarity index 100% rename from tools/clang/test/CodeGenSPIRV/spv.inline.builtin.output.nonstatic.error.hlsl rename to tools/clang/test/CodeGenSPIRV/inline-spirv/spv.inline.builtin.output.nonstatic.error.hlsl diff --git a/tools/clang/test/CodeGenSPIRV/spv.inline.builtin.redefine.error.hlsl b/tools/clang/test/CodeGenSPIRV/inline-spirv/spv.inline.builtin.redefine.error.hlsl similarity index 100% rename from tools/clang/test/CodeGenSPIRV/spv.inline.builtin.redefine.error.hlsl rename to tools/clang/test/CodeGenSPIRV/inline-spirv/spv.inline.builtin.redefine.error.hlsl diff --git a/tools/clang/test/CodeGenSPIRV/spv.inline.capability.hlsl b/tools/clang/test/CodeGenSPIRV/inline-spirv/spv.inline.capability.hlsl similarity index 100% rename from tools/clang/test/CodeGenSPIRV/spv.inline.capability.hlsl rename to tools/clang/test/CodeGenSPIRV/inline-spirv/spv.inline.capability.hlsl diff --git a/tools/clang/test/CodeGenSPIRV/spv.inline.executionmode.hlsl b/tools/clang/test/CodeGenSPIRV/inline-spirv/spv.inline.executionmode.hlsl similarity index 100% rename from tools/clang/test/CodeGenSPIRV/spv.inline.executionmode.hlsl rename to tools/clang/test/CodeGenSPIRV/inline-spirv/spv.inline.executionmode.hlsl diff --git a/tools/clang/test/CodeGenSPIRV/spv.inline.executionmode.undefined.hlsl b/tools/clang/test/CodeGenSPIRV/inline-spirv/spv.inline.executionmode.undefined.hlsl similarity index 100% rename from tools/clang/test/CodeGenSPIRV/spv.inline.executionmode.undefined.hlsl rename to tools/clang/test/CodeGenSPIRV/inline-spirv/spv.inline.executionmode.undefined.hlsl diff --git a/tools/clang/test/CodeGenSPIRV/spv.inline.extension.hlsl b/tools/clang/test/CodeGenSPIRV/inline-spirv/spv.inline.extension.hlsl similarity index 100% rename from tools/clang/test/CodeGenSPIRV/spv.inline.extension.hlsl rename to tools/clang/test/CodeGenSPIRV/inline-spirv/spv.inline.extension.hlsl diff --git a/tools/clang/test/CodeGenSPIRV/spv.inline.extension.unused.hlsl b/tools/clang/test/CodeGenSPIRV/inline-spirv/spv.inline.extension.unused.hlsl similarity index 100% rename from tools/clang/test/CodeGenSPIRV/spv.inline.extension.unused.hlsl rename to tools/clang/test/CodeGenSPIRV/inline-spirv/spv.inline.extension.unused.hlsl diff --git a/tools/clang/test/CodeGenSPIRV/spv.inline.type.alignment.hlsl b/tools/clang/test/CodeGenSPIRV/inline-spirv/spv.inline.type.alignment.hlsl similarity index 100% rename from tools/clang/test/CodeGenSPIRV/spv.inline.type.alignment.hlsl rename to tools/clang/test/CodeGenSPIRV/inline-spirv/spv.inline.type.alignment.hlsl diff --git a/tools/clang/test/CodeGenSPIRV/spv.inline.type.enum-class.hlsl b/tools/clang/test/CodeGenSPIRV/inline-spirv/spv.inline.type.enum-class.hlsl similarity index 100% rename from tools/clang/test/CodeGenSPIRV/spv.inline.type.enum-class.hlsl rename to tools/clang/test/CodeGenSPIRV/inline-spirv/spv.inline.type.enum-class.hlsl diff --git a/tools/clang/test/CodeGenSPIRV/spv.inline.type.hlsl b/tools/clang/test/CodeGenSPIRV/inline-spirv/spv.inline.type.hlsl similarity index 100% rename from tools/clang/test/CodeGenSPIRV/spv.inline.type.hlsl rename to tools/clang/test/CodeGenSPIRV/inline-spirv/spv.inline.type.hlsl diff --git a/tools/clang/test/CodeGenSPIRV/spv.inline.type.literal.error.hlsl b/tools/clang/test/CodeGenSPIRV/inline-spirv/spv.inline.type.literal.error.hlsl similarity index 100% rename from tools/clang/test/CodeGenSPIRV/spv.inline.type.literal.error.hlsl rename to tools/clang/test/CodeGenSPIRV/inline-spirv/spv.inline.type.literal.error.hlsl diff --git a/tools/clang/test/CodeGenSPIRV/spv.intrinsic.reference.error.hlsl b/tools/clang/test/CodeGenSPIRV/inline-spirv/spv.intrinsic.reference.error.hlsl similarity index 100% rename from tools/clang/test/CodeGenSPIRV/spv.intrinsic.reference.error.hlsl rename to tools/clang/test/CodeGenSPIRV/inline-spirv/spv.intrinsic.reference.error.hlsl diff --git a/tools/clang/test/CodeGenSPIRV/spv.intrinsic.result_id.hlsl b/tools/clang/test/CodeGenSPIRV/inline-spirv/spv.intrinsic.result_id.hlsl similarity index 100% rename from tools/clang/test/CodeGenSPIRV/spv.intrinsic.result_id.hlsl rename to tools/clang/test/CodeGenSPIRV/inline-spirv/spv.intrinsic.result_id.hlsl diff --git a/tools/clang/test/CodeGenSPIRV/spv.intrinsicDecorate.hlsl b/tools/clang/test/CodeGenSPIRV/inline-spirv/spv.intrinsicDecorate.hlsl similarity index 100% rename from tools/clang/test/CodeGenSPIRV/spv.intrinsicDecorate.hlsl rename to tools/clang/test/CodeGenSPIRV/inline-spirv/spv.intrinsicDecorate.hlsl diff --git a/tools/clang/test/CodeGenSPIRV/spv.intrinsicExecutionMode.hlsl b/tools/clang/test/CodeGenSPIRV/inline-spirv/spv.intrinsicExecutionMode.hlsl similarity index 100% rename from tools/clang/test/CodeGenSPIRV/spv.intrinsicExecutionMode.hlsl rename to tools/clang/test/CodeGenSPIRV/inline-spirv/spv.intrinsicExecutionMode.hlsl diff --git a/tools/clang/test/CodeGenSPIRV/spv.intrinsicExecutionMode.template.hlsl b/tools/clang/test/CodeGenSPIRV/inline-spirv/spv.intrinsicExecutionMode.template.hlsl similarity index 100% rename from tools/clang/test/CodeGenSPIRV/spv.intrinsicExecutionMode.template.hlsl rename to tools/clang/test/CodeGenSPIRV/inline-spirv/spv.intrinsicExecutionMode.template.hlsl diff --git a/tools/clang/test/CodeGenSPIRV/spv.intrinsicExecutionModeId.hlsl b/tools/clang/test/CodeGenSPIRV/inline-spirv/spv.intrinsicExecutionModeId.hlsl similarity index 100% rename from tools/clang/test/CodeGenSPIRV/spv.intrinsicExecutionModeId.hlsl rename to tools/clang/test/CodeGenSPIRV/inline-spirv/spv.intrinsicExecutionModeId.hlsl diff --git a/tools/clang/test/CodeGenSPIRV/spv.intrinsicInstruction.hlsl b/tools/clang/test/CodeGenSPIRV/inline-spirv/spv.intrinsicInstruction.hlsl similarity index 100% rename from tools/clang/test/CodeGenSPIRV/spv.intrinsicInstruction.hlsl rename to tools/clang/test/CodeGenSPIRV/inline-spirv/spv.intrinsicInstruction.hlsl diff --git a/tools/clang/test/CodeGenSPIRV/spv.intrinsicLiteral.hlsl b/tools/clang/test/CodeGenSPIRV/inline-spirv/spv.intrinsicLiteral.hlsl similarity index 100% rename from tools/clang/test/CodeGenSPIRV/spv.intrinsicLiteral.hlsl rename to tools/clang/test/CodeGenSPIRV/inline-spirv/spv.intrinsicLiteral.hlsl diff --git a/tools/clang/test/CodeGenSPIRV/spv.intrinsicLiteralVariable.error.hlsl b/tools/clang/test/CodeGenSPIRV/inline-spirv/spv.intrinsicLiteralVariable.error.hlsl similarity index 100% rename from tools/clang/test/CodeGenSPIRV/spv.intrinsicLiteralVariable.error.hlsl rename to tools/clang/test/CodeGenSPIRV/inline-spirv/spv.intrinsicLiteralVariable.error.hlsl diff --git a/tools/clang/test/CodeGenSPIRV/spv.intrinsicLiteralVariable.hlsl b/tools/clang/test/CodeGenSPIRV/inline-spirv/spv.intrinsicLiteralVariable.hlsl similarity index 100% rename from tools/clang/test/CodeGenSPIRV/spv.intrinsicLiteralVariable.hlsl rename to tools/clang/test/CodeGenSPIRV/inline-spirv/spv.intrinsicLiteralVariable.hlsl diff --git a/tools/clang/test/CodeGenSPIRV/spv.intrinsicStorageClass.hlsl b/tools/clang/test/CodeGenSPIRV/inline-spirv/spv.intrinsicStorageClass.hlsl similarity index 100% rename from tools/clang/test/CodeGenSPIRV/spv.intrinsicStorageClass.hlsl rename to tools/clang/test/CodeGenSPIRV/inline-spirv/spv.intrinsicStorageClass.hlsl diff --git a/tools/clang/test/CodeGenSPIRV/spv.intrinsicTypeInteger.hlsl b/tools/clang/test/CodeGenSPIRV/inline-spirv/spv.intrinsicTypeInteger.hlsl similarity index 100% rename from tools/clang/test/CodeGenSPIRV/spv.intrinsicTypeInteger.hlsl rename to tools/clang/test/CodeGenSPIRV/inline-spirv/spv.intrinsicTypeInteger.hlsl diff --git a/tools/clang/test/CodeGenSPIRV/workgroupspirvpointer.const.hlsl b/tools/clang/test/CodeGenSPIRV/inline-spirv/workgroupspirvpointer.const.hlsl similarity index 100% rename from tools/clang/test/CodeGenSPIRV/workgroupspirvpointer.const.hlsl rename to tools/clang/test/CodeGenSPIRV/inline-spirv/workgroupspirvpointer.const.hlsl diff --git a/tools/clang/test/CodeGenSPIRV/workgroupspirvpointer.hlsl b/tools/clang/test/CodeGenSPIRV/inline-spirv/workgroupspirvpointer.hlsl similarity index 100% rename from tools/clang/test/CodeGenSPIRV/workgroupspirvpointer.hlsl rename to tools/clang/test/CodeGenSPIRV/inline-spirv/workgroupspirvpointer.hlsl diff --git a/tools/clang/test/CodeGenSPIRV/workgroupspirvpointer.varpointer.hlsl b/tools/clang/test/CodeGenSPIRV/inline-spirv/workgroupspirvpointer.varpointer.hlsl similarity index 100% rename from tools/clang/test/CodeGenSPIRV/workgroupspirvpointer.varpointer.hlsl rename to tools/clang/test/CodeGenSPIRV/inline-spirv/workgroupspirvpointer.varpointer.hlsl From a7b7a0c6d1412c8b0a19cb2410c1c8840fb34949 Mon Sep 17 00:00:00 2001 From: Joshua Batista Date: Wed, 22 Jan 2025 11:29:15 -0800 Subject: [PATCH 06/20] Bump SM version to 6.9 (#7075) Bumping SM version to 6.9. Though 6.9 is a while away from release, we intend to use 6.9 as a temporary marker for a pre-release version. Co-authored-by: Greg Roth --- include/dxc/DXIL/DxilConstants.h | 2 +- include/dxc/DXIL/DxilShaderModel.h | 5 +- include/dxc/Support/HLSLOptions.td | 2 +- lib/DXIL/DxilShaderModel.cpp | 189 +++++++++++++++++------------ utils/hct/hctdb_instrhelp.py | 2 +- 5 files changed, 116 insertions(+), 84 deletions(-) diff --git a/include/dxc/DXIL/DxilConstants.h b/include/dxc/DXIL/DxilConstants.h index 2449915754..aeb214f48d 100644 --- a/include/dxc/DXIL/DxilConstants.h +++ b/include/dxc/DXIL/DxilConstants.h @@ -29,7 +29,7 @@ namespace DXIL { const unsigned kDxilMajor = 1; /* hctdb_instrhelp.get_dxil_version_minor()*/ // VALRULE-TEXT:BEGIN -const unsigned kDxilMinor = 8; +const unsigned kDxilMinor = 9; // VALRULE-TEXT:END inline unsigned MakeDxilVersion(unsigned DxilMajor, unsigned DxilMinor) { diff --git a/include/dxc/DXIL/DxilShaderModel.h b/include/dxc/DXIL/DxilShaderModel.h index 293797ba12..c89c079bb9 100644 --- a/include/dxc/DXIL/DxilShaderModel.h +++ b/include/dxc/DXIL/DxilShaderModel.h @@ -33,7 +33,7 @@ class ShaderModel { // clang-format on // VALRULE-TEXT:BEGIN static const unsigned kHighestMajor = 6; - static const unsigned kHighestMinor = 8; + static const unsigned kHighestMinor = 9; // VALRULE-TEXT:END static const unsigned kOfflineMinor = 0xF; @@ -74,6 +74,7 @@ class ShaderModel { bool IsSM66Plus() const { return IsSMAtLeast(6, 6); } bool IsSM67Plus() const { return IsSMAtLeast(6, 7); } bool IsSM68Plus() const { return IsSMAtLeast(6, 8); } + bool IsSM69Plus() const { return IsSMAtLeast(6, 9); } // VALRULE-TEXT:END const char *GetName() const { return m_pszName; } const char *GetKindName() const; @@ -121,7 +122,7 @@ class ShaderModel { bool m_bTypedUavs, unsigned m_UAVRegsLim); /* hctdb_instrhelp.get_num_shader_models()*/ // VALRULE-TEXT:BEGIN - static const unsigned kNumShaderModels = 92; + static const unsigned kNumShaderModels = 101; // VALRULE-TEXT:END static const ShaderModel ms_ShaderModels[kNumShaderModels]; diff --git a/include/dxc/Support/HLSLOptions.td b/include/dxc/Support/HLSLOptions.td index b7bf232070..130e19a525 100644 --- a/include/dxc/Support/HLSLOptions.td +++ b/include/dxc/Support/HLSLOptions.td @@ -430,7 +430,7 @@ def fvk_bind_counter_heap : MultiArg<["-"], "fvk-bind-counter-heap", 2>, MetaVar def target_profile : JoinedOrSeparate<["-", "/"], "T">, Flags<[CoreOption]>, Group, MetaVarName<"">, /* hctdb_instrhelp.get_target_profiles()*/ // VALRULE-TEXT:BEGIN - HelpText<"Set target profile. \n\t: ps_6_0, ps_6_1, ps_6_2, ps_6_3, ps_6_4, ps_6_5, ps_6_6, ps_6_7, ps_6_8, \n\t\t vs_6_0, vs_6_1, vs_6_2, vs_6_3, vs_6_4, vs_6_5, vs_6_6, vs_6_7, vs_6_8, \n\t\t gs_6_0, gs_6_1, gs_6_2, gs_6_3, gs_6_4, gs_6_5, gs_6_6, gs_6_7, gs_6_8, \n\t\t hs_6_0, hs_6_1, hs_6_2, hs_6_3, hs_6_4, hs_6_5, hs_6_6, hs_6_7, hs_6_8, \n\t\t ds_6_0, ds_6_1, ds_6_2, ds_6_3, ds_6_4, ds_6_5, ds_6_6, ds_6_7, ds_6_8, \n\t\t cs_6_0, cs_6_1, cs_6_2, cs_6_3, cs_6_4, cs_6_5, cs_6_6, cs_6_7, cs_6_8, \n\t\t lib_6_1, lib_6_2, lib_6_3, lib_6_4, lib_6_5, lib_6_6, lib_6_7, lib_6_8, \n\t\t ms_6_5, ms_6_6, ms_6_7, ms_6_8, \n\t\t as_6_5, as_6_6, as_6_7, as_6_8, \n\t\t ">; + HelpText<"Set target profile. \n\t: ps_6_0, ps_6_1, ps_6_2, ps_6_3, ps_6_4, ps_6_5, ps_6_6, ps_6_7, ps_6_8, ps_6_9, \n\t\t vs_6_0, vs_6_1, vs_6_2, vs_6_3, vs_6_4, vs_6_5, vs_6_6, vs_6_7, vs_6_8, vs_6_9, \n\t\t gs_6_0, gs_6_1, gs_6_2, gs_6_3, gs_6_4, gs_6_5, gs_6_6, gs_6_7, gs_6_8, gs_6_9, \n\t\t hs_6_0, hs_6_1, hs_6_2, hs_6_3, hs_6_4, hs_6_5, hs_6_6, hs_6_7, hs_6_8, hs_6_9, \n\t\t ds_6_0, ds_6_1, ds_6_2, ds_6_3, ds_6_4, ds_6_5, ds_6_6, ds_6_7, ds_6_8, ds_6_9, \n\t\t cs_6_0, cs_6_1, cs_6_2, cs_6_3, cs_6_4, cs_6_5, cs_6_6, cs_6_7, cs_6_8, cs_6_9, \n\t\t lib_6_1, lib_6_2, lib_6_3, lib_6_4, lib_6_5, lib_6_6, lib_6_7, lib_6_8, lib_6_9, \n\t\t ms_6_5, ms_6_6, ms_6_7, ms_6_8, ms_6_9, \n\t\t as_6_5, as_6_6, as_6_7, as_6_8, as_6_9, \n\t\t ">; // VALRULE-TEXT:END def entrypoint : JoinedOrSeparate<["-", "/"], "E">, Flags<[CoreOption, RewriteOption]>, Group, HelpText<"Entry point name">; diff --git a/lib/DXIL/DxilShaderModel.cpp b/lib/DXIL/DxilShaderModel.cpp index 22bf5f545b..8c94b3f5a1 100644 --- a/lib/DXIL/DxilShaderModel.cpp +++ b/lib/DXIL/DxilShaderModel.cpp @@ -64,6 +64,7 @@ bool ShaderModel::IsValidForDxil() const { case 6: case 7: case 8: + case 9: // VALRULE-TEXT:END return true; case kOfflineMinor: @@ -91,85 +92,94 @@ const ShaderModel *ShaderModel::Get(Kind Kind, unsigned Major, unsigned Minor) { {1542, 10}, // ps_6_6 {1543, 11}, // ps_6_7 {1544, 12}, // ps_6_8 - {66560, 13}, // vs_4_0 - {66561, 14}, // vs_4_1 - {66816, 15}, // vs_5_0 - {66817, 16}, // vs_5_1 - {67072, 17}, // vs_6_0 - {67073, 18}, // vs_6_1 - {67074, 19}, // vs_6_2 - {67075, 20}, // vs_6_3 - {67076, 21}, // vs_6_4 - {67077, 22}, // vs_6_5 - {67078, 23}, // vs_6_6 - {67079, 24}, // vs_6_7 - {67080, 25}, // vs_6_8 - {132096, 26}, // gs_4_0 - {132097, 27}, // gs_4_1 - {132352, 28}, // gs_5_0 - {132353, 29}, // gs_5_1 - {132608, 30}, // gs_6_0 - {132609, 31}, // gs_6_1 - {132610, 32}, // gs_6_2 - {132611, 33}, // gs_6_3 - {132612, 34}, // gs_6_4 - {132613, 35}, // gs_6_5 - {132614, 36}, // gs_6_6 - {132615, 37}, // gs_6_7 - {132616, 38}, // gs_6_8 - {197888, 39}, // hs_5_0 - {197889, 40}, // hs_5_1 - {198144, 41}, // hs_6_0 - {198145, 42}, // hs_6_1 - {198146, 43}, // hs_6_2 - {198147, 44}, // hs_6_3 - {198148, 45}, // hs_6_4 - {198149, 46}, // hs_6_5 - {198150, 47}, // hs_6_6 - {198151, 48}, // hs_6_7 - {198152, 49}, // hs_6_8 - {263424, 50}, // ds_5_0 - {263425, 51}, // ds_5_1 - {263680, 52}, // ds_6_0 - {263681, 53}, // ds_6_1 - {263682, 54}, // ds_6_2 - {263683, 55}, // ds_6_3 - {263684, 56}, // ds_6_4 - {263685, 57}, // ds_6_5 - {263686, 58}, // ds_6_6 - {263687, 59}, // ds_6_7 - {263688, 60}, // ds_6_8 - {328704, 61}, // cs_4_0 - {328705, 62}, // cs_4_1 - {328960, 63}, // cs_5_0 - {328961, 64}, // cs_5_1 - {329216, 65}, // cs_6_0 - {329217, 66}, // cs_6_1 - {329218, 67}, // cs_6_2 - {329219, 68}, // cs_6_3 - {329220, 69}, // cs_6_4 - {329221, 70}, // cs_6_5 - {329222, 71}, // cs_6_6 - {329223, 72}, // cs_6_7 - {329224, 73}, // cs_6_8 - {394753, 74}, // lib_6_1 - {394754, 75}, // lib_6_2 - {394755, 76}, // lib_6_3 - {394756, 77}, // lib_6_4 - {394757, 78}, // lib_6_5 - {394758, 79}, // lib_6_6 - {394759, 80}, // lib_6_7 - {394760, 81}, // lib_6_8 + {1545, 13}, // ps_6_9 + {66560, 14}, // vs_4_0 + {66561, 15}, // vs_4_1 + {66816, 16}, // vs_5_0 + {66817, 17}, // vs_5_1 + {67072, 18}, // vs_6_0 + {67073, 19}, // vs_6_1 + {67074, 20}, // vs_6_2 + {67075, 21}, // vs_6_3 + {67076, 22}, // vs_6_4 + {67077, 23}, // vs_6_5 + {67078, 24}, // vs_6_6 + {67079, 25}, // vs_6_7 + {67080, 26}, // vs_6_8 + {67081, 27}, // vs_6_9 + {132096, 28}, // gs_4_0 + {132097, 29}, // gs_4_1 + {132352, 30}, // gs_5_0 + {132353, 31}, // gs_5_1 + {132608, 32}, // gs_6_0 + {132609, 33}, // gs_6_1 + {132610, 34}, // gs_6_2 + {132611, 35}, // gs_6_3 + {132612, 36}, // gs_6_4 + {132613, 37}, // gs_6_5 + {132614, 38}, // gs_6_6 + {132615, 39}, // gs_6_7 + {132616, 40}, // gs_6_8 + {132617, 41}, // gs_6_9 + {197888, 42}, // hs_5_0 + {197889, 43}, // hs_5_1 + {198144, 44}, // hs_6_0 + {198145, 45}, // hs_6_1 + {198146, 46}, // hs_6_2 + {198147, 47}, // hs_6_3 + {198148, 48}, // hs_6_4 + {198149, 49}, // hs_6_5 + {198150, 50}, // hs_6_6 + {198151, 51}, // hs_6_7 + {198152, 52}, // hs_6_8 + {198153, 53}, // hs_6_9 + {263424, 54}, // ds_5_0 + {263425, 55}, // ds_5_1 + {263680, 56}, // ds_6_0 + {263681, 57}, // ds_6_1 + {263682, 58}, // ds_6_2 + {263683, 59}, // ds_6_3 + {263684, 60}, // ds_6_4 + {263685, 61}, // ds_6_5 + {263686, 62}, // ds_6_6 + {263687, 63}, // ds_6_7 + {263688, 64}, // ds_6_8 + {263689, 65}, // ds_6_9 + {328704, 66}, // cs_4_0 + {328705, 67}, // cs_4_1 + {328960, 68}, // cs_5_0 + {328961, 69}, // cs_5_1 + {329216, 70}, // cs_6_0 + {329217, 71}, // cs_6_1 + {329218, 72}, // cs_6_2 + {329219, 73}, // cs_6_3 + {329220, 74}, // cs_6_4 + {329221, 75}, // cs_6_5 + {329222, 76}, // cs_6_6 + {329223, 77}, // cs_6_7 + {329224, 78}, // cs_6_8 + {329225, 79}, // cs_6_9 + {394753, 80}, // lib_6_1 + {394754, 81}, // lib_6_2 + {394755, 82}, // lib_6_3 + {394756, 83}, // lib_6_4 + {394757, 84}, // lib_6_5 + {394758, 85}, // lib_6_6 + {394759, 86}, // lib_6_7 + {394760, 87}, // lib_6_8 + {394761, 88}, // lib_6_9 // lib_6_x is for offline linking only, and relaxes restrictions - {394767, 82}, // lib_6_x - {853509, 83}, // ms_6_5 - {853510, 84}, // ms_6_6 - {853511, 85}, // ms_6_7 - {853512, 86}, // ms_6_8 - {919045, 87}, // as_6_5 - {919046, 88}, // as_6_6 - {919047, 89}, // as_6_7 - {919048, 90}, // as_6_8 + {394767, 89}, // lib_6_x + {853509, 90}, // ms_6_5 + {853510, 91}, // ms_6_6 + {853511, 92}, // ms_6_7 + {853512, 93}, // ms_6_8 + {853513, 94}, // ms_6_9 + {919045, 95}, // as_6_5 + {919046, 96}, // as_6_6 + {919047, 97}, // as_6_7 + {919048, 98}, // as_6_8 + {919049, 99}, // as_6_9 }; unsigned hash = (unsigned)Kind << 16 | Major << 8 | Minor; auto pred = [](const std::pair &elem, unsigned val) { @@ -303,6 +313,12 @@ const ShaderModel *ShaderModel::GetByName(llvm::StringRef Name) { break; } else return GetInvalid(); + case '9': + if (Major == 6) { + Minor = 9; + break; + } else + return GetInvalid(); // VALRULE-TEXT:END case 'x': if (kind == Kind::Library && Major == 6) { @@ -354,8 +370,11 @@ void ShaderModel::GetDxilVersion(unsigned &DxilMajor, case 8: DxilMinor = 8; break; + case 9: + DxilMinor = 9; + break; case kOfflineMinor: // Always update this to highest dxil version - DxilMinor = 8; + DxilMinor = 9; break; // VALRULE-TEXT:END default: @@ -401,6 +420,9 @@ void ShaderModel::GetMinValidatorVersion(unsigned &ValMajor, case 8: ValMinor = 8; break; + case 9: + ValMinor = 9; + break; // VALRULE-TEXT:END case kOfflineMinor: ValMajor = 0; @@ -541,6 +563,7 @@ const ShaderModel ShaderModel::ms_ShaderModels[kNumShaderModels] = { SM(Kind::Pixel, 6, 6, "ps_6_6", 32, 8, true, true, UINT_MAX), SM(Kind::Pixel, 6, 7, "ps_6_7", 32, 8, true, true, UINT_MAX), SM(Kind::Pixel, 6, 8, "ps_6_8", 32, 8, true, true, UINT_MAX), + SM(Kind::Pixel, 6, 9, "ps_6_9", 32, 8, true, true, UINT_MAX), SM(Kind::Vertex, 4, 0, "vs_4_0", 16, 16, false, false, 0), SM(Kind::Vertex, 4, 1, "vs_4_1", 32, 32, false, false, 0), SM(Kind::Vertex, 5, 0, "vs_5_0", 32, 32, true, true, 64), @@ -554,6 +577,7 @@ const ShaderModel ShaderModel::ms_ShaderModels[kNumShaderModels] = { SM(Kind::Vertex, 6, 6, "vs_6_6", 32, 32, true, true, UINT_MAX), SM(Kind::Vertex, 6, 7, "vs_6_7", 32, 32, true, true, UINT_MAX), SM(Kind::Vertex, 6, 8, "vs_6_8", 32, 32, true, true, UINT_MAX), + SM(Kind::Vertex, 6, 9, "vs_6_9", 32, 32, true, true, UINT_MAX), SM(Kind::Geometry, 4, 0, "gs_4_0", 16, 32, false, false, 0), SM(Kind::Geometry, 4, 1, "gs_4_1", 32, 32, false, false, 0), SM(Kind::Geometry, 5, 0, "gs_5_0", 32, 32, true, true, 64), @@ -567,6 +591,7 @@ const ShaderModel ShaderModel::ms_ShaderModels[kNumShaderModels] = { SM(Kind::Geometry, 6, 6, "gs_6_6", 32, 32, true, true, UINT_MAX), SM(Kind::Geometry, 6, 7, "gs_6_7", 32, 32, true, true, UINT_MAX), SM(Kind::Geometry, 6, 8, "gs_6_8", 32, 32, true, true, UINT_MAX), + SM(Kind::Geometry, 6, 9, "gs_6_9", 32, 32, true, true, UINT_MAX), SM(Kind::Hull, 5, 0, "hs_5_0", 32, 32, true, true, 64), SM(Kind::Hull, 5, 1, "hs_5_1", 32, 32, true, true, 64), SM(Kind::Hull, 6, 0, "hs_6_0", 32, 32, true, true, UINT_MAX), @@ -578,6 +603,7 @@ const ShaderModel ShaderModel::ms_ShaderModels[kNumShaderModels] = { SM(Kind::Hull, 6, 6, "hs_6_6", 32, 32, true, true, UINT_MAX), SM(Kind::Hull, 6, 7, "hs_6_7", 32, 32, true, true, UINT_MAX), SM(Kind::Hull, 6, 8, "hs_6_8", 32, 32, true, true, UINT_MAX), + SM(Kind::Hull, 6, 9, "hs_6_9", 32, 32, true, true, UINT_MAX), SM(Kind::Domain, 5, 0, "ds_5_0", 32, 32, true, true, 64), SM(Kind::Domain, 5, 1, "ds_5_1", 32, 32, true, true, 64), SM(Kind::Domain, 6, 0, "ds_6_0", 32, 32, true, true, UINT_MAX), @@ -589,6 +615,7 @@ const ShaderModel ShaderModel::ms_ShaderModels[kNumShaderModels] = { SM(Kind::Domain, 6, 6, "ds_6_6", 32, 32, true, true, UINT_MAX), SM(Kind::Domain, 6, 7, "ds_6_7", 32, 32, true, true, UINT_MAX), SM(Kind::Domain, 6, 8, "ds_6_8", 32, 32, true, true, UINT_MAX), + SM(Kind::Domain, 6, 9, "ds_6_9", 32, 32, true, true, UINT_MAX), SM(Kind::Compute, 4, 0, "cs_4_0", 0, 0, false, false, 0), SM(Kind::Compute, 4, 1, "cs_4_1", 0, 0, false, false, 0), SM(Kind::Compute, 5, 0, "cs_5_0", 0, 0, true, true, 64), @@ -602,6 +629,7 @@ const ShaderModel ShaderModel::ms_ShaderModels[kNumShaderModels] = { SM(Kind::Compute, 6, 6, "cs_6_6", 0, 0, true, true, UINT_MAX), SM(Kind::Compute, 6, 7, "cs_6_7", 0, 0, true, true, UINT_MAX), SM(Kind::Compute, 6, 8, "cs_6_8", 0, 0, true, true, UINT_MAX), + SM(Kind::Compute, 6, 9, "cs_6_9", 0, 0, true, true, UINT_MAX), SM(Kind::Library, 6, 1, "lib_6_1", 32, 32, true, true, UINT_MAX), SM(Kind::Library, 6, 2, "lib_6_2", 32, 32, true, true, UINT_MAX), SM(Kind::Library, 6, 3, "lib_6_3", 32, 32, true, true, UINT_MAX), @@ -610,6 +638,7 @@ const ShaderModel ShaderModel::ms_ShaderModels[kNumShaderModels] = { SM(Kind::Library, 6, 6, "lib_6_6", 32, 32, true, true, UINT_MAX), SM(Kind::Library, 6, 7, "lib_6_7", 32, 32, true, true, UINT_MAX), SM(Kind::Library, 6, 8, "lib_6_8", 32, 32, true, true, UINT_MAX), + SM(Kind::Library, 6, 9, "lib_6_9", 32, 32, true, true, UINT_MAX), // lib_6_x is for offline linking only, and relaxes restrictions SM(Kind::Library, 6, kOfflineMinor, "lib_6_x", 32, 32, true, true, UINT_MAX), @@ -617,10 +646,12 @@ const ShaderModel ShaderModel::ms_ShaderModels[kNumShaderModels] = { SM(Kind::Mesh, 6, 6, "ms_6_6", 0, 0, true, true, UINT_MAX), SM(Kind::Mesh, 6, 7, "ms_6_7", 0, 0, true, true, UINT_MAX), SM(Kind::Mesh, 6, 8, "ms_6_8", 0, 0, true, true, UINT_MAX), + SM(Kind::Mesh, 6, 9, "ms_6_9", 0, 0, true, true, UINT_MAX), SM(Kind::Amplification, 6, 5, "as_6_5", 0, 0, true, true, UINT_MAX), SM(Kind::Amplification, 6, 6, "as_6_6", 0, 0, true, true, UINT_MAX), SM(Kind::Amplification, 6, 7, "as_6_7", 0, 0, true, true, UINT_MAX), SM(Kind::Amplification, 6, 8, "as_6_8", 0, 0, true, true, UINT_MAX), + SM(Kind::Amplification, 6, 9, "as_6_9", 0, 0, true, true, UINT_MAX), // Values before Invalid must remain sorted by Kind, then Major, then Minor. SM(Kind::Invalid, 0, 0, "invalid", 0, 0, false, false, 0), // VALRULE-TEXT:END diff --git a/utils/hct/hctdb_instrhelp.py b/utils/hct/hctdb_instrhelp.py index 919610e420..7b0de52db4 100644 --- a/utils/hct/hctdb_instrhelp.py +++ b/utils/hct/hctdb_instrhelp.py @@ -1537,7 +1537,7 @@ def get_interpretation_table(): highest_major = 6 -highest_minor = 8 +highest_minor = 9 highest_shader_models = {4: 1, 5: 1, 6: highest_minor} From b14832986cf4bd5d39a21123f9fed29c88474054 Mon Sep 17 00:00:00 2001 From: Joshua Batista Date: Wed, 22 Jan 2025 12:55:23 -0800 Subject: [PATCH 07/20] [Release] Add section for upcoming release (#7072) Add section for upcoming release --- docs/ReleaseNotes.md | 4 ++++ 1 file changed, 4 insertions(+) diff --git a/docs/ReleaseNotes.md b/docs/ReleaseNotes.md index b47f17b600..b88792d302 100644 --- a/docs/ReleaseNotes.md +++ b/docs/ReleaseNotes.md @@ -17,6 +17,10 @@ The included licenses apply to the following files: ## Changelog +### Upcoming Release + +Place release notes for the upcoming release below this line and remove this line upon naming this release. + ### Version 1.8.2502 - The incomplete WaveMatrix implementation has been removed. From 25faa88053437c30e599b9e5323c94addc2397a1 Mon Sep 17 00:00:00 2001 From: Greg Roth Date: Wed, 22 Jan 2025 14:23:33 -0700 Subject: [PATCH 08/20] Fix and/or intrinsics with non-int parameters (#7060) And/or intrinsics were set to allow any parameters, which is consistent with the behavior of the && and || operators they were meant to replace, however this meant that if they were passed floating point values, those values would be applied to the binary and/or operands, which isn't allowed. Instead, they should be converted to booleans to be consistent with the behavior of && and ||. This can be done simply by restricting the parameters to booleans which forces the appropriate conversions. Adds tests for for using bools, ints, and floats in scalars, different sized vectors, and matrices as parameters to or and and. Fixes: #7057 FIxes: #6995 --- tools/clang/test/CodeGenHLSL/vector-and.hlsl | 141 +++++++++++++++ tools/clang/test/CodeGenHLSL/vector-or.hlsl | 164 ++++++++++++++++++ .../test/CodeGenSPIRV/intrinsics.and.hlsl | 45 +++++ .../test/CodeGenSPIRV/intrinsics.or.hlsl | 34 ++++ utils/hct/gen_intrin_main.txt | 4 +- 5 files changed, 386 insertions(+), 2 deletions(-) create mode 100644 tools/clang/test/CodeGenHLSL/vector-and.hlsl create mode 100644 tools/clang/test/CodeGenHLSL/vector-or.hlsl diff --git a/tools/clang/test/CodeGenHLSL/vector-and.hlsl b/tools/clang/test/CodeGenHLSL/vector-and.hlsl new file mode 100644 index 0000000000..d6ae00af2f --- /dev/null +++ b/tools/clang/test/CodeGenHLSL/vector-and.hlsl @@ -0,0 +1,141 @@ +// RUN: %dxc -T ps_6_0 -HV 2021 -DTYPE=bool %s | FileCheck %s --check-prefixes=CHECK,I32 +// RUN: %dxc -T ps_6_0 -HV 2018 -DTYPE=bool %s | FileCheck %s --check-prefixes=CHECK,I32 +// RUN: %dxc -T ps_6_0 -HV 2021 -DTYPE=int %s | FileCheck %s --check-prefixes=CHECK,I32 +// RUN: %dxc -T ps_6_0 -HV 2018 -DTYPE=int %s | FileCheck %s --check-prefixes=CHECK,I32 +// RUN: %dxc -T ps_6_0 -HV 2021 -DTYPE=float %s | FileCheck %s --check-prefixes=CHECK,F32 +// RUN: %dxc -T ps_6_0 -HV 2018 -DTYPE=float %s | FileCheck %s --check-prefixes=CHECK,F32 + +// I32: %dx.types.ResRet.[[TY:i32]] = type { [[TYPE:i32]] +// F32: %dx.types.ResRet.[[TY:f32]] = type { [[TYPE:float]] + +// CHECK-LABEL: define void @main + +ByteAddressBuffer buf; + +float4 main() : SV_Target { + + // CHECK-DAG: [[SAR:%.*]] = call %dx.types.ResRet.[[TY]] @dx.op.bufferLoad.[[TY]](i32 68, %dx.types.Handle %{{.*}}, i32 0 + // CHECK-DAG: [[SAX:%.*]] = extractvalue %dx.types.ResRet.[[TY]] [[SAR]], 0 + // I32-DAG: [[SA:%.*]] = icmp ne i32 [[SAX]], 0 + // F32-DAG: [[SA:%.*]] = fcmp fast une float [[SAX]], 0.000000e+00 + + // CHECK-DAG: [[SBR:%.*]] = call %dx.types.ResRet.[[TY]] @dx.op.bufferLoad.[[TY]](i32 68, %dx.types.Handle %{{.*}}, i32 8 + // CHECK-DAG: [[SBX:%.*]] = extractvalue %dx.types.ResRet.[[TY]] [[SBR]], 0 + // I32-DAG: [[SB:%.*]] = icmp ne i32 [[SBX]], 0 + // F32-DAG: [[SB:%.*]] = fcmp fast une float [[SBX]], 0.000000e+00 + + TYPE sb = buf.Load(8); + TYPE sa = buf.Load(0); + + // CHECK: and i1 [[SB]], [[SA]] + TYPE res = and(sa, sb); + + // CHECK-DAG: [[V1AR:%.*]] = call %dx.types.ResRet.[[TY]] @dx.op.bufferLoad.[[TY]](i32 68, %dx.types.Handle %{{.*}}, i32 16 + // CHECK-DAG: [[V1AX:%.*]] = extractvalue %dx.types.ResRet.[[TY]] [[V1AR]], 0 + // I32-DAG: [[V1A:%.*]] = icmp ne i32 [[V1AX]], 0 + // F32-DAG: [[V1A:%.*]] = fcmp fast une float [[V1AX]], 0.000000e+00 + + // CHECK-DAG: [[V1BR:%.*]] = call %dx.types.ResRet.[[TY]] @dx.op.bufferLoad.[[TY]](i32 68, %dx.types.Handle %{{.*}}, i32 24 + // CHECK-DAG: [[V1BX:%.*]] = extractvalue %dx.types.ResRet.[[TY]] [[V1BR]], 0 + // I32-DAG: [[V1B:%.*]] = icmp ne i32 [[V1BX]], 0 + // F32-DAG: [[V1B:%.*]] = fcmp fast une float [[V1BX]], 0.000000e+00 + + vector v1b = buf.Load< vector >(24); + vector v1a = buf.Load< vector >(16); + + // CHECK: and i1 [[V1B]], [[V1A]] + vector res1 = and(v1a, v1b); + + // CHECK-DAG: [[V3AR:%.*]] = call %dx.types.ResRet.[[TY]] @dx.op.bufferLoad.[[TY]](i32 68, %dx.types.Handle %{{.*}}, i32 32 + // CHECK-DAG: [[V3AX0:%.*]] = extractvalue %dx.types.ResRet.[[TY]] [[V3AR]], 0 + // CHECK-DAG: [[V3AX1:%.*]] = extractvalue %dx.types.ResRet.[[TY]] [[V3AR]], 1 + // CHECK-DAG: [[V3AX2:%.*]] = extractvalue %dx.types.ResRet.[[TY]] [[V3AR]], 2 + + // I32-DAG: [[V3A0:%.*]] = icmp ne i32 [[V3AX0]], 0 + // I32-DAG: [[V3A1:%.*]] = icmp ne i32 [[V3AX1]], 0 + // I32-DAG: [[V3A2:%.*]] = icmp ne i32 [[V3AX2]], 0 + + // F32-DAG: [[V3A0:%.*]] = fcmp fast une float [[V3AX0]], 0.000000e+00 + // F32-DAG: [[V3A1:%.*]] = fcmp fast une float [[V3AX1]], 0.000000e+00 + // F32-DAG: [[V3A2:%.*]] = fcmp fast une float [[V3AX2]], 0.000000e+00 + + // CHECK-DAG: [[V3BR:%.*]] = call %dx.types.ResRet.[[TY]] @dx.op.bufferLoad.[[TY]](i32 68, %dx.types.Handle %{{.*}}, i32 56 + // CHECK-DAG: [[V3BX0:%.*]] = extractvalue %dx.types.ResRet.[[TY]] [[V3BR]], 0 + // CHECK-DAG: [[V3BX1:%.*]] = extractvalue %dx.types.ResRet.[[TY]] [[V3BR]], 1 + // CHECK-DAG: [[V3BX2:%.*]] = extractvalue %dx.types.ResRet.[[TY]] [[V3BR]], 2 + + // I32-DAG: [[V3B0:%.*]] = icmp ne i32 [[V3BX0]], 0 + // I32-DAG: [[V3B1:%.*]] = icmp ne i32 [[V3BX1]], 0 + // I32-DAG: [[V3B2:%.*]] = icmp ne i32 [[V3BX2]], 0 + + // F32-DAG: [[V3B0:%.*]] = fcmp fast une float [[V3BX0]], 0.000000e+00 + // F32-DAG: [[V3B1:%.*]] = fcmp fast une float [[V3BX1]], 0.000000e+00 + // F32-DAG: [[V3B2:%.*]] = fcmp fast une float [[V3BX2]], 0.000000e+00 + + vector v3b = buf.Load< vector >(56); + vector v3a = buf.Load< vector >(32); + + // CHECK: and i1 [[V3B0]], [[V3A0]] + // CHECK: and i1 [[V3B1]], [[V3A1]] + // CHECK: and i1 [[V3B2]], [[V3A2]] + vector res3 = and(v3a, v3b); + + // CHECK-DAG: [[MAR:%.*]] = call %dx.types.ResRet.[[TY]] @dx.op.bufferLoad.[[TY]](i32 68, %dx.types.Handle %{{.*}}, i32 80 + // CHECK-DAG: [[MAX0:%.*]] = extractvalue %dx.types.ResRet.[[TY]] [[MAR]], 0 + // CHECK-DAG: [[MAX1:%.*]] = extractvalue %dx.types.ResRet.[[TY]] [[MAR]], 1 + // CHECK-DAG: [[MAX2:%.*]] = extractvalue %dx.types.ResRet.[[TY]] [[MAR]], 2 + // CHECK-DAG: [[MAX3:%.*]] = extractvalue %dx.types.ResRet.[[TY]] [[MAR]], 3 + // CHECK-DAG: [[MAR:%.*]] = call %dx.types.ResRet.[[TY]] @dx.op.bufferLoad.[[TY]](i32 68, %dx.types.Handle %{{.*}}, i32 96 + // CHECK-DAG: [[MAX4:%.*]] = extractvalue %dx.types.ResRet.[[TY]] [[MAR]], 0 + // CHECK-DAG: [[MAX5:%.*]] = extractvalue %dx.types.ResRet.[[TY]] [[MAR]], 1 + + // I32-DAG: [[MA0:%.*]] = icmp ne i32 [[MAX0]], 0 + // I32-DAG: [[MA1:%.*]] = icmp ne i32 [[MAX1]], 0 + // I32-DAG: [[MA2:%.*]] = icmp ne i32 [[MAX2]], 0 + // I32-DAG: [[MA3:%.*]] = icmp ne i32 [[MAX3]], 0 + // I32-DAG: [[MA4:%.*]] = icmp ne i32 [[MAX4]], 0 + // I32-DAG: [[MA5:%.*]] = icmp ne i32 [[MAX5]], 0 + + // F32-DAG: [[MA0:%.*]] = fcmp fast une float [[MAX0]], 0.000000e+00 + // F32-DAG: [[MA1:%.*]] = fcmp fast une float [[MAX1]], 0.000000e+00 + // F32-DAG: [[MA2:%.*]] = fcmp fast une float [[MAX2]], 0.000000e+00 + // F32-DAG: [[MA3:%.*]] = fcmp fast une float [[MAX3]], 0.000000e+00 + // F32-DAG: [[MA4:%.*]] = fcmp fast une float [[MAX4]], 0.000000e+00 + // F32-DAG: [[MA5:%.*]] = fcmp fast une float [[MAX5]], 0.000000e+00 + + // CHECK-DAG: [[MBR:%.*]] = call %dx.types.ResRet.[[TY]] @dx.op.bufferLoad.[[TY]](i32 68, %dx.types.Handle %{{.*}}, i32 128 + // CHECK-DAG: [[MBX0:%.*]] = extractvalue %dx.types.ResRet.[[TY]] [[MBR]], 0 + // CHECK-DAG: [[MBX1:%.*]] = extractvalue %dx.types.ResRet.[[TY]] [[MBR]], 1 + // CHECK-DAG: [[MBX2:%.*]] = extractvalue %dx.types.ResRet.[[TY]] [[MBR]], 2 + // CHECK-DAG: [[MBX3:%.*]] = extractvalue %dx.types.ResRet.[[TY]] [[MBR]], 3 + // CHECK-DAG: [[MBR:%.*]] = call %dx.types.ResRet.[[TY]] @dx.op.bufferLoad.[[TY]](i32 68, %dx.types.Handle %{{.*}}, i32 144 + // CHECK-DAG: [[MBX4:%.*]] = extractvalue %dx.types.ResRet.[[TY]] [[MBR]], 0 + // CHECK-DAG: [[MBX5:%.*]] = extractvalue %dx.types.ResRet.[[TY]] [[MBR]], 1 + + // I32-DAG: [[MB0:%.*]] = icmp ne i32 [[MBX0]], 0 + // I32-DAG: [[MB1:%.*]] = icmp ne i32 [[MBX1]], 0 + // I32-DAG: [[MB2:%.*]] = icmp ne i32 [[MBX2]], 0 + // I32-DAG: [[MB3:%.*]] = icmp ne i32 [[MBX3]], 0 + // I32-DAG: [[MB4:%.*]] = icmp ne i32 [[MBX4]], 0 + // I32-DAG: [[MB5:%.*]] = icmp ne i32 [[MBX5]], 0 + + // F32-DAG: [[MB0:%.*]] = fcmp fast une float [[MBX0]], 0.000000e+00 + // F32-DAG: [[MB1:%.*]] = fcmp fast une float [[MBX1]], 0.000000e+00 + // F32-DAG: [[MB2:%.*]] = fcmp fast une float [[MBX2]], 0.000000e+00 + // F32-DAG: [[MB3:%.*]] = fcmp fast une float [[MBX3]], 0.000000e+00 + // F32-DAG: [[MB4:%.*]] = fcmp fast une float [[MBX4]], 0.000000e+00 + // F32-DAG: [[MB5:%.*]] = fcmp fast une float [[MBX5]], 0.000000e+00 + + matrix matb = buf.Load< matrix >(128); + matrix mata = buf.Load< matrix >(80); + + // CHECK: and i1 [[MB0]], [[MA0]] + // CHECK: and i1 [[MB1]], [[MA1]] + // CHECK: and i1 [[MB2]], [[MA2]] + // CHECK: and i1 [[MB3]], [[MA3]] + // CHECK: and i1 [[MB4]], [[MA4]] + // CHECK: and i1 [[MB5]], [[MA5]] + matrix resmat = and(mata, matb); + + return float4(res3 + resmat[0] + resmat[1], res + res1.x); +} diff --git a/tools/clang/test/CodeGenHLSL/vector-or.hlsl b/tools/clang/test/CodeGenHLSL/vector-or.hlsl new file mode 100644 index 0000000000..2fe6c72434 --- /dev/null +++ b/tools/clang/test/CodeGenHLSL/vector-or.hlsl @@ -0,0 +1,164 @@ +// RUN: %dxc -T ps_6_0 -HV 2021 -DTYPE=bool %s | FileCheck %s --check-prefixes=CHECK,I32 +// RUN: %dxc -T ps_6_0 -HV 2018 -DTYPE=bool %s | FileCheck %s --check-prefixes=CHECK,I32 +// RUN: %dxc -T ps_6_0 -HV 2021 -DTYPE=int %s | FileCheck %s --check-prefixes=CHECK,I32 +// RUN: %dxc -T ps_6_0 -HV 2018 -DTYPE=int %s | FileCheck %s --check-prefixes=CHECK,I32 +// RUN: %dxc -T ps_6_0 -HV 2021 -DTYPE=float %s | FileCheck %s --check-prefixes=CHECK,F32 +// RUN: %dxc -T ps_6_0 -HV 2018 -DTYPE=float %s | FileCheck %s --check-prefixes=CHECK,F32 + +// I32: %dx.types.ResRet.[[TY:i32]] = type { [[TYPE:i32]] +// F32: %dx.types.ResRet.[[TY:f32]] = type { [[TYPE:float]] + +// CHECK-LABEL: define void @main + +ByteAddressBuffer buf; + +float4 main() : SV_Target { + + // CHECK: [[SAR:%.*]] = call %dx.types.ResRet.[[TY]] @dx.op.bufferLoad.[[TY]](i32 68, %dx.types.Handle %{{.*}}, i32 0 + // F32: [[SAX:%.*]] = extractvalue %dx.types.ResRet.[[TY]] [[SAR]], 0 + // I32: [[SA:%.*]] = extractvalue %dx.types.ResRet.[[TY]] [[SAR]], 0 + + // CHECK: [[SBR:%.*]] = call %dx.types.ResRet.[[TY]] @dx.op.bufferLoad.[[TY]](i32 68, %dx.types.Handle %{{.*}}, i32 8 + // F32: [[SBX:%.*]] = extractvalue %dx.types.ResRet.[[TY]] [[SBR]], 0 + // I32: [[SB:%.*]] = extractvalue %dx.types.ResRet.[[TY]] [[SBR]], 0 + + // F32: [[SA:%.*]] = fcmp fast une float [[SAX]], 0.000000e+00 + // F32: [[SB:%.*]] = fcmp fast une float [[SBX]], 0.000000e+00 + + TYPE sa = buf.Load(0); + TYPE sb = buf.Load(8); + + // I32: or i32 [[SB]], [[SA]] + // F32: or i1 [[SA]], [[SB]] + + TYPE res = or(sb, sa); + + // CHECK: [[V1AR:%.*]] = call %dx.types.ResRet.[[TY]] @dx.op.bufferLoad.[[TY]](i32 68, %dx.types.Handle %{{.*}}, i32 16 + // F32: [[V1AX:%.*]] = extractvalue %dx.types.ResRet.[[TY]] [[V1AR]], 0 + // I32: [[V1A:%.*]] = extractvalue %dx.types.ResRet.[[TY]] [[V1AR]], 0 + + // CHECK: [[V1BR:%.*]] = call %dx.types.ResRet.[[TY]] @dx.op.bufferLoad.[[TY]](i32 68, %dx.types.Handle %{{.*}}, i32 24 + // F32: [[V1BX:%.*]] = extractvalue %dx.types.ResRet.[[TY]] [[V1BR]], 0 + // I32: [[V1B:%.*]] = extractvalue %dx.types.ResRet.[[TY]] [[V1BR]], 0 + + // F32: [[V1B:%.*]] = fcmp fast une float [[V1BX]], 0.000000e+00 + // F32: [[V1A:%.*]] = fcmp fast une float [[V1AX]], 0.000000e+00 + + vector v1a = buf.Load< vector >(16); + vector v1b = buf.Load< vector >(24); + + // I32: or i32 [[V1B]], [[V1A]] + // F32: or i1 [[V1A]], [[V1B]] + + vector res1 = or(v1a, v1b); + + // CHECK: [[V3AR:%.*]] = call %dx.types.ResRet.[[TY]] @dx.op.bufferLoad.[[TY]](i32 68, %dx.types.Handle %{{.*}}, i32 32 + // F32: [[V3AX0:%.*]] = extractvalue %dx.types.ResRet.[[TY]] [[V3AR]], 0 + // F32: [[V3AX1:%.*]] = extractvalue %dx.types.ResRet.[[TY]] [[V3AR]], 1 + // F32: [[V3AX2:%.*]] = extractvalue %dx.types.ResRet.[[TY]] [[V3AR]], 2 + + // I32: [[V3A0:%.*]] = extractvalue %dx.types.ResRet.[[TY]] [[V3AR]], 0 + // I32: [[V3A1:%.*]] = extractvalue %dx.types.ResRet.[[TY]] [[V3AR]], 1 + // I32: [[V3A2:%.*]] = extractvalue %dx.types.ResRet.[[TY]] [[V3AR]], 2 + + // CHECK: [[V3BR:%.*]] = call %dx.types.ResRet.[[TY]] @dx.op.bufferLoad.[[TY]](i32 68, %dx.types.Handle %{{.*}}, i32 56 + // F32: [[V3BX0:%.*]] = extractvalue %dx.types.ResRet.[[TY]] [[V3BR]], 0 + // F32: [[V3BX1:%.*]] = extractvalue %dx.types.ResRet.[[TY]] [[V3BR]], 1 + // F32: [[V3BX2:%.*]] = extractvalue %dx.types.ResRet.[[TY]] [[V3BR]], 2 + + // I32: [[V3B0:%.*]] = extractvalue %dx.types.ResRet.[[TY]] [[V3BR]], 0 + // I32: [[V3B1:%.*]] = extractvalue %dx.types.ResRet.[[TY]] [[V3BR]], 1 + // I32: [[V3B2:%.*]] = extractvalue %dx.types.ResRet.[[TY]] [[V3BR]], 2 + + // F32: [[V3B0:%.*]] = fcmp fast une float [[V3BX0]], 0.000000e+00 + // F32: [[V3B1:%.*]] = fcmp fast une float [[V3BX1]], 0.000000e+00 + // F32: [[V3B2:%.*]] = fcmp fast une float [[V3BX2]], 0.000000e+00 + + // F32: [[V3A0:%.*]] = fcmp fast une float [[V3AX0]], 0.000000e+00 + // F32: [[V3A1:%.*]] = fcmp fast une float [[V3AX1]], 0.000000e+00 + // F32: [[V3A2:%.*]] = fcmp fast une float [[V3AX2]], 0.000000e+00 + + vector v3a = buf.Load< vector >(32); + vector v3b = buf.Load< vector >(56); + + // I32: or i32 [[V3B0]], [[V3A0]] + // I32: or i32 [[V3B1]], [[V3A1]] + // I32: or i32 [[V3B2]], [[V3A2]] + + // F32: or i1 [[V3A0]], [[V3B0]] + // F32: or i1 [[V3A1]], [[V3B1]] + // F32: or i1 [[V3A2]], [[V3B2]] + + vector res3 = or(v3a, v3b); + + // CHECK: [[MAR:%.*]] = call %dx.types.ResRet.[[TY]] @dx.op.bufferLoad.[[TY]](i32 68, %dx.types.Handle %{{.*}}, i32 80 + // F32: [[MAX0:%.*]] = extractvalue %dx.types.ResRet.[[TY]] [[MAR]], 0 + // F32: [[MAX1:%.*]] = extractvalue %dx.types.ResRet.[[TY]] [[MAR]], 1 + // F32: [[MAX2:%.*]] = extractvalue %dx.types.ResRet.[[TY]] [[MAR]], 2 + // F32: [[MAX3:%.*]] = extractvalue %dx.types.ResRet.[[TY]] [[MAR]], 3 + + // I32: [[MA0:%.*]] = extractvalue %dx.types.ResRet.[[TY]] [[MAR]], 0 + // I32: [[MA1:%.*]] = extractvalue %dx.types.ResRet.[[TY]] [[MAR]], 1 + // I32: [[MA2:%.*]] = extractvalue %dx.types.ResRet.[[TY]] [[MAR]], 2 + // I32: [[MA3:%.*]] = extractvalue %dx.types.ResRet.[[TY]] [[MAR]], 3 + + // CHECK: [[MAR:%.*]] = call %dx.types.ResRet.[[TY]] @dx.op.bufferLoad.[[TY]](i32 68, %dx.types.Handle %{{.*}}, i32 96 + // F32: [[MAX4:%.*]] = extractvalue %dx.types.ResRet.[[TY]] [[MAR]], 0 + // F32: [[MAX5:%.*]] = extractvalue %dx.types.ResRet.[[TY]] [[MAR]], 1 + + // I32: [[MA4:%.*]] = extractvalue %dx.types.ResRet.[[TY]] [[MAR]], 0 + // I32: [[MA5:%.*]] = extractvalue %dx.types.ResRet.[[TY]] [[MAR]], 1 + + // CHECK: [[MBR:%.*]] = call %dx.types.ResRet.[[TY]] @dx.op.bufferLoad.[[TY]](i32 68, %dx.types.Handle %{{.*}}, i32 128 + // F32: [[MBX0:%.*]] = extractvalue %dx.types.ResRet.[[TY]] [[MBR]], 0 + // F32: [[MBX1:%.*]] = extractvalue %dx.types.ResRet.[[TY]] [[MBR]], 1 + // F32: [[MBX2:%.*]] = extractvalue %dx.types.ResRet.[[TY]] [[MBR]], 2 + // F32: [[MBX3:%.*]] = extractvalue %dx.types.ResRet.[[TY]] [[MBR]], 3 + + // I32: [[MB0:%.*]] = extractvalue %dx.types.ResRet.[[TY]] [[MBR]], 0 + // I32: [[MB1:%.*]] = extractvalue %dx.types.ResRet.[[TY]] [[MBR]], 1 + // I32: [[MB2:%.*]] = extractvalue %dx.types.ResRet.[[TY]] [[MBR]], 2 + // I32: [[MB3:%.*]] = extractvalue %dx.types.ResRet.[[TY]] [[MBR]], 3 + + // CHECK: [[MBR:%.*]] = call %dx.types.ResRet.[[TY]] @dx.op.bufferLoad.[[TY]](i32 68, %dx.types.Handle %{{.*}}, i32 144 + // F32: [[MBX4:%.*]] = extractvalue %dx.types.ResRet.[[TY]] [[MBR]], 0 + // F32: [[MBX5:%.*]] = extractvalue %dx.types.ResRet.[[TY]] [[MBR]], 1 + + // I32: [[MB4:%.*]] = extractvalue %dx.types.ResRet.[[TY]] [[MBR]], 0 + // I32: [[MB5:%.*]] = extractvalue %dx.types.ResRet.[[TY]] [[MBR]], 1 + + // F32: [[MB0:%.*]] = fcmp fast une float [[MBX0]], 0.000000e+00 + // F32: [[MB1:%.*]] = fcmp fast une float [[MBX1]], 0.000000e+00 + // F32: [[MB2:%.*]] = fcmp fast une float [[MBX2]], 0.000000e+00 + // F32: [[MB3:%.*]] = fcmp fast une float [[MBX3]], 0.000000e+00 + // F32: [[MB4:%.*]] = fcmp fast une float [[MBX4]], 0.000000e+00 + // F32: [[MB5:%.*]] = fcmp fast une float [[MBX5]], 0.000000e+00 + + // F32: [[MA0:%.*]] = fcmp fast une float [[MAX0]], 0.000000e+00 + // F32: [[MA1:%.*]] = fcmp fast une float [[MAX1]], 0.000000e+00 + // F32: [[MA2:%.*]] = fcmp fast une float [[MAX2]], 0.000000e+00 + // F32: [[MA3:%.*]] = fcmp fast une float [[MAX3]], 0.000000e+00 + // F32: [[MA4:%.*]] = fcmp fast une float [[MAX4]], 0.000000e+00 + // F32: [[MA5:%.*]] = fcmp fast une float [[MAX5]], 0.000000e+00 + + matrix mata = buf.Load< matrix >(80); + matrix matb = buf.Load< matrix >(128); + + // I32: or i32 [[MB0]], [[MA0]] + // I32: or i32 [[MB1]], [[MA1]] + // I32: or i32 [[MB2]], [[MA2]] + // I32: or i32 [[MB3]], [[MA3]] + // I32: or i32 [[MB4]], [[MA4]] + // I32: or i32 [[MB5]], [[MA5]] + + // F32: or i1 [[MA0]], [[MB0]] + // F32: or i1 [[MA1]], [[MB1]] + // F32: or i1 [[MA2]], [[MB2]] + // F32: or i1 [[MA3]], [[MB3]] + // F32: or i1 [[MA4]], [[MB4]] + // F32: or i1 [[MA5]], [[MB5]] + + matrix resmat = or(mata, matb); + + return float4(res3 + resmat[0] + resmat[1], res + res1.x); +} diff --git a/tools/clang/test/CodeGenSPIRV/intrinsics.and.hlsl b/tools/clang/test/CodeGenSPIRV/intrinsics.and.hlsl index a983e20c45..970faf6a8d 100644 --- a/tools/clang/test/CodeGenSPIRV/intrinsics.and.hlsl +++ b/tools/clang/test/CodeGenSPIRV/intrinsics.and.hlsl @@ -1,6 +1,9 @@ // RUN: %dxc -T ps_6_0 -E main -HV 2021 -fcgl %s -spirv | FileCheck %s // RUN: %dxc -T ps_6_0 -E main -HV 2018 -fcgl %s -spirv | FileCheck %s +// CHECK-DAG: [[v3_0:%[0-9]+]] = OpConstantComposite %v3int %int_0 %int_0 %int_0 +// CHECK-DAG: [[v3_1:%[0-9]+]] = OpConstantComposite %v3int %int_1 %int_1 %int_1 + void main() { // CHECK-LABEL: %bb_entry = OpLabel @@ -33,4 +36,46 @@ void main() { // CHECK-NEXT: [[and3:%[0-9]+]] = OpLogicalAnd %bool [[a1]] [[b1]] // CHECK-NEXT: {{%[0-9]+}} = OpCompositeConstruct %v2bool [[and3]] %true bool2 t = bool2(and(a, b), true); + + int a_0, b_0, c_0; + // Plain assign (scalar) +// CHECK: [[a0_int:%[0-9]+]] = OpLoad %int %a_0 +// CHECK-NEXT: [[a0:%[0-9]+]] = OpINotEqual %bool [[a0_int]] %int_0 +// CHECK-NEXT: [[b0_int:%[0-9]+]] = OpLoad %int %b_0 +// CHECK-NEXT: [[b0:%[0-9]+]] = OpINotEqual %bool [[b0_int]] %int_0 +// CHECK-NEXT: [[and0:%[0-9]+]] = OpLogicalAnd %bool [[a0]] [[b0]] +// CHECK-NEXT: [[sel:%[0-9]+]] = OpSelect %int [[and0]] %int_1 %int_0 +// CHECK-NEXT: OpStore %c_0 [[sel]] + c_0 = and(a_0, b_0); + + int1 i_0, j_0, k_0; + int3 o_0, p_0, q_0; + // Plain assign (vector) +// CHECK-NEXT: [[i0_int:%[0-9]+]] = OpLoad %int %i_0 +// CHECK-NEXT: [[i0:%[0-9]+]] = OpINotEqual %bool [[i0_int]] %int_0 +// CHECK-NEXT: [[j0_int:%[0-9]+]] = OpLoad %int %j_0 +// CHECK-NEXT: [[j0:%[0-9]+]] = OpINotEqual %bool [[j0_int]] %int_0 +// CHECK-NEXT: [[and1:%[0-9]+]] = OpLogicalAnd %bool [[i0]] [[j0]] +// CHECK-NEXT: [[sel:%[0-9]+]] = OpSelect %int [[and1]] %int_1 %int_0 +// CHECK-NEXT: OpStore %k_0 [[sel]] + k_0 = and(i_0, j_0); + +// CHECK-NEXT: [[o0_int:%[0-9]+]] = OpLoad %v3int %o_0 +// CHECK-NEXT: [[o0:%[0-9]+]] = OpINotEqual %v3bool [[o0_int]] [[v3_0]] +// CHECK-NEXT: [[p0_int:%[0-9]+]] = OpLoad %v3int %p_0 +// CHECK-NEXT: [[p0:%[0-9]+]] = OpINotEqual %v3bool [[p0_int]] [[v3_0]] +// CHECK-NEXT: [[and2:%[0-9]+]] = OpLogicalAnd %v3bool [[o0]] [[p0]] +// CHECK-NEXT: [[sel:%[0-9]+]] = OpSelect %v3int [[and2]] [[v3_1]] [[v3_0]] +// CHECK-NEXT: OpStore %q_0 [[sel]] + q_0 = and(o_0, p_0); + +// The result of '&&' could be 'const bool'. In such cases, make sure +// the result type is correct. +// CHECK: [[a0_int:%[0-9]+]] = OpLoad %int %a_0 +// CHECK-NEXT: [[a0:%[0-9]+]] = OpINotEqual %bool [[a0_int]] %int_0 +// CHECK-NEXT: [[b0_int:%[0-9]+]] = OpLoad %int %b_0 +// CHECK-NEXT: [[b0:%[0-9]+]] = OpINotEqual %bool [[b0_int]] %int_0 +// CHECK-NEXT: [[and0:%[0-9]+]] = OpLogicalAnd %bool [[a0]] [[b0]] +// CHECK-NEXT: {{%[0-9]+}} = OpCompositeConstruct %v2bool [[and0]] %true + t = bool2(and(a_0, b_0), true); } diff --git a/tools/clang/test/CodeGenSPIRV/intrinsics.or.hlsl b/tools/clang/test/CodeGenSPIRV/intrinsics.or.hlsl index a61272211e..39a39062f1 100644 --- a/tools/clang/test/CodeGenSPIRV/intrinsics.or.hlsl +++ b/tools/clang/test/CodeGenSPIRV/intrinsics.or.hlsl @@ -1,6 +1,8 @@ // RUN: %dxc -T ps_6_0 -E main -HV 2021 -fcgl %s -spirv | FileCheck %s // RUN: %dxc -T ps_6_0 -E main -HV 2018 -fcgl %s -spirv | FileCheck %s +// CHECK: [[v3_0:%[0-9]+]] = OpConstantComposite %v3int %int_0 %int_0 %int_0 + void main() { // CHECK-LABEL: %bb_entry = OpLabel @@ -25,4 +27,36 @@ void main() { // CHECK-NEXT: OpStore %q [[or2]] k = or(i, j); q = or(o, p); + + int r, s; + bool t; + // Plain assign (scalar) +// CHECK: [[r0_int:%[0-9]+]] = OpLoad %int %r +// CHECK-NEXT: [[r0:%[0-9]+]] = OpINotEqual %bool [[r0_int]] %int_0 +// CHECK-NEXT: [[s0_int:%[0-9]+]] = OpLoad %int %s +// CHECK-NEXT: [[s0:%[0-9]+]] = OpINotEqual %bool [[s0_int]] %int_0 +// CHECK-NEXT: [[or0:%[0-9]+]] = OpLogicalOr %bool [[r0]] [[s0]] +// CHECK-NEXT: OpStore %t [[or0]] + t = or(r, s); + + int1 u, v; + bool1 w; + // Plain assign (vector) +// CHECK-NEXT: [[u0_int:%[0-9]+]] = OpLoad %int %u +// CHECK-NEXT: [[u0:%[0-9]+]] = OpINotEqual %bool [[u0_int]] %int_0 +// CHECK-NEXT: [[v0_int:%[0-9]+]] = OpLoad %int %v +// CHECK-NEXT: [[v0:%[0-9]+]] = OpINotEqual %bool [[v0_int]] %int_0 +// CHECK-NEXT: [[or1:%[0-9]+]] = OpLogicalOr %bool [[u0]] [[v0]] +// CHECK-NEXT: OpStore %w [[or1]] + w = or(u, v); + + int3 x, y; + bool3 z; +// CHECK-NEXT: [[x0_int:%[0-9]+]] = OpLoad %v3int %x +// CHECK-NEXT: [[x0:%[0-9]+]] = OpINotEqual %v3bool [[x0_int]] [[v3_0]] +// CHECK-NEXT: [[y0_int:%[0-9]+]] = OpLoad %v3int %y +// CHECK-NEXT: [[y0:%[0-9]+]] = OpINotEqual %v3bool [[y0_int]] [[v3_0]] +// CHECK-NEXT: [[or2:%[0-9]+]] = OpLogicalOr %v3bool [[x0]] [[y0]] +// CHECK-NEXT: OpStore %z [[or2]] + z = or(x, y); } diff --git a/utils/hct/gen_intrin_main.txt b/utils/hct/gen_intrin_main.txt index 40b5f6d96a..8ee8d0bc99 100644 --- a/utils/hct/gen_intrin_main.txt +++ b/utils/hct/gen_intrin_main.txt @@ -369,8 +369,8 @@ resource [[hidden]] CreateResourceFromHeap(in uint index); // Replacement for vector logical &&, ||, and ternary conditional operators, // For use when HLSL changes to support short-circuiting and only scalar // conditions to maintain clarity. -$match<1, 0> bool<> [[rn]] and(in any<> x, in $type1 y); -$match<1, 0> bool<> [[rn]] or(in any<> x, in $type1 y); +$match<1, 0> bool<> [[rn]] and(in bool<> x, in $type1 y); +$match<1, 0> bool<> [[rn]] or(in bool<> x, in $type1 y); $type2 [[rn]] select(in bool<> cond, in $match<1, 2> any<> t, in $type2 f); $type2 [[rn]] select(in bool cond, in any_sampler t, in $type2 f); From 10b3328a915e30973449aa03a1816d289e4fd42f Mon Sep 17 00:00:00 2001 From: Joshua Batista Date: Fri, 24 Jan 2025 17:23:20 -0800 Subject: [PATCH 09/20] Add Preview Hashing sentinel value to dxil validator (#7053) When the dxil validator detects that the target shader model is a preview shader model that hasn't officially been released, it should sign the dxil container with a special sentinel hash: 02020202020202020202020202020202 Fixes https://github.com/microsoft/DirectXShaderCompiler/issues/7083 --------- Co-authored-by: github-actions[bot] --- include/dxc/DXIL/DxilShaderModel.h | 13 +- include/dxc/DxilContainer/DxilContainer.h | 3 + include/dxc/Support/ErrorCodes.h | 5 + lib/DXIL/DxilShaderModel.cpp | 27 +++- .../DxilContainerValidation.cpp | 23 ++- .../clang/tools/dxcvalidator/dxcvalidator.cpp | 32 +++- tools/clang/unittests/HLSL/ValidationTest.cpp | 144 ++++++++++++++++-- utils/hct/hctdb.py | 5 + utils/hct/hctdb_instrhelp.py | 21 ++- 9 files changed, 250 insertions(+), 23 deletions(-) diff --git a/include/dxc/DXIL/DxilShaderModel.h b/include/dxc/DXIL/DxilShaderModel.h index c89c079bb9..c5c5d66884 100644 --- a/include/dxc/DXIL/DxilShaderModel.h +++ b/include/dxc/DXIL/DxilShaderModel.h @@ -26,7 +26,7 @@ class ShaderModel { public: using Kind = DXIL::ShaderKind; - // Major/Minor version of highest shader model + // Major/Minor version of highest recognized shader model // clang-format off // Python lines need to be not formatted. /* hctdb_instrhelp.get_highest_shader_model()*/ @@ -35,6 +35,15 @@ class ShaderModel { static const unsigned kHighestMajor = 6; static const unsigned kHighestMinor = 9; // VALRULE-TEXT:END + + // Major/Minor version of highest released shader model + /* hctdb_instrhelp.get_highest_released_shader_model()*/ + // clang-format on + // VALRULE-TEXT:BEGIN + static const unsigned kHighestReleasedMajor = 6; + static const unsigned kHighestReleasedMinor = 8; + // VALRULE-TEXT:END + static const unsigned kOfflineMinor = 0xF; bool IsPS() const { return m_Kind == Kind::Pixel; } @@ -86,6 +95,8 @@ class ShaderModel { static const ShaderModel *Get(Kind Kind, unsigned Major, unsigned Minor); static const ShaderModel *GetByName(llvm::StringRef Name); static const char *GetKindName(Kind kind); + static bool IsPreReleaseShaderModel(int Major, int Minor); + static Kind GetKindFromName(llvm::StringRef Name); static DXIL::ShaderKind KindFromFullName(llvm::StringRef Name); static const llvm::StringRef FullNameFromKind(DXIL::ShaderKind sk); static const char *GetNodeLaunchTypeName(DXIL::NodeLaunchType launchTy); diff --git a/include/dxc/DxilContainer/DxilContainer.h b/include/dxc/DxilContainer/DxilContainer.h index 80e6308430..8b7d85954b 100644 --- a/include/dxc/DxilContainer/DxilContainer.h +++ b/include/dxc/DxilContainer/DxilContainer.h @@ -36,6 +36,9 @@ struct DxilContainerHash { uint8_t Digest[DxilContainerHashSize]; }; +static const DxilContainerHash PreviewByPassHash = {2, 2, 2, 2, 2, 2, 2, 2, + 2, 2, 2, 2, 2, 2, 2, 2}; + enum class DxilShaderHashFlags : uint32_t { None = 0, // No flags defined. IncludesSource = 1, // This flag indicates that the shader hash was computed diff --git a/include/dxc/Support/ErrorCodes.h b/include/dxc/Support/ErrorCodes.h index 7a5830fe8f..5239c8118c 100644 --- a/include/dxc/Support/ErrorCodes.h +++ b/include/dxc/Support/ErrorCodes.h @@ -153,3 +153,8 @@ // 0X80AA001E - External validator (DXIL.dll) required, and missing. #define DXC_E_VALIDATOR_MISSING \ DXC_MAKE_HRESULT(DXC_SEVERITY_ERROR, FACILITY_DXC, (0x001E)) + +// 0X80AA001F - DXIL container Program Version mismatches Dxil module shader +// model +#define DXC_E_INCORRECT_PROGRAM_VERSION \ + DXC_MAKE_HRESULT(DXC_SEVERITY_ERROR, FACILITY_DXC, (0x001F)) \ No newline at end of file diff --git a/lib/DXIL/DxilShaderModel.cpp b/lib/DXIL/DxilShaderModel.cpp index 8c94b3f5a1..861e8586e0 100644 --- a/lib/DXIL/DxilShaderModel.cpp +++ b/lib/DXIL/DxilShaderModel.cpp @@ -193,11 +193,21 @@ const ShaderModel *ShaderModel::Get(Kind Kind, unsigned Major, unsigned Minor) { // VALRULE-TEXT:END } -const ShaderModel *ShaderModel::GetByName(llvm::StringRef Name) { - // [ps|vs|gs|hs|ds|cs|ms|as]_[major]_[minor] +bool ShaderModel::IsPreReleaseShaderModel(int major, int minor) { + if (DXIL::CompareVersions(major, minor, kHighestReleasedMajor, + kHighestReleasedMinor) <= 0) + return false; + + // now compare against highest recognized + if (DXIL::CompareVersions(major, minor, kHighestMajor, kHighestMinor) <= 0) + return true; + return false; +} + +ShaderModel::Kind ShaderModel::GetKindFromName(llvm::StringRef Name) { Kind kind; if (Name.empty()) { - return GetInvalid(); + return Kind::Invalid; } switch (Name[0]) { @@ -229,8 +239,17 @@ const ShaderModel *ShaderModel::GetByName(llvm::StringRef Name) { kind = Kind::Amplification; break; default: - return GetInvalid(); + return Kind::Invalid; } + return kind; +} + +const ShaderModel *ShaderModel::GetByName(llvm::StringRef Name) { + // [ps|vs|gs|hs|ds|cs|ms|as]_[major]_[minor] + Kind kind = GetKindFromName(Name); + if (kind == Kind::Invalid) + return GetInvalid(); + unsigned Idx = 3; if (kind != Kind::Library) { if (Name[1] != 's' || Name[2] != '_') diff --git a/lib/DxilValidation/DxilContainerValidation.cpp b/lib/DxilValidation/DxilContainerValidation.cpp index 2276b0d3de..890e90e354 100644 --- a/lib/DxilValidation/DxilContainerValidation.cpp +++ b/lib/DxilValidation/DxilContainerValidation.cpp @@ -1033,8 +1033,29 @@ HRESULT ValidateDxilContainerParts(llvm::Module *pModule, case DFCC_ResourceDef: case DFCC_ShaderStatistics: case DFCC_PrivateData: + break; case DFCC_DXIL: - case DFCC_ShaderDebugInfoDXIL: + case DFCC_ShaderDebugInfoDXIL: { + const DxilProgramHeader *pProgramHeader = + reinterpret_cast(GetDxilPartData(pPart)); + if (!pProgramHeader) + continue; + + int PV = pProgramHeader->ProgramVersion; + int major = (PV >> 4) & 0xF; // Extract the major version (next 4 bits) + int minor = PV & 0xF; // Extract the minor version (lowest 4 bits) + + int moduleMajor = pDxilModule->GetShaderModel()->GetMajor(); + int moduleMinor = pDxilModule->GetShaderModel()->GetMinor(); + if (moduleMajor != major || moduleMinor != minor) { + ValCtx.EmitFormatError(ValidationRule::SmProgramVersion, + {std::to_string(major), std::to_string(minor), + std::to_string(moduleMajor), + std::to_string(moduleMinor)}); + return DXC_E_INCORRECT_PROGRAM_VERSION; + } + continue; + } case DFCC_ShaderDebugName: continue; diff --git a/tools/clang/tools/dxcvalidator/dxcvalidator.cpp b/tools/clang/tools/dxcvalidator/dxcvalidator.cpp index b8b71ece62..60ad35036f 100644 --- a/tools/clang/tools/dxcvalidator/dxcvalidator.cpp +++ b/tools/clang/tools/dxcvalidator/dxcvalidator.cpp @@ -20,6 +20,7 @@ #include "dxc/dxcapi.h" #include "dxcvalidator.h" +#include "dxc/DXIL/DxilShaderModel.h" #include "dxc/DxilRootSignature/DxilRootSignature.h" #include "dxc/Support/FileIOHelper.h" #include "dxc/Support/Global.h" @@ -32,7 +33,13 @@ using namespace llvm; using namespace hlsl; -static void HashAndUpdate(DxilContainerHeader *Container) { +static void HashAndUpdate(DxilContainerHeader *Container, bool isPreRelease) { + if (isPreRelease) { + // If preview bypass is enabled, use the preview hash. + memcpy(Container->Hash.Digest, PreviewByPassHash.Digest, + sizeof(PreviewByPassHash.Digest)); + return; + } // Compute hash and update stored hash. // Hash the container from this offset to the end. static const uint32_t DXBCHashStartOffset = @@ -45,8 +52,26 @@ static void HashAndUpdate(DxilContainerHeader *Container) { static void HashAndUpdateOrCopy(uint32_t Flags, IDxcBlob *Shader, IDxcBlob **Hashed) { + bool isPreRelease = false; + const DxilContainerHeader *DxilContainer = + IsDxilContainerLike(Shader->GetBufferPointer(), Shader->GetBufferSize()); + if (!DxilContainer) + return; + + const DxilProgramHeader *ProgramHeader = + GetDxilProgramHeader(DxilContainer, DFCC_DXIL); + + // ProgramHeader may be null here, when hashing a root signature container + if (ProgramHeader) { + int PV = ProgramHeader->ProgramVersion; + int major = (PV >> 4) & 0xF; // Extract the major version (next 4 bits) + int minor = PV & 0xF; // Extract the minor version (lowest 4 bits) + isPreRelease = ShaderModel::IsPreReleaseShaderModel(major, minor); + } + if (Flags & DxcValidatorFlags_InPlaceEdit) { - HashAndUpdate((DxilContainerHeader *)Shader->GetBufferPointer()); + HashAndUpdate((DxilContainerHeader *)Shader->GetBufferPointer(), + isPreRelease); *Hashed = Shader; Shader->AddRef(); } else { @@ -55,7 +80,8 @@ static void HashAndUpdateOrCopy(uint32_t Flags, IDxcBlob *Shader, unsigned long CB; IFT(HashedBlobStream->Write(Shader->GetBufferPointer(), Shader->GetBufferSize(), &CB)); - HashAndUpdate((DxilContainerHeader *)HashedBlobStream->GetPtr()); + HashAndUpdate((DxilContainerHeader *)HashedBlobStream->GetPtr(), + isPreRelease); IFT(HashedBlobStream.QueryInterface(Hashed)); } } diff --git a/tools/clang/unittests/HLSL/ValidationTest.cpp b/tools/clang/unittests/HLSL/ValidationTest.cpp index 08f67f35d0..8008541bfa 100644 --- a/tools/clang/unittests/HLSL/ValidationTest.cpp +++ b/tools/clang/unittests/HLSL/ValidationTest.cpp @@ -31,6 +31,7 @@ #include "dxc/Support/FileIOHelper.h" #include "dxc/Support/Global.h" +#include "dxc/DXIL/DxilShaderModel.h" #include "dxc/Test/DxcTestUtils.h" #include "dxc/Test/HlslTestUtils.h" @@ -300,6 +301,8 @@ class ValidationTest : public ::testing::Test { TEST_METHOD(ValidateWithHash) TEST_METHOD(ValidateVersionNotAllowed) + TEST_METHOD(ValidatePreviewBypassHash) + TEST_METHOD(ValidateProgramVersionAgainstDxilModule) TEST_METHOD(CreateHandleNotAllowedSM66) TEST_METHOD(AtomicsConsts) @@ -537,18 +540,10 @@ class ValidationTest : public ::testing::Test { pLookFors, pReplacements, pErrorMsgs, bRegex); } - bool RewriteAssemblyToText(IDxcBlobEncoding *pSource, LPCSTR pShaderModel, - LPCWSTR *pArguments, UINT32 argCount, - const DxcDefine *pDefines, UINT32 defineCount, - llvm::ArrayRef pLookFors, - llvm::ArrayRef pReplacements, - IDxcBlob **pBlob, bool bRegex = false) { - CComPtr pProgram; - std::string disassembly; - if (!CompileSource(pSource, pShaderModel, pArguments, argCount, pDefines, - defineCount, &pProgram)) - return false; - DisassembleProgram(pProgram, &disassembly); + void PerformReplacementOnDisassembly(std::string disassembly, + llvm::ArrayRef pLookFors, + llvm::ArrayRef pReplacements, + IDxcBlob **pBlob, bool bRegex = false) { for (unsigned i = 0; i < pLookFors.size(); ++i) { LPCSTR pLookFor = pLookFors[i]; bool bOptional = false; @@ -605,6 +600,22 @@ class ValidationTest : public ::testing::Test { } } Utf8ToBlob(m_dllSupport, disassembly.c_str(), pBlob); + } + + bool RewriteAssemblyToText(IDxcBlobEncoding *pSource, LPCSTR pShaderModel, + LPCWSTR *pArguments, UINT32 argCount, + const DxcDefine *pDefines, UINT32 defineCount, + llvm::ArrayRef pLookFors, + llvm::ArrayRef pReplacements, + IDxcBlob **pBlob, bool bRegex = false) { + CComPtr pProgram; + std::string disassembly; + if (!CompileSource(pSource, pShaderModel, pArguments, argCount, pDefines, + defineCount, &pProgram)) + return false; + DisassembleProgram(pProgram, &disassembly); + PerformReplacementOnDisassembly(disassembly, pLookFors, pReplacements, + pBlob, bRegex); return true; } @@ -4114,7 +4125,7 @@ TEST_F(ValidationTest, ValidatePrintfNotAllowed) { } TEST_F(ValidationTest, ValidateWithHash) { - if (m_ver.SkipDxilVersion(1, 8)) + if (m_ver.SkipDxilVersion(1, ShaderModel::kHighestReleasedMinor)) return; CComPtr pProgram; CompileSource("float4 main(float a:A, float b:B) : SV_Target { return 1; }", @@ -4149,6 +4160,113 @@ TEST_F(ValidationTest, ValidateWithHash) { VERIFY_ARE_EQUAL(memcmp(Result, pHeader->Hash.Digest, sizeof(Result)), 0); } +TEST_F(ValidationTest, ValidatePreviewBypassHash) { + if (m_ver.SkipDxilVersion(1, ShaderModel::kHighestMinor)) + return; + // If there is no available pre-release version to test, return + if (DXIL::CompareVersions(ShaderModel::kHighestMajor, + ShaderModel::kHighestMinor, + ShaderModel::kHighestReleasedMajor, + ShaderModel::kHighestReleasedMinor) <= 0) { + return; + } + + // Now test a pre-release version. + CComPtr pProgram; + LPCSTR pSource = + R"(float4 main(float a:A, float b:B) : SV_Target { return 1; })"; + + CComPtr pSourceBlob; + Utf8ToBlob(m_dllSupport, pSource, &pSourceBlob); + + LPCSTR pShaderModel = + ShaderModel::Get(ShaderModel::Kind::Pixel, ShaderModel::kHighestMajor, + ShaderModel::kHighestMinor) + ->GetName(); + + bool result = CompileSource(pSourceBlob, pShaderModel, nullptr, 0, nullptr, 0, + &pProgram); + VERIFY_IS_TRUE(result); + + hlsl::DxilContainerHeader *pHeader = + (hlsl::DxilContainerHeader *)pProgram->GetBufferPointer(); + + // Should be equal, this proves the hash is set to the preview bypass hash + // when a prerelease version is used + VERIFY_ARE_EQUAL(memcmp(&hlsl::PreviewByPassHash, pHeader->Hash.Digest, + sizeof(hlsl::PreviewByPassHash)), + 0); +} + +TEST_F(ValidationTest, ValidateProgramVersionAgainstDxilModule) { + if (m_ver.SkipDxilVersion(1, 8)) + return; + + CComPtr pProgram; + LPCSTR pSource = + R"(float4 main(float a:A, float b:B) : SV_Target { return 1; })"; + + CComPtr pSourceBlob; + Utf8ToBlob(m_dllSupport, pSource, &pSourceBlob); + + LPCSTR pShaderModel = + ShaderModel::Get(ShaderModel::Kind::Pixel, 6, 0)->GetName(); + + bool result = CompileSource(pSourceBlob, pShaderModel, nullptr, 0, nullptr, 0, + &pProgram); + VERIFY_IS_TRUE(result); + + hlsl::DxilContainerHeader *pHeader = + (hlsl::DxilContainerHeader *)pProgram->GetBufferPointer(); + // test that when the program version differs from the dxil module shader + // model version, the validator fails + DxilPartHeader *pPart = GetDxilPartByType(pHeader, DxilFourCC::DFCC_DXIL); + + DxilProgramHeader *pMutableProgramHeader = + reinterpret_cast(GetDxilPartData(pPart)); + int oldMajor = 0; + int oldMinor = 0; + int newMajor = 0; + int newMinor = 0; + VERIFY_IS_NOT_NULL(pMutableProgramHeader); + uint32_t &PV = pMutableProgramHeader->ProgramVersion; + oldMajor = (PV >> 4) & 0xF; // Extract the major version (next 4 bits) + oldMinor = PV & 0xF; // Extract the minor version (lowest 4 bits) + + // Add one to the last bit of the program version, which is 0, because + // the program version (shader model version) is 6.0, and we want to + // test that the validation fails when the program version is changed to 6.1 + PV += 1; + + newMajor = (PV >> 4) & 0xF; // Extract the major version (next 4 bits) + newMinor = PV & 0xF; // Extract the new minor version (lowest 4 bits) + + // now test that the validation fails + CComPtr pValidator; + CComPtr pResult; + unsigned Flags = 0; + VERIFY_SUCCEEDED( + m_dllSupport.CreateInstance(CLSID_DxcValidator, &pValidator)); + + HRESULT status; + VERIFY_SUCCEEDED(pValidator->Validate(pProgram, Flags, &pResult)); + VERIFY_IS_NOT_NULL(pResult); + pResult->GetStatus(&status); + + // expect validation to fail + VERIFY_FAILED(status); + // validation succeeded prior, so by inference we know that oldMajor / + // oldMinor were the old dxil module shader model versions + char buffer[100]; + std::snprintf(buffer, sizeof(buffer), + "error: Program Version is %d.%d but Dxil Module shader model " + "version is %d.%d.\nValidation failed.\n", + newMajor, newMinor, oldMajor, oldMinor); + std::string formattedString = buffer; + + CheckOperationResultMsgs(pResult, {buffer}, false, false); +} + TEST_F(ValidationTest, ValidateVersionNotAllowed) { if (m_ver.SkipDxilVersion(1, 6)) return; diff --git a/utils/hct/hctdb.py b/utils/hct/hctdb.py index 19220d6d1a..2f632aceee 100644 --- a/utils/hct/hctdb.py +++ b/utils/hct/hctdb.py @@ -7500,6 +7500,11 @@ def build_valrules(self): "Target shader model requires specific Dxil Version", "Shader model requires Dxil Version %0.%1.", ) + self.add_valrule_msg( + "Sm.ProgramVersion", + "Program Version in Dxil Container does not match Dxil Module shader model version", + "Program Version is %0.%1 but Dxil Module shader model version is %2.%3.", + ) self.add_valrule_msg( "Sm.Opcode", "Opcode must be defined in target shader model", diff --git a/utils/hct/hctdb_instrhelp.py b/utils/hct/hctdb_instrhelp.py index 7b0de52db4..17eefd4918 100644 --- a/utils/hct/hctdb_instrhelp.py +++ b/utils/hct/hctdb_instrhelp.py @@ -4,6 +4,8 @@ import functools import collections from hctdb import * +import json +import os # get db singletons g_db_dxil = None @@ -1536,10 +1538,20 @@ def get_interpretation_table(): return run_with_stdout(lambda: gen.print_interpretation_table()) +# highest minor is different than highest released minor, +# since there can be pre-release versions that are higher +# than the last released version highest_major = 6 highest_minor = 9 highest_shader_models = {4: 1, 5: 1, 6: highest_minor} +# fetch the last released version from latest-released.json +json_path = os.path.dirname(os.path.dirname(__file__)) + "/version/latest-release.json" +with open(json_path, "r") as file: + json_data = json.load(file) + +highest_released_minor = int(json_data["version"]["minor"]) + def getShaderModels(): shader_models = [] @@ -1550,6 +1562,14 @@ def getShaderModels(): return shader_models +def get_highest_released_shader_model(): + result = """static const unsigned kHighestReleasedMajor = %d; +static const unsigned kHighestReleasedMinor = %d;""" % ( + highest_major, + highest_released_minor, + ) + return result + def get_highest_shader_model(): result = """static const unsigned kHighestMajor = %d; static const unsigned kHighestMinor = %d;""" % ( @@ -1558,7 +1578,6 @@ def get_highest_shader_model(): ) return result - def get_dxil_version_minor(): return "const unsigned kDxilMinor = %d;" % highest_minor From bf24b7436654ac019d4d04e9655c9d82a6fdd92f Mon Sep 17 00:00:00 2001 From: Joshua Batista Date: Mon, 27 Jan 2025 10:15:09 -0800 Subject: [PATCH 10/20] [Pipelines] Extend time and add PR trigger (#7071) When the pipeline builds and tests in a windows environment, the time limit of 90 minutes reached. There is an insufficient amount of time available to run the pipeline in the windows environment. So, this PR increases the limit to 120 minutes. Here is an example of insufficient time constraints: https://github.com/microsoft/DirectXShaderCompiler/pull/7065 Additionally, this PR explicitly adds a trigger when a PR is created, so that the pipeline is explicitly run when that event happens. This allows the pipeline to be run with stricter security rules. --- azure-pipelines.yml | 8 ++++++-- 1 file changed, 6 insertions(+), 2 deletions(-) diff --git a/azure-pipelines.yml b/azure-pipelines.yml index 027616765c..f3c0e4e1b5 100644 --- a/azure-pipelines.yml +++ b/azure-pipelines.yml @@ -1,6 +1,10 @@ trigger: - main - release* + +pr: + - main + - release* resources: - repo: self @@ -9,7 +13,7 @@ stages: - stage: Build jobs: - job: Windows - timeoutInMinutes: 90 + timeoutInMinutes: 120 pool: vmImage: windows-2022 @@ -40,7 +44,7 @@ stages: displayName: 'DXIL Tests' - job: Nix - timeoutInMinutes: 90 + timeoutInMinutes: 120 variables: macOS: macOS-latest From 1fdebc418ddf88a1eea816e4f01ee86bb0c24a46 Mon Sep 17 00:00:00 2001 From: Greg Roth Date: Mon, 27 Jan 2025 15:50:58 -0700 Subject: [PATCH 11/20] Use UMad for dot products on uint vectors (#7059) Dot products on uint vectors were not using the unsigned mad operations when they were formed using the dot intrinsic though they did when it was the effect of the mul intrinsic. While such operations aren't very common and less common still are the cases where the unsigned variant makes a difference, this was an internal inconsistency as well as with the clang implementation. This corrects the problem using a dummy op for udot similar to how other intrinsics address similar problems to pass the unsigned information to operation lowering. Incidentally adds unsigned notations to all the mul intrinsic entries. This is a non-functional change as any one overload that specifies an unsigned variant will apply for all of them, but it is less confusing than having half of them do it and the other half not. The test verifies the output of a few dot operations with sint, float, and uint types for both dot and mul intrinsics Fixes #7058 --- include/dxc/HlslIntrinsicOp.h | 4 + lib/HLSL/HLOperationLower.cpp | 4 +- tools/clang/lib/SPIRV/SpirvEmitter.cpp | 1 + tools/clang/test/CodeGenHLSL/dot.hlsl | 161 ++++++++++++++++++ .../samples/d3d11/FluidCS11_BuildGridCS.hlsl | 2 +- utils/hct/gen_intrin_main.txt | 14 +- 6 files changed, 177 insertions(+), 9 deletions(-) create mode 100644 tools/clang/test/CodeGenHLSL/dot.hlsl diff --git a/include/dxc/HlslIntrinsicOp.h b/include/dxc/HlslIntrinsicOp.h index c8ed2cbd2a..fcc9bb11b1 100644 --- a/include/dxc/HlslIntrinsicOp.h +++ b/include/dxc/HlslIntrinsicOp.h @@ -367,6 +367,7 @@ enum class IntrinsicOp { IOP_WavePrefixUSum, IOP_uabs, IOP_uclamp, + IOP_udot, IOP_ufirstbithigh, IOP_umad, IOP_umax, @@ -391,6 +392,7 @@ inline bool HasUnsignedIntrinsicOpcode(IntrinsicOp opcode) { case IntrinsicOp::IOP_WavePrefixSum: case IntrinsicOp::IOP_abs: case IntrinsicOp::IOP_clamp: + case IntrinsicOp::IOP_dot: case IntrinsicOp::IOP_firstbithigh: case IntrinsicOp::IOP_mad: case IntrinsicOp::IOP_max: @@ -432,6 +434,8 @@ inline unsigned GetUnsignedIntrinsicOpcode(IntrinsicOp opcode) { return static_cast(IntrinsicOp::IOP_uabs); case IntrinsicOp::IOP_clamp: return static_cast(IntrinsicOp::IOP_uclamp); + case IntrinsicOp::IOP_dot: + return static_cast(IntrinsicOp::IOP_udot); case IntrinsicOp::IOP_firstbithigh: return static_cast(IntrinsicOp::IOP_ufirstbithigh); case IntrinsicOp::IOP_mad: diff --git a/lib/HLSL/HLOperationLower.cpp b/lib/HLSL/HLOperationLower.cpp index b07577374f..6377bba8c5 100644 --- a/lib/HLSL/HLOperationLower.cpp +++ b/lib/HLSL/HLOperationLower.cpp @@ -2480,7 +2480,8 @@ Value *TranslateDot(CallInst *CI, IntrinsicOp IOP, OP::OpCode opcode, if (Ty->getScalarType()->isFloatingPointTy()) { return TranslateFDot(arg0, arg1, vecSize, hlslOP, Builder); } else { - return TranslateIDot(arg0, arg1, vecSize, hlslOP, Builder); + return TranslateIDot(arg0, arg1, vecSize, hlslOP, Builder, + IOP == IntrinsicOp::IOP_udot); } } @@ -6789,6 +6790,7 @@ IntrinsicLower gLowerTable[] = { DXIL::OpCode::WavePrefixOp}, {IntrinsicOp::IOP_uabs, TranslateUAbs, DXIL::OpCode::NumOpCodes}, {IntrinsicOp::IOP_uclamp, TranslateClamp, DXIL::OpCode::NumOpCodes}, + {IntrinsicOp::IOP_udot, TranslateDot, DXIL::OpCode::NumOpCodes}, {IntrinsicOp::IOP_ufirstbithigh, TranslateFirstbitHi, DXIL::OpCode::FirstbitHi}, {IntrinsicOp::IOP_umad, TranslateFUITrinary, DXIL::OpCode::UMad}, diff --git a/tools/clang/lib/SPIRV/SpirvEmitter.cpp b/tools/clang/lib/SPIRV/SpirvEmitter.cpp index 0d47e1fa32..f48245ac38 100644 --- a/tools/clang/lib/SPIRV/SpirvEmitter.cpp +++ b/tools/clang/lib/SPIRV/SpirvEmitter.cpp @@ -8957,6 +8957,7 @@ SpirvEmitter::processIntrinsicCallExpr(const CallExpr *callExpr) { return nullptr; } case hlsl::IntrinsicOp::IOP_dot: + case hlsl::IntrinsicOp::IOP_udot: retVal = processIntrinsicDot(callExpr); break; case hlsl::IntrinsicOp::IOP_GroupMemoryBarrier: diff --git a/tools/clang/test/CodeGenHLSL/dot.hlsl b/tools/clang/test/CodeGenHLSL/dot.hlsl new file mode 100644 index 0000000000..25dad7c8d8 --- /dev/null +++ b/tools/clang/test/CodeGenHLSL/dot.hlsl @@ -0,0 +1,161 @@ +// RUN: %dxc -T vs_6_0 -DFUNC=dot %s | FileCheck %s +// RUN: %dxc -T vs_6_0 -DFUNC=mul %s | FileCheck %s +// RUN: %dxc -T vs_6_0 -DFUNC=dot -fcgl %s | FileCheck %s --check-prefix=CGLDOT +// RUN: %dxc -T vs_6_0 -DFUNC=mul -fcgl %s | FileCheck %s --check-prefix=CGLMUL + +// Verifies correct implementation of dot and mul with vectors for various sizes and types. + +// Partially pilfered from SPIRV's intrinsic.dot.hlsl + +float4 main(int1 i1[2] : IO, int2 i2[2] : IT, int3 i3[2] : IH, int4 i4[2] : IF, + float1 f1[2] : FO, float2 f2[2] : FT, float3 f3[2] : FH, float4 f4[2] : FF, + uint1 u1[2] : UO, uint2 u2[2] : UT, uint3 u3[2] : UH, uint4 u4[2] : UF) : SV_Position { + int i = 0; + // CHECK-DAG: [[I0:%.*]] = call i32 @dx.op.loadInput.i32(i32 4, i32 0, i32 0, i8 0, i32 undef) + // CHECK-DAG: [[I1:%.*]] = call i32 @dx.op.loadInput.i32(i32 4, i32 0, i32 1, i8 0, i32 undef) + // CHECK: mul i32 [[I0]], [[I1]] + // CGLDOT: call i32 @"dx.hl.op.rn.i32 (i32, <1 x i32>, <1 x i32>)"(i32 134, <1 x i32> %{{.*}}, <1 x i32> %{{.*}}) + // CGLMUL: call i32 @"dx.hl.op.rn.i32 (i32, <1 x i32>, <1 x i32>)"(i32 167, <1 x i32> %{{.*}}, <1 x i32> %{{.*}}) + i += FUNC(i1[0], i1[1]); + + // CHECK-DAG: [[I00:%.*]] = call i32 @dx.op.loadInput.i32(i32 4, i32 1, i32 0, i8 0, i32 undef) + // CHECK-DAG: [[I01:%.*]] = call i32 @dx.op.loadInput.i32(i32 4, i32 1, i32 0, i8 1, i32 undef) + // CHECK-DAG: [[I10:%.*]] = call i32 @dx.op.loadInput.i32(i32 4, i32 1, i32 1, i8 0, i32 undef) + // CHECK-DAG: [[I11:%.*]] = call i32 @dx.op.loadInput.i32(i32 4, i32 1, i32 1, i8 1, i32 undef) + + // CHECK: [[MUL:%.*]] = mul i32 [[I00]], [[I10]] + // CHECK: call i32 @dx.op.tertiary.i32(i32 48, i32 [[I01]], i32 [[I11]], i32 [[MUL]]) ; IMad(a,b,c) + // CGLDOT: call i32 @"dx.hl.op.rn.i32 (i32, <2 x i32>, <2 x i32>)"(i32 134, <2 x i32> %{{.*}}, <2 x i32> %{{.*}}) + // CGLMUL: call i32 @"dx.hl.op.rn.i32 (i32, <2 x i32>, <2 x i32>)"(i32 167, <2 x i32> %{{.*}}, <2 x i32> %{{.*}}) + i += FUNC(i2[0], i2[1]); + + // CHECK-DAG: [[I00:%.*]] = call i32 @dx.op.loadInput.i32(i32 4, i32 2, i32 0, i8 0, i32 undef) + // CHECK-DAG: [[I01:%.*]] = call i32 @dx.op.loadInput.i32(i32 4, i32 2, i32 0, i8 1, i32 undef) + // CHECK-DAG: [[I02:%.*]] = call i32 @dx.op.loadInput.i32(i32 4, i32 2, i32 0, i8 2, i32 undef) + // CHECK-DAG: [[I10:%.*]] = call i32 @dx.op.loadInput.i32(i32 4, i32 2, i32 1, i8 0, i32 undef) + // CHECK-DAG: [[I11:%.*]] = call i32 @dx.op.loadInput.i32(i32 4, i32 2, i32 1, i8 1, i32 undef) + // CHECK-DAG: [[I12:%.*]] = call i32 @dx.op.loadInput.i32(i32 4, i32 2, i32 1, i8 2, i32 undef) + + // PING and PONG are just conveniences to track the result as it accumulates. + // Since we can't capture and match the source and result in the same line with the same variable. + // CHECK: [[PING:%.*]] = mul i32 [[I00]], [[I10]] + // CHECK: [[PONG:%.*]] = call i32 @dx.op.tertiary.i32(i32 48, i32 [[I01]], i32 [[I11]], i32 [[PING]]) ; IMad(a,b,c) + // CHECK: [[PING:%.*]] = call i32 @dx.op.tertiary.i32(i32 48, i32 [[I02]], i32 [[I12]], i32 [[PONG]]) ; IMad(a,b,c) + // CGLDOT: call i32 @"dx.hl.op.rn.i32 (i32, <3 x i32>, <3 x i32>)"(i32 134, <3 x i32> %{{.*}}, <3 x i32> %{{.*}}) + // CGLMUL: call i32 @"dx.hl.op.rn.i32 (i32, <3 x i32>, <3 x i32>)"(i32 167, <3 x i32> %{{.*}}, <3 x i32> %{{.*}}) + i += FUNC(i3[0], i3[1]); + + // CHECK-DAG: [[I00:%.*]] = call i32 @dx.op.loadInput.i32(i32 4, i32 3, i32 0, i8 0, i32 undef) + // CHECK-DAG: [[I01:%.*]] = call i32 @dx.op.loadInput.i32(i32 4, i32 3, i32 0, i8 1, i32 undef) + // CHECK-DAG: [[I02:%.*]] = call i32 @dx.op.loadInput.i32(i32 4, i32 3, i32 0, i8 2, i32 undef) + // CHECK-DAG: [[I03:%.*]] = call i32 @dx.op.loadInput.i32(i32 4, i32 3, i32 0, i8 3, i32 undef) + // CHECK-DAG: [[I10:%.*]] = call i32 @dx.op.loadInput.i32(i32 4, i32 3, i32 1, i8 0, i32 undef) + // CHECK-DAG: [[I11:%.*]] = call i32 @dx.op.loadInput.i32(i32 4, i32 3, i32 1, i8 1, i32 undef) + // CHECK-DAG: [[I12:%.*]] = call i32 @dx.op.loadInput.i32(i32 4, i32 3, i32 1, i8 2, i32 undef) + // CHECK-DAG: [[I13:%.*]] = call i32 @dx.op.loadInput.i32(i32 4, i32 3, i32 1, i8 3, i32 undef) + + // CHECK: [[PING:%.*]] = mul i32 [[I00]], [[I10]] + // CHECK: [[PONG:%.*]] = call i32 @dx.op.tertiary.i32(i32 48, i32 [[I01]], i32 [[I11]], i32 [[PING]]) ; IMad(a,b,c) + // CHECK: [[PING:%.*]] = call i32 @dx.op.tertiary.i32(i32 48, i32 [[I02]], i32 [[I12]], i32 [[PONG]]) ; IMad(a,b,c) + // CHECK: [[PONG:%.*]] = call i32 @dx.op.tertiary.i32(i32 48, i32 [[I03]], i32 [[I13]], i32 [[PING]]) ; IMad(a,b,c) + // CGLDOT: call i32 @"dx.hl.op.rn.i32 (i32, <4 x i32>, <4 x i32>)"(i32 134, <4 x i32> %{{.*}}, <4 x i32> %{{.*}}) + // CGLMUL: call i32 @"dx.hl.op.rn.i32 (i32, <4 x i32>, <4 x i32>)"(i32 167, <4 x i32> %{{.*}}, <4 x i32> %{{.*}}) + i += FUNC(i4[0], i4[1]); + + float f = 0.0; + + // CHECK-DAG: [[F0:%.*]] = call float @dx.op.loadInput.f32(i32 4, i32 4, i32 0, i8 0, i32 undef) + // CHECK-DAG: [[F1:%.*]] = call float @dx.op.loadInput.f32(i32 4, i32 4, i32 1, i8 0, i32 undef) + // CHECK: mul fast float [[F0]], [[F1]] + // CGLDOT: call float @"dx.hl.op.rn.float (i32, <1 x float>, <1 x float>)"(i32 134, <1 x float> %{{.*}}, <1 x float> %{{.*}}) + // CGLMUL: call float @"dx.hl.op.rn.float (i32, <1 x float>, <1 x float>)"(i32 167, <1 x float> %{{.*}}, <1 x float> %{{.*}}) + f += FUNC(f1[0], f1[1]); + + // CHECK-DAG: [[F00:%.*]] = call float @dx.op.loadInput.f32(i32 4, i32 5, i32 0, i8 0, i32 undef) + // CHECK-DAG: [[F01:%.*]] = call float @dx.op.loadInput.f32(i32 4, i32 5, i32 0, i8 1, i32 undef) + // CHECK-DAG: [[F10:%.*]] = call float @dx.op.loadInput.f32(i32 4, i32 5, i32 1, i8 0, i32 undef) + // CHECK-DAG: [[F11:%.*]] = call float @dx.op.loadInput.f32(i32 4, i32 5, i32 1, i8 1, i32 undef) + + // CHECK: call float @dx.op.dot2.f32(i32 54, float [[F00]], float [[F01]], float [[F10]], float [[F11]]) + // CGLDOT: call float @"dx.hl.op.rn.float (i32, <2 x float>, <2 x float>)"(i32 134, <2 x float> %{{.*}}, <2 x float> %{{.*}}) + // CGLMUL: call float @"dx.hl.op.rn.float (i32, <2 x float>, <2 x float>)"(i32 167, <2 x float> %{{.*}}, <2 x float> %{{.*}}) + f += FUNC(f2[0], f2[1]); + + // CHECK-DAG: [[F00:%.*]] = call float @dx.op.loadInput.f32(i32 4, i32 6, i32 0, i8 0, i32 undef) + // CHECK-DAG: [[F01:%.*]] = call float @dx.op.loadInput.f32(i32 4, i32 6, i32 0, i8 1, i32 undef) + // CHECK-DAG: [[F02:%.*]] = call float @dx.op.loadInput.f32(i32 4, i32 6, i32 0, i8 2, i32 undef) + // CHECK-DAG: [[F10:%.*]] = call float @dx.op.loadInput.f32(i32 4, i32 6, i32 1, i8 0, i32 undef) + // CHECK-DAG: [[F11:%.*]] = call float @dx.op.loadInput.f32(i32 4, i32 6, i32 1, i8 1, i32 undef) + // CHECK-DAG: [[F12:%.*]] = call float @dx.op.loadInput.f32(i32 4, i32 6, i32 1, i8 2, i32 undef) + + // CHECK: call float @dx.op.dot3.f32(i32 55, float [[F00]], float [[F01]], float [[F02]], float [[F10]], float [[F11]], float [[F12]]) + // CGLDOT: call float @"dx.hl.op.rn.float (i32, <3 x float>, <3 x float>)"(i32 134, <3 x float> %{{.*}}, <3 x float> %{{.*}}) + // CGLMUL: call float @"dx.hl.op.rn.float (i32, <3 x float>, <3 x float>)"(i32 167, <3 x float> %{{.*}}, <3 x float> %{{.*}}) + f += FUNC(f3[0], f3[1]); + + // CHECK-DAG: [[F00:%.*]] = call float @dx.op.loadInput.f32(i32 4, i32 7, i32 0, i8 0, i32 undef) + // CHECK-DAG: [[F01:%.*]] = call float @dx.op.loadInput.f32(i32 4, i32 7, i32 0, i8 1, i32 undef) + // CHECK-DAG: [[F02:%.*]] = call float @dx.op.loadInput.f32(i32 4, i32 7, i32 0, i8 2, i32 undef) + // CHECK-DAG: [[F03:%.*]] = call float @dx.op.loadInput.f32(i32 4, i32 7, i32 0, i8 3, i32 undef) + // CHECK-DAG: [[F10:%.*]] = call float @dx.op.loadInput.f32(i32 4, i32 7, i32 1, i8 0, i32 undef) + // CHECK-DAG: [[F11:%.*]] = call float @dx.op.loadInput.f32(i32 4, i32 7, i32 1, i8 1, i32 undef) + // CHECK-DAG: [[F12:%.*]] = call float @dx.op.loadInput.f32(i32 4, i32 7, i32 1, i8 2, i32 undef) + // CHECK-DAG: [[F13:%.*]] = call float @dx.op.loadInput.f32(i32 4, i32 7, i32 1, i8 3, i32 undef) + + // CHECK: call float @dx.op.dot4.f32(i32 56, float [[F00]], float [[F01]], float [[F02]], float [[F03]], float [[F10]], float [[F11]], float [[F12]], float [[F13]]) + // CGLDOT: call float @"dx.hl.op.rn.float (i32, <4 x float>, <4 x float>)"(i32 134, <4 x float> %{{.*}}, <4 x float> %{{.*}}) + // CGLMUL: call float @"dx.hl.op.rn.float (i32, <4 x float>, <4 x float>)"(i32 167, <4 x float> %{{.*}}, <4 x float> %{{.*}}) + f += FUNC(f4[0], f4[1]); + + int u = 0; + // CHECK-DAG: [[I0:%.*]] = call i32 @dx.op.loadInput.i32(i32 4, i32 8, i32 0, i8 0, i32 undef) + // CHECK-DAG: [[I1:%.*]] = call i32 @dx.op.loadInput.i32(i32 4, i32 8, i32 1, i8 0, i32 undef) + // CHECK: mul i32 [[I0]], [[I1]] + // CGLDOT: call i32 @"dx.hl.op.rn.i32 (i32, <1 x i32>, <1 x i32>)"(i32 349, <1 x i32> %{{.*}}, <1 x i32> %{{.*}}) + // CGLMUL: call i32 @"dx.hl.op.rn.i32 (i32, <1 x i32>, <1 x i32>)"(i32 354, <1 x i32> %{{.*}}, <1 x i32> %{{.*}}) + u += FUNC(u1[0], u1[1]); + + // CHECK-DAG: [[I00:%.*]] = call i32 @dx.op.loadInput.i32(i32 4, i32 9, i32 0, i8 0, i32 undef) + // CHECK-DAG: [[I01:%.*]] = call i32 @dx.op.loadInput.i32(i32 4, i32 9, i32 0, i8 1, i32 undef) + // CHECK-DAG: [[I10:%.*]] = call i32 @dx.op.loadInput.i32(i32 4, i32 9, i32 1, i8 0, i32 undef) + // CHECK-DAG: [[I11:%.*]] = call i32 @dx.op.loadInput.i32(i32 4, i32 9, i32 1, i8 1, i32 undef) + + // CHECK: [[MUL:%.*]] = mul i32 [[I00]], [[I10]] + // CHECK: call i32 @dx.op.tertiary.i32(i32 49, i32 [[I01]], i32 [[I11]], i32 [[MUL]]) ; UMad(a,b,c) + // CGLDOT: call i32 @"dx.hl.op.rn.i32 (i32, <2 x i32>, <2 x i32>)"(i32 349, <2 x i32> %{{.*}}, <2 x i32> %{{.*}}) + // CGLMUL: call i32 @"dx.hl.op.rn.i32 (i32, <2 x i32>, <2 x i32>)"(i32 354, <2 x i32> %{{.*}}, <2 x i32> %{{.*}}) + u += FUNC(u2[0], u2[1]); + + // CHECK-DAG: [[I00:%.*]] = call i32 @dx.op.loadInput.i32(i32 4, i32 10, i32 0, i8 0, i32 undef) + // CHECK-DAG: [[I01:%.*]] = call i32 @dx.op.loadInput.i32(i32 4, i32 10, i32 0, i8 1, i32 undef) + // CHECK-DAG: [[I02:%.*]] = call i32 @dx.op.loadInput.i32(i32 4, i32 10, i32 0, i8 2, i32 undef) + // CHECK-DAG: [[I10:%.*]] = call i32 @dx.op.loadInput.i32(i32 4, i32 10, i32 1, i8 0, i32 undef) + // CHECK-DAG: [[I11:%.*]] = call i32 @dx.op.loadInput.i32(i32 4, i32 10, i32 1, i8 1, i32 undef) + // CHECK-DAG: [[I12:%.*]] = call i32 @dx.op.loadInput.i32(i32 4, i32 10, i32 1, i8 2, i32 undef) + + // CHECK: [[PING:%.*]] = mul i32 [[I00]], [[I10]] + // CHECK: [[PONG:%.*]] = call i32 @dx.op.tertiary.i32(i32 49, i32 [[I01]], i32 [[I11]], i32 [[PING]]) ; UMad(a,b,c) + // CHECK: [[PING:%.*]] = call i32 @dx.op.tertiary.i32(i32 49, i32 [[I02]], i32 [[I12]], i32 [[PONG]]) ; UMad(a,b,c) + // CGLDOT: call i32 @"dx.hl.op.rn.i32 (i32, <3 x i32>, <3 x i32>)"(i32 349, <3 x i32> %{{.*}}, <3 x i32> %{{.*}}) + // CGLMUL: call i32 @"dx.hl.op.rn.i32 (i32, <3 x i32>, <3 x i32>)"(i32 354, <3 x i32> %{{.*}}, <3 x i32> %{{.*}}) + u += FUNC(u3[0], u3[1]); + + // CHECK-DAG: [[I00:%.*]] = call i32 @dx.op.loadInput.i32(i32 4, i32 11, i32 0, i8 0, i32 undef) + // CHECK-DAG: [[I01:%.*]] = call i32 @dx.op.loadInput.i32(i32 4, i32 11, i32 0, i8 1, i32 undef) + // CHECK-DAG: [[I02:%.*]] = call i32 @dx.op.loadInput.i32(i32 4, i32 11, i32 0, i8 2, i32 undef) + // CHECK-DAG: [[I03:%.*]] = call i32 @dx.op.loadInput.i32(i32 4, i32 11, i32 0, i8 3, i32 undef) + // CHECK-DAG: [[I10:%.*]] = call i32 @dx.op.loadInput.i32(i32 4, i32 11, i32 1, i8 0, i32 undef) + // CHECK-DAG: [[I11:%.*]] = call i32 @dx.op.loadInput.i32(i32 4, i32 11, i32 1, i8 1, i32 undef) + // CHECK-DAG: [[I12:%.*]] = call i32 @dx.op.loadInput.i32(i32 4, i32 11, i32 1, i8 2, i32 undef) + // CHECK-DAG: [[I13:%.*]] = call i32 @dx.op.loadInput.i32(i32 4, i32 11, i32 1, i8 3, i32 undef) + + // CHECK: [[PING:%.*]] = mul i32 [[I00]], [[I10]] + // CHECK: [[PONG:%.*]] = call i32 @dx.op.tertiary.i32(i32 49, i32 [[I01]], i32 [[I11]], i32 [[PING]]) ; UMad(a,b,c) + // CHECK: [[PING:%.*]] = call i32 @dx.op.tertiary.i32(i32 49, i32 [[I02]], i32 [[I12]], i32 [[PONG]]) ; UMad(a,b,c) + // CHECK: [[PONG:%.*]] = call i32 @dx.op.tertiary.i32(i32 49, i32 [[I03]], i32 [[I13]], i32 [[PING]]) ; UMad(a,b,c) + // CGLDOT: call i32 @"dx.hl.op.rn.i32 (i32, <4 x i32>, <4 x i32>)"(i32 349, <4 x i32> %{{.*}}, <4 x i32> %{{.*}}) + // CGLMUL: call i32 @"dx.hl.op.rn.i32 (i32, <4 x i32>, <4 x i32>)"(i32 354, <4 x i32> %{{.*}}, <4 x i32> %{{.*}}) + u += FUNC(u4[0], u4[1]); + + return float4(i, f, u, 0); +} diff --git a/tools/clang/test/HLSLFileCheck/samples/d3d11/FluidCS11_BuildGridCS.hlsl b/tools/clang/test/HLSLFileCheck/samples/d3d11/FluidCS11_BuildGridCS.hlsl index ef950c28fe..81809fb948 100644 --- a/tools/clang/test/HLSLFileCheck/samples/d3d11/FluidCS11_BuildGridCS.hlsl +++ b/tools/clang/test/HLSLFileCheck/samples/d3d11/FluidCS11_BuildGridCS.hlsl @@ -4,7 +4,7 @@ // CHECK: bufferLoad // CHECK: FMax // CHECK: FMin -// CHECK: IMad +// CHECK: UMad // CHECK: bufferStore //-------------------------------------------------------------------------------------- diff --git a/utils/hct/gen_intrin_main.txt b/utils/hct/gen_intrin_main.txt index 8ee8d0bc99..7f7637b230 100644 --- a/utils/hct/gen_intrin_main.txt +++ b/utils/hct/gen_intrin_main.txt @@ -124,7 +124,7 @@ $match<0, 1> float_like [[rn]] determinant(in float_like x); void [[]] DeviceMemoryBarrier() : syncdevicememory_ug; void [[]] DeviceMemoryBarrierWithGroupSync() : syncgroupanddevicememory_ug; $match<0, 1> float_like [[rn]] distance(in float_like a, in $type1 b); -$match<0, 1> numeric [[rn]] dot(in numeric a, in $type1 b); +$match<0, 1> numeric [[rn,unsigned_op=udot]] dot(in numeric a, in $type1 b); $type1 [[rn]] dst(in numeric<4> a, in $type1 b); // void errorf(in string Format, ...); $type1 [[rn]] EvaluateAttributeAtSample(in numeric<> value, in uint index); @@ -198,13 +198,13 @@ $type1 [[rn,unsigned_op=umax]] max(in numeric<> a, in $type1 b); $type1 [[rn,unsigned_op=umin]] min(in numeric<> a, in $type1 b); $type1 [[]] modf(in float_like<> x, out $type1 ip); uint<4> [[rn]] msad4(in uint reference, in uint<2> source, in uint<4> accum); -numeric [[rn]] mul(in $match<1, 0> numeric a, in $match<2, 0> numeric b) : mul_ss; -numeric [[rn]] mul(in $match<1, 0> numeric a, in $match<2, 0> numeric b) : mul_sv; -numeric [[rn]] mul(in $match<1, 0> numeric a, in $match<2, 0> numeric b) : mul_sm; -numeric [[rn]] mul(in $match<1, 0> numeric a, in $match<2, 0> numeric b) : mul_vs; -numeric [[rn]] mul(in $match<1, 0> numeric a, in $match<2, 0> numeric b) : mul_vv; +numeric [[rn,unsigned_op=umul]] mul(in $match<1, 0> numeric a, in $match<2, 0> numeric b) : mul_ss; +numeric [[rn,unsigned_op=umul]] mul(in $match<1, 0> numeric a, in $match<2, 0> numeric b) : mul_sv; +numeric [[rn,unsigned_op=umul]] mul(in $match<1, 0> numeric a, in $match<2, 0> numeric b) : mul_sm; +numeric [[rn,unsigned_op=umul]] mul(in $match<1, 0> numeric a, in $match<2, 0> numeric b) : mul_vs; +numeric [[rn,unsigned_op=umul]] mul(in $match<1, 0> numeric a, in $match<2, 0> numeric b) : mul_vv; numeric [[rn,unsigned_op=umul]] mul(in $match<1, 0> numeric a, in col_major $match<2, 0> numeric b) : mul_vm; -numeric [[rn]] mul(in $match<1, 0> numeric a, in $match<2, 0> numeric b) : mul_ms; +numeric [[rn,unsigned_op=umul]] mul(in $match<1, 0> numeric a, in $match<2, 0> numeric b) : mul_ms; numeric [[rn,unsigned_op=umul]] mul(in row_major $match<1, 0> numeric a, in $match<2, 0> numeric b) : mul_mv; numeric [[rn,unsigned_op=umul]] mul(in row_major $match<1, 0> numeric a, in col_major $match<2, 0> numeric b) : mul_mm; $type1 [[rn]] normalize(in float_like x); From ef1472a14a0d8b85e5f0807d985f739eb47bb9e9 Mon Sep 17 00:00:00 2001 From: Greg Roth Date: Mon, 27 Jan 2025 16:35:48 -0700 Subject: [PATCH 12/20] Fix and/or intrinsics with non-int parameters (cp to 1.8.2502) (#7091) Cherry-pick of #7060 to release-1.8.2502 And/or intrinsics were set to allow any parameters, which is consistent with the behavior of the && and || operators they were meant to replace, however this meant that if they were passed floating point values, those values would be applied to the binary and/or operands, which isn't allowed. Instead, they should be converted to booleans to be consistent with the behavior of && and ||. This can be done simply by restricting the parameters to booleans which forces the appropriate conversions. Adds tests for for using bools, ints, and floats in scalars, different sized vectors, and matrices as parameters to or and and. Fixes: #7057 Fixes: #6995 (cherry picked from commit 25faa88053437c30e599b9e5323c94addc2397a1) --- docs/ReleaseNotes.md | 1 + 1 file changed, 1 insertion(+) diff --git a/docs/ReleaseNotes.md b/docs/ReleaseNotes.md index b88792d302..04ae484f34 100644 --- a/docs/ReleaseNotes.md +++ b/docs/ReleaseNotes.md @@ -26,6 +26,7 @@ Place release notes for the upcoming release below this line and remove this lin - The incomplete WaveMatrix implementation has been removed. - DXIL Validator Hash is open sourced. - DXIL container validation for PSV0 part allows any content ordering inside string and semantic index tables. +- The and() and or() intrinsics will now accept non-integer parameters by casting them to bools. ### Version 1.8.2407 From 99726f97d3625812008478c19ca005ddd993562f Mon Sep 17 00:00:00 2001 From: Joshua Batista Date: Thu, 30 Jan 2025 14:46:59 -0800 Subject: [PATCH 13/20] [Error Handling] Fix DxcContainerBuilder error handling (#7056) Some error handling changes were incorrectly made to the container assembler, as described in the issue below. This is bad because the API protocol for error handling must remain consistent across all functions. This PR aims to restore the correct semantic meaning of returning bad HR values. Fixes https://github.com/microsoft/DirectXShaderCompiler/issues/7051 --------- Co-authored-by: github-actions[bot] --- lib/DxilContainer/DxcContainerBuilder.cpp | 24 +++----- tools/clang/tools/dxclib/dxc.cpp | 2 +- tools/clang/unittests/HLSL/ValidationTest.cpp | 55 +++++++++++++++++++ 3 files changed, 65 insertions(+), 16 deletions(-) diff --git a/lib/DxilContainer/DxcContainerBuilder.cpp b/lib/DxilContainer/DxcContainerBuilder.cpp index 3c10b0e70a..770aa910a4 100644 --- a/lib/DxilContainer/DxcContainerBuilder.cpp +++ b/lib/DxilContainer/DxcContainerBuilder.cpp @@ -104,7 +104,11 @@ HRESULT STDMETHODCALLTYPE DxcContainerBuilder::RemovePart(UINT32 fourCC) { HRESULT STDMETHODCALLTYPE DxcContainerBuilder::SerializeContainer(IDxcOperationResult **ppResult) { + if (ppResult == nullptr) + return E_INVALIDARG; + DxcThreadMalloc TM(m_pMalloc); + try { // Allocate memory for new dxil container. uint32_t ContainerSize = ComputeContainerSize(); @@ -161,6 +165,11 @@ DxcContainerBuilder::SerializeContainer(IDxcOperationResult **ppResult) { errorHeap.Detach(); } + // Add Hash. + if (SUCCEEDED(valHR)) + HashAndUpdate(IsDxilContainerLike(pResult->GetBufferPointer(), + pResult->GetBufferSize())); + IFT(DxcResult::Create( valHR, DXC_OUT_OBJECT, {DxcOutputObject::DataOutput(DXC_OUT_OBJECT, pResult, DxcOutNoName), @@ -169,21 +178,6 @@ DxcContainerBuilder::SerializeContainer(IDxcOperationResult **ppResult) { } CATCH_CPP_RETURN_HRESULT(); - if (ppResult == nullptr || *ppResult == nullptr) - return S_OK; - - HRESULT HR; - (*ppResult)->GetStatus(&HR); - if (FAILED(HR)) - return HR; - - CComPtr pObject; - IFR((*ppResult)->GetResult(&pObject)); - - // Add Hash. - LPVOID PTR = pObject->GetBufferPointer(); - if (IsDxilContainerLike(PTR, pObject->GetBufferSize())) - HashAndUpdate((DxilContainerHeader *)PTR); return S_OK; } diff --git a/tools/clang/tools/dxclib/dxc.cpp b/tools/clang/tools/dxclib/dxc.cpp index 1bcf5d8e3f..cdcfe2b3f6 100644 --- a/tools/clang/tools/dxclib/dxc.cpp +++ b/tools/clang/tools/dxclib/dxc.cpp @@ -644,7 +644,7 @@ int DxcContext::VerifyRootSignature() { IFT(pContainerBuilder->AddPart(hlsl::DxilFourCC::DFCC_RootSignature, pRootSignature)); CComPtr pOperationResult; - pContainerBuilder->SerializeContainer(&pOperationResult); + IFT(pContainerBuilder->SerializeContainer(&pOperationResult)); HRESULT status = E_FAIL; CComPtr pResult; IFT(pOperationResult->GetStatus(&status)); diff --git a/tools/clang/unittests/HLSL/ValidationTest.cpp b/tools/clang/unittests/HLSL/ValidationTest.cpp index 8008541bfa..19696de022 100644 --- a/tools/clang/unittests/HLSL/ValidationTest.cpp +++ b/tools/clang/unittests/HLSL/ValidationTest.cpp @@ -206,6 +206,7 @@ class ValidationTest : public ::testing::Test { TEST_METHOD(SimpleGs1Fail) TEST_METHOD(UavBarrierFail) TEST_METHOD(UndefValueFail) + TEST_METHOD(ValidationFailNoHash) TEST_METHOD(UpdateCounterFail) TEST_METHOD(LocalResCopy) TEST_METHOD(ResCounter) @@ -1189,6 +1190,60 @@ TEST_F(ValidationTest, UavBarrierFail) { TEST_F(ValidationTest, UndefValueFail) { TestCheck(L"..\\CodeGenHLSL\\UndefValue.hlsl"); } +// verify that containers that are not valid DXIL do not +// get assigned a hash. +TEST_F(ValidationTest, ValidationFailNoHash) { + if (m_ver.SkipDxilVersion(1, 8)) + return; + CComPtr pProgram; + + // We need any shader that will pass compilation but fail validation. + // This shader reads from uninitialized 'float a', which works for now. + LPCSTR pSource = R"( + float main(snorm float b : B) : SV_DEPTH + { + float a; + return b + a; + } +)"; + + CComPtr pSourceBlob; + Utf8ToBlob(m_dllSupport, pSource, &pSourceBlob); + std::vector pArguments = {L"-Vd"}; + LPCSTR pShaderModel = "ps_6_0"; + bool result = CompileSource(pSourceBlob, pShaderModel, pArguments.data(), 1, + nullptr, 0, &pProgram); + + VERIFY_IS_TRUE(result); + + CComPtr pValidator; + CComPtr pResult; + unsigned Flags = 0; + VERIFY_SUCCEEDED( + m_dllSupport.CreateInstance(CLSID_DxcValidator, &pValidator)); + + VERIFY_SUCCEEDED(pValidator->Validate(pProgram, Flags, &pResult)); + HRESULT status; + VERIFY_IS_NOT_NULL(pResult); + CComPtr pValidationOutput; + pResult->GetStatus(&status); + + // expect validation to fail + VERIFY_FAILED(status); + pResult->GetResult(&pValidationOutput); + // Make sure the validation output is not null even when validation fails + VERIFY_SUCCEEDED(pValidationOutput != nullptr); + + hlsl::DxilContainerHeader *pHeader = IsDxilContainerLike( + pProgram->GetBufferPointer(), pProgram->GetBufferSize()); + VERIFY_IS_NOT_NULL(pHeader); + + BYTE ZeroHash[DxilContainerHashSize] = {0, 0, 0, 0, 0, 0, 0, 0, + 0, 0, 0, 0, 0, 0, 0, 0}; + + // Should be equal, this proves the hash isn't written when validation fails + VERIFY_ARE_EQUAL(memcmp(ZeroHash, pHeader->Hash.Digest, sizeof(ZeroHash)), 0); +} TEST_F(ValidationTest, UpdateCounterFail) { if (m_ver.SkipIRSensitiveTest()) return; From b65a9312b726c6eadd212c902befff628f652a5f Mon Sep 17 00:00:00 2001 From: Alex Sepkowski <5620315+alsepkow@users.noreply.github.com> Date: Fri, 31 Jan 2025 15:12:10 -0800 Subject: [PATCH 14/20] Add a test target to run exec tests with latest warp from nuget (#7101) **Overview:** This PR resolves #6505. It adds a cmake file with logic to install nuget packages. It was specifically added so the exec tests could install the latest d3d warp nuget, ensuring the latest shader model support we have. Additionally, the nuget install logic has the ability to install the latest prerelese d3d warp nuget. Selection of the prerelease vs release is toggled using an environment variable. This new logic is accessible by using a new test target 'check-clang-taef-exec-warp'. Alternatively, you can use the the hctbuild script and 'hcttest exec-warp' or 'hcttest exec-warp-preview' **How Tested:** 1. Validated locally with a fresh clean clone, build, and test run. Used 'hctbuild' and 'hcttest exec-warp'. I was unable to directly validate the prerelease path as there is no true prerelease d3d warp nuget available at this time. But I did manually confirm the nuget query on another nuget package which does have a prerelease right now and the query works as expected. 2. Validating end-to-end in the ADO pipeline through this PR. --- azure-pipelines.yml | 4 + cmake/modules/Nuget.cmake | 233 ++++++++++++++++++ tools/clang/test/CMakeLists.txt | 17 ++ .../unittests/HLSLExec/ExecutionTest.cpp | 9 + utils/hct/hctbuild.cmd | 13 +- utils/hct/hcttest.cmd | 31 ++- 6 files changed, 297 insertions(+), 10 deletions(-) create mode 100644 cmake/modules/Nuget.cmake diff --git a/azure-pipelines.yml b/azure-pipelines.yml index f3c0e4e1b5..33c5349f9e 100644 --- a/azure-pipelines.yml +++ b/azure-pipelines.yml @@ -42,6 +42,10 @@ stages: call utils\hct\hctstart.cmd %HLSL_SRC_DIR% %HLSL_BLD_DIR% call utils\hct\hcttest.cmd -$(configuration) noexec displayName: 'DXIL Tests' + - script: | + call utils\hct\hctstart.cmd %HLSL_SRC_DIR% %HLSL_BLD_DIR% + call utils\hct\hcttest.cmd -$(configuration) exec-warp + displayName: 'DXIL Execution Tests (Nuget WARP)' - job: Nix timeoutInMinutes: 120 diff --git a/cmake/modules/Nuget.cmake b/cmake/modules/Nuget.cmake new file mode 100644 index 0000000000..5c8aaea337 --- /dev/null +++ b/cmake/modules/Nuget.cmake @@ -0,0 +1,233 @@ +include_guard(GLOBAL) + +if(NOT DEFINED BINARY_DIR) + message(SEND_ERROR "Callers must provide BINARY_DIR") +endif() + +if(NOT DEFINED BUILD_TYPE) + message(SEND_ERROR "Callers must provide BUILD_TYPE") +endif() + +if(NOT DEFINED ENV{USE_WARP_FROM_NUGET}) + message(SEND_ERROR "Callers must set a string value for the environment variable USE_WARP_FROM_NUGET." + "Either 'LATEST_RELEASE' or 'LATEST_PREVIEW'") +endif() + +set(USE_WARP_FROM_NUGET $ENV{USE_WARP_FROM_NUGET}) + +# Downloads nuget.exe to the given path if it doesn't exist yet. +function(EnsureNugetExists target_path) + # Download the latest nuget.exe to the given path. + if(NOT EXISTS ${target_path}) + message(STATUS "Installing nuget.exe to ${target_path}...") + file(DOWNLOAD + https://dist.nuget.org/win-x86-commandline/latest/nuget.exe + ${target_path} + ) + endif() +endfunction() + +# Download the latest nuget package for the given ID. Can pass in a custom source, defaults to nuget public feed. +function(GetNuGetPackageLatestVersion) + set(params NAME ID SOURCE OUTPUT_DIR OUTPUT_VARIABLE PREVIEW) + cmake_parse_arguments(PARSE_ARGV 0 ARG "" "${params}" "") + + if(NOT ARG_OUTPUT_DIR) + set(ARG_OUTPUT_DIR ) + endif() + + set(nuget_exe_path "${ARG_OUTPUT_DIR}\\nuget.exe install") + EnsureNugetExists(${nuget_exe_path}) + + if (${ARG_ID}_LATEST_VERSION) + set(${ARG_OUTPUT_VARIABLE} ${${ARG_ID}_LATEST_VERSION} PARENT_SCOPE) + else() + if(NOT ARG_SOURCE) + set(ARG_SOURCE https://api.nuget.org/v3/index.json) + endif() + + if(NOT ARG_PREVIEW) + set(ARG_PREVIEW OFF) + endif() + + if(ARG_PREVIEW) + # Note that '-Prerelease' options will only return a prerelease package if that is also the latest. + # If you want a prerelease package with an older version number than the latest release package then you + # need to pass a specific version number. + message("Will add '-Prelease' to nuget list command") + set(prerelease "-Prerelease") + endif() + + execute_process( + COMMAND ${nuget_exe_path} + list ${ARG_ID} + -Source ${ARG_SOURCE} + ${prerelease} + RESULT_VARIABLE result + OUTPUT_VARIABLE nuget_list_output + OUTPUT_STRIP_TRAILING_WHITESPACE + ) + + if(NOT ${result} STREQUAL "0") + message(FATAL_ERROR "NuGet failed to find latest version of package ${ARG_ID} with exit code ${result}.") + endif() + + # Get last line of running nuget.exe list . + string(REPLACE "\n" ";" nuget_list_output ${nuget_list_output}) + list(POP_BACK nuget_list_output nuget_list_last_line) + if(nuget_list_last_line STREQUAL "No packages found.") + message(FATAL_ERROR "NuGet failed to find latest version of package ${ARG_ID}.") + endif() + + # The last line should have the format + string(REPLACE " " ";" nuget_list_last_line ${nuget_list_last_line}) + list(POP_BACK nuget_list_last_line nuget_version) + + if(NOT nuget_version) + message(FATAL_ERROR "NuGet failed to find latest version of package ${ARG_ID}.") + endif() + + message("Nuget found version:${nuget_version} for ${ARG_ID}") + + # Save output variable and cache the result so subsequent calls to the version-unspecified package + # are faster. + set(${ARG_OUTPUT_VARIABLE} ${nuget_version} PARENT_SCOPE) + set(${ARG_ID}_LATEST_VERSION ${nuget_version} CACHE INTERNAL "") + endif() +endfunction() + +# Installs a NuGet package under OUTPUT_DIR. +# +# FetchNuGetPackage( +# ID Microsoft.Direct3D.WARP +# VERSION 1.0.13 +# SOURCE https://api.nuget.org/v3/index.json +# ) +# +# This function sets a variable _SOURCE_DIR (e.g. Microsoft.Direct3D.WARP_SOURCE_DIR in above example) to the +# extract NuGet package contents. +function(FetchNuGetPackage) + set(params NAME ID VERSION SOURCE OUTPUT_DIR RELEASE_TYPE) + cmake_parse_arguments(PARSE_ARGV 0 ARG "" "${params}" "") + + # The NAME parameter is optional: if it's not set then the package ID is used as the name. The + # reason for having a separate NAME is to allow a consistent identifier for packages whose ID + # changes with each release (e.g. GDK). + if(NOT ARG_NAME) + set(ARG_NAME ${ARG_ID}) + endif() + + if(NOT ARG_OUTPUT_DIR) + set(ARG_OUTPUT_DIR ${BINARY_DIR}/temp) + endif() + + set(nuget_exe_path ${ARG_OUTPUT_DIR}/nuget.exe) + + if(NOT ARG_SOURCE) + set(ARG_SOURCE https://api.nuget.org/v3/index.json) + endif() + + if(NOT ARG_RELEASE_TYPE) + set(ARG_RELEASE_TYPE "LATEST_RELEASE") + endif() + + set(PREVIEW OFF) + + if(${ARG_RELEASE_TYPE} STREQUAL "LATEST_PREVIEW") + set(PREVIEW ON) + endif() + + # Default to latest version + if(NOT ARG_VERSION) + GetNuGetPackageLatestVersion( + ID ${ARG_ID} + SOURCE ${ARG_SOURCE} + PREVIEW ${PREVIEW} + OUTPUT_DIR ${ARG_OUTPUT_DIR} + OUTPUT_VARIABLE ARG_VERSION + ) + endif() + + set(nupkg_path ${ARG_OUTPUT_DIR}/${ARG_ID}.${ARG_VERSION}/${ARG_ID}.${ARG_VERSION}.nupkg) + + if(NOT EXISTS ${nupkg_path}) + message(STATUS "NuGet: adding package ${ARG_ID}.${ARG_VERSION}") + + EnsureNugetExists(${nuget_exe_path}) + + set(retry_count 0) + set(max_retries 10) + set(retry_delay 10) + set(result 1) + + # Run NuGet CLI to download the package. + while(NOT ${result} STREQUAL "0" AND ${retry_count} LESS ${max_retries}) + message(STATUS "'${nuget_exe_path}' install '${ARG_ID}' -Version '${ARG_VERSION}' -Source '${ARG_SOURCE}' -OutputDirectory '${ARG_OUTPUT_DIR}' -DependencyVersion Ignore -Verbosity quiet") + execute_process( + COMMAND + ${nuget_exe_path} + install ${ARG_ID} + -Version ${ARG_VERSION} + -Source ${ARG_SOURCE} + -OutputDirectory ${ARG_OUTPUT_DIR} + -DependencyVersion Ignore + -Verbosity quiet + RESULT_VARIABLE result + ) + if(NOT ${result} STREQUAL "0") + math(EXPR retry_count "${retry_count} + 1") + + message(STATUS "Nuget failed: '${result}'. Retrying in ${retry_delay} seconds...") + execute_process( + COMMAND + ${CMAKE_COMMAND} -E sleep ${retry_delay} + ) + endif() + endwhile() + + if(NOT ${result} STREQUAL "0") + message(FATAL_ERROR "NuGet failed: '${result}' Package '${ARG_NAME}' (${ARG_ID}.${ARG_VERSION})") + endif() + endif() + + # Set output variable. The NAME parameter is optional: if it's not set then the package ID is used as the + # name. The reason for having a separate NAME is for packages whose IDs change (e.g. GDK) so that callers + # can use a fixed name in dependents. Example, targets can reference gdk_SOURCE_DIR with the snippet below + # instead of having to reference Microsoft.GDK.PC.230300_SOURCE_DIR. + # + # FetchNuGetPackage( + # NAME gdk + # ID Microsoft.GDK.PC.220300 + # VERSION 10.0.22621.3049 + # ) + set(${ARG_NAME}_SOURCE_DIR ${ARG_OUTPUT_DIR}/${ARG_ID}.${ARG_VERSION} PARENT_SCOPE) +endfunction() + +# Begin the 'main' logic of this file. Previous code is all defintions. +message("USE_WARP_FROM_NUGET: ${USE_WARP_FROM_NUGET}") +if(${USE_WARP_FROM_NUGET} STREQUAL "LATEST_RELEASE" OR ${USE_WARP_FROM_NUGET} STREQUAL "LATEST_PREVIEW") + + message("Fetching warp from nuget") + + FetchNuGetPackage(ID Microsoft.Direct3D.WARP OUTPUT_DIR ${BINARY_DIR}/temp RELEASE_TYPE ${USE_WARP_FROM_NUGET}) + + if(${CMAKE_SYSTEM_PROCESSOR} STREQUAL "AMD64") + set(ARCH "x64") + endif() + if(${CMAKE_SYSTEM_PROCESSOR} STREQUAL "X86") + set(ARCH "win32") + endif() + if(${CMAKE_SYSTEM_PROCESSOR} STREQUAL "ARM64") + set(ARCH "arm64") + endif() + + set(WARP_SOURCE_PATH "${Microsoft.Direct3D.WARP_SOURCE_DIR}/build/native/bin/${ARCH}") + set(WARP_DEST_PATH "${BINARY_DIR}/${BUILD_TYPE}/bin/") + message("Copying d3d10warp.dll and d3d10warp.pdb \n" + " from: ${WARP_SOURCE_PATH}\n" + " to: ${WARP_DEST_PATH}") + file(COPY "${WARP_SOURCE_PATH}/d3d10warp.dll" + DESTINATION "${WARP_DEST_PATH}") + file(COPY "${WARP_SOURCE_PATH}/d3d10warp.pdb" + DESTINATION "${WARP_DEST_PATH}") +endif() diff --git a/tools/clang/test/CMakeLists.txt b/tools/clang/test/CMakeLists.txt index 0e4733ebb7..6b0aa32fbc 100644 --- a/tools/clang/test/CMakeLists.txt +++ b/tools/clang/test/CMakeLists.txt @@ -147,6 +147,15 @@ if (WIN32) ) set(TAEF_EXEC_ADAPTER "" CACHE STRING "adapter for taef exec test") + # Use a custom target so we can depend on it and re-run the cmake logic which downloads warp + # from nuget if requested. + add_custom_target(WarpFromNuget + COMMAND "${CMAKE_COMMAND}" + -DCMAKE_SYSTEM_PROCESSOR=${CMAKE_SYSTEM_PROCESSOR} + -DBUILD_TYPE=${CMAKE_BUILD_TYPE} + -DBINARY_DIR=${CMAKE_BINARY_DIR} + -P "${CMAKE_SOURCE_DIR}/cmake/modules/nuget.cmake") + add_lit_target("check-clang-taef-exec" "Running lit suite hlsl execution test" ${CMAKE_CURRENT_SOURCE_DIR}/taef_exec PARAMS ${CLANG_TEST_PARAMS} @@ -154,5 +163,13 @@ if (WIN32) DEPENDS ExecHLSLTests dxexp ARGS ${CLANG_TEST_EXTRA_ARGS} ) + + add_lit_target("check-clang-taef-exec-warp" "Running lit suite hlsl execution test with D3D WARP from nuget" + ${CMAKE_CURRENT_SOURCE_DIR}/taef_exec + PARAMS ${CLANG_TEST_PARAMS} + adapter=${TAEF_EXEC_ADAPTER} + DEPENDS ExecHLSLTests dxexp WarpFromNuget + ARGS ${CLANG_TEST_EXTRA_ARGS} + ) endif() # HLSL Change End diff --git a/tools/clang/unittests/HLSLExec/ExecutionTest.cpp b/tools/clang/unittests/HLSLExec/ExecutionTest.cpp index f22e99e467..79a28d26f1 100644 --- a/tools/clang/unittests/HLSLExec/ExecutionTest.cpp +++ b/tools/clang/unittests/HLSLExec/ExecutionTest.cpp @@ -819,6 +819,15 @@ class ExecutionTest { return false; } + + if (GetModuleHandle("d3d10warp.dll") != NULL) { + CHAR szFullModuleFilePath[MAX_PATH] = ""; + GetModuleFileName(GetModuleHandle("d3d10warp.dll"), + szFullModuleFilePath, sizeof(szFullModuleFilePath)); + WEX::Logging::Log::Comment(WEX::Common::String().Format( + L"WARP driver loaded from: %S", szFullModuleFilePath)); + } + } else { CComPtr hardwareAdapter; WEX::Common::String AdapterValue; diff --git a/utils/hct/hctbuild.cmd b/utils/hct/hctbuild.cmd index 19e32c4731..ff6dbfa22a 100644 --- a/utils/hct/hctbuild.cmd +++ b/utils/hct/hctbuild.cmd @@ -46,6 +46,7 @@ set WINSDK_MIN_VERSION=10.0.17763.0 set INSTALL_DIR= set DEFAULT_EXEC_ADAPTER=-DTAEF_EXEC_ADAPTER= set LIT_ARGS= +set FRESH= :parse_args if "%1"=="" ( @@ -208,6 +209,10 @@ if "%1"=="-sanitizer" ( set CMAKE_OPTS=%CMAKE_OPTS% -DLLVM_USE_SANITIZER:STRING=Address shift /1 & goto :parse_args ) +if "%1"=="-fresh" ( + set FRESH="--fresh" + shift /1 & goto :parse_args +) rem Begin SPIRV change @@ -468,13 +473,13 @@ cd /d %3 if "%DO_SETUP%"=="1" ( echo Creating solution files for %2, logging to %3\cmake-log.txt if "%BUILD_GENERATOR%"=="Ninja" ( - echo Running "%CMAKE_PATH%" -DCMAKE_BUILD_TYPE:STRING=%1 %CMAKE_OPTS% -G %4 %HLSL_SRC_DIR% > %3\cmake-log.txt - "%CMAKE_PATH%" -DCMAKE_BUILD_TYPE:STRING=%1 %CMAKE_OPTS% -G %4 %HLSL_SRC_DIR% >> %3\cmake-log.txt 2>&1 + echo Running "%CMAKE_PATH%" %FRESH% -DCMAKE_BUILD_TYPE:STRING=%1 %CMAKE_OPTS% -G %4 %HLSL_SRC_DIR% > %3\cmake-log.txt + "%CMAKE_PATH%" %FRESH% -DCMAKE_BUILD_TYPE:STRING=%1 %CMAKE_OPTS% -G %4 %HLSL_SRC_DIR% >> %3\cmake-log.txt 2>&1 ) else ( rem BUILD_TYPE is mostly ignored in this path as VS generates multiple targets rem it is still needed to satisfy cmake file expectations - echo Running "%CMAKE_PATH%" -DCMAKE_BUILD_TYPE:STRING=%1 %CMAKE_OPTS% -G %4 %5 %HLSL_SRC_DIR% > %3\cmake-log.txt - "%CMAKE_PATH%" -DCMAKE_BUILD_TYPE:STRING=%1 %CMAKE_OPTS% -G %4 %5 %HLSL_SRC_DIR% >> %3\cmake-log.txt 2>&1 + echo Running "%CMAKE_PATH%" %FRESH% -DCMAKE_BUILD_TYPE:STRING=%1 %CMAKE_OPTS% -G %4 %5 %HLSL_SRC_DIR% > %3\cmake-log.txt + "%CMAKE_PATH%" %FRESH% -DCMAKE_BUILD_TYPE:STRING=%1 %CMAKE_OPTS% -G %4 %5 %HLSL_SRC_DIR% >> %3\cmake-log.txt 2>&1 ) if %SHOW_CMAKE_LOG%==1 ( echo ------- Start of %3\cmake-log.txt ------- diff --git a/utils/hct/hcttest.cmd b/utils/hct/hcttest.cmd index 1810efcee4..3e536d199f 100644 --- a/utils/hct/hcttest.cmd +++ b/utils/hct/hcttest.cmd @@ -37,6 +37,8 @@ set TEST_MANUAL_FILE_CHECK=0 set SINGLE_FILE_CHECK_NAME=0 set CUSTOM_BIN_SET= set USE_AGILITY_SDK= +set USE_WARP_FROM_NUGET= +set EXEC_TEST_TARGET="check-clang-taef-exec" rem Begin SPIRV change set TEST_SPIRV=0 @@ -132,6 +134,22 @@ if "%1"=="-clean" ( set TEST_ALL=0 set TEST_EXEC=1 set TEST_EXEC_REQUIRED=1 +) else if "%1"=="exec-warp" ( + rem If exec-warp is explicitly supplied, hcttest will fail if machine is not configured + rem to run execution tests, otherwise, execution tests would be skipped. + set TEST_ALL=0 + set TEST_EXEC=1 + set USE_WARP_FROM_NUGET=LATEST_RELEASE + set TEST_EXEC_REQUIRED=1 + set EXEC_TEST_TARGET="check-clang-taef-exec-warp" +) else if "%1"=="exec-warp-preview" ( + rem If exec-warp-preview is explicitly supplied, hcttest will fail if machine is not configured + rem to run execution tests, otherwise, execution tests would be skipped. + set TEST_ALL=0 + set TEST_EXEC=1 + set USE_WARP_FROM_NUGET=LATEST_PREVIEW + set TEST_EXEC_REQUIRED=1 + set EXEC_TEST_TARGET="check-clang-taef-exec-warp" ) else if "%1"=="exec-filter" ( set TEST_ALL=0 set TEST_EXEC=1 @@ -333,21 +351,22 @@ if "%TEST_USE_LIT%"=="1" ( if defined EXEC_ADAPTER ( py %HLSL_SRC_DIR%/utils/lit/lit.py -v --no-progress-bar --param build_mode=%BUILD_CONFIG% --param clang_site_config=%HLSL_BLD_DIR%/tools/clang/test/lit.site.cfg --param clang_taef_exec_site_config=%HLSL_BLD_DIR%/tools/clang/test/taef_exec/lit.site.cfg %EXEC_ADAPTER% %HLSL_SRC_DIR%/tools/clang/test/taef_exec ) else ( - cmake --build %HLSL_BLD_DIR% --config %BUILD_CONFIG% --target check-clang-taef-exec + cmake --build %HLSL_BLD_DIR% --config %BUILD_CONFIG% --target %EXEC_TEST_TARGET% ) set RES_EXEC=!ERRORLEVEL! ) ) - set TEST_CLANG=0 - set TEST_DXILCONV=0 - set TEST_SPIRV=0 - set TEST_EXEC=0 - set TEST_CMD=0 rem No other tests to run - skip copying and move on to report the results if not exist "%HCT_EXTRAS%\hcttest-extras.cmd" ( goto :report_results ) + + set TEST_CLANG=0 + set TEST_DXILCONV=0 + set TEST_SPIRV=0 + set TEST_EXEC=0 + set TEST_CMD=0 ) if not exist %TEST_DIR% (mkdir %TEST_DIR%) From c2ed9ad4ee775f3de903ce757c994aecc59a5306 Mon Sep 17 00:00:00 2001 From: Steven Perron Date: Mon, 3 Feb 2025 10:01:15 -0500 Subject: [PATCH 15/20] [SPIRV] Avoid invlid SIPR-V (#7107) 1. Array with the block decoration are not suppose to have an array stride. 2. The CullPrimitive output should be an array of bools. --- tools/clang/lib/SPIRV/DeclResultIdMapper.cpp | 24 ++++++++++++------- tools/clang/lib/SPIRV/SpirvEmitter.cpp | 3 ++- .../meshshading.ext.cullprimative.hlsl | 6 ++--- .../sm6_6.descriptorheap.acbuffer.hlsl | 3 +++ ...ype.rwstructured-buffer.array.counter.hlsl | 4 ++++ ...ctured-buffer.array.unbounded.counter.hlsl | 4 ++++ 6 files changed, 31 insertions(+), 13 deletions(-) diff --git a/tools/clang/lib/SPIRV/DeclResultIdMapper.cpp b/tools/clang/lib/SPIRV/DeclResultIdMapper.cpp index f84a0e34ed..fd0fa8a3d0 100644 --- a/tools/clang/lib/SPIRV/DeclResultIdMapper.cpp +++ b/tools/clang/lib/SPIRV/DeclResultIdMapper.cpp @@ -392,10 +392,16 @@ bool isBooleanStageIOVar(const NamedDecl *decl, QualType type, // [[vk::builtin(...)]] makes the decl a built-in stage variable. // IsFrontFace (if used as PSIn) is the only known boolean built-in stage // variable. - const bool isBooleanBuiltin = - (decl->getAttr() != nullptr) || - (semanticKind == hlsl::Semantic::Kind::IsFrontFace && - sigPointKind == hlsl::SigPoint::Kind::PSIn); + bool isBooleanBuiltin = false; + + if ((decl->getAttr() != nullptr)) + isBooleanBuiltin = true; + else if (semanticKind == hlsl::Semantic::Kind::IsFrontFace && + sigPointKind == hlsl::SigPoint::Kind::PSIn) { + isBooleanBuiltin = true; + } else if (semanticKind == hlsl::Semantic::Kind::CullPrimitive) { + isBooleanBuiltin = true; + } // TODO: support boolean matrix stage I/O variable if needed. QualType elemType = {}; @@ -1816,24 +1822,24 @@ void DeclResultIdMapper::createCounterVar( } const SpirvType *counterType = spvContext.getACSBufferCounterType(); + llvm::Optional noArrayStride; QualType declType = decl->getType(); if (declType->isArrayType()) { // Vulkan does not support multi-dimentional arrays of resource, so we // assume the array is a single dimensional array. assert(!declType->getArrayElementTypeNoTypeQual()->isArrayType()); - uint32_t arrayStride = 4; + if (const auto *constArrayType = astContext.getAsConstantArrayType(declType)) { counterType = spvContext.getArrayType( - counterType, constArrayType->getSize().getZExtValue(), arrayStride); + counterType, constArrayType->getSize().getZExtValue(), noArrayStride); } else { assert(declType->isIncompleteArrayType()); - counterType = spvContext.getRuntimeArrayType(counterType, arrayStride); + counterType = spvContext.getRuntimeArrayType(counterType, noArrayStride); } } else if (isResourceDescriptorHeap(decl->getType()) || isSamplerDescriptorHeap(decl->getType())) { - counterType = - spvContext.getRuntimeArrayType(counterType, /* arrayStride= */ 4); + counterType = spvContext.getRuntimeArrayType(counterType, noArrayStride); } // {RW|Append|Consume}StructuredBuffer are all in Uniform storage class. diff --git a/tools/clang/lib/SPIRV/SpirvEmitter.cpp b/tools/clang/lib/SPIRV/SpirvEmitter.cpp index f48245ac38..9addb6d22c 100644 --- a/tools/clang/lib/SPIRV/SpirvEmitter.cpp +++ b/tools/clang/lib/SPIRV/SpirvEmitter.cpp @@ -8095,7 +8095,8 @@ void SpirvEmitter::assignToMSOutAttribute( // All other attribute writes are handled below. auto *varInstr = declIdMapper.getStageVarInstruction(decl); QualType valueType = value->getAstResultType(); - if (valueType->isBooleanType()) { + if (valueType->isBooleanType() && + semanticInfo.getKind() != hlsl::Semantic::Kind::CullPrimitive) { // Externally visible variables are changed to uint, so we need to cast the // value to uint. value = castToInt(value, valueType, astContext.UnsignedIntTy, loc); diff --git a/tools/clang/test/CodeGenSPIRV/meshshading.ext.cullprimative.hlsl b/tools/clang/test/CodeGenSPIRV/meshshading.ext.cullprimative.hlsl index d45f4bdb6e..2a143afab2 100644 --- a/tools/clang/test/CodeGenSPIRV/meshshading.ext.cullprimative.hlsl +++ b/tools/clang/test/CodeGenSPIRV/meshshading.ext.cullprimative.hlsl @@ -9,7 +9,7 @@ struct MeshletPrimitiveOut // specification says that externally visible variables cannot be bool. // CHECK: OpDecorate [[var:%[0-9]+]] BuiltIn CullPrimitiveEXT // CHECK: OpDecorate [[var]] PerPrimitiveEXT -// CHECK: [[var]] = OpVariable %_ptr_Output__arr_uint_uint_2 Output +// CHECK: [[var]] = OpVariable %_ptr_Output__arr_bool_uint_2 Output struct VertOut { @@ -28,7 +28,7 @@ void main(uint svGroupIndex : SV_GROUPINDEX, // Make sure that the references to m_cullPrimitive use uints. // CHECK: [[idx:%[0-9]+]] = OpLoad %uint %gl_LocalInvocationIndex -// CHECK: [[ac:%[0-9]+]] = OpAccessChain %_ptr_Output_uint [[var]] [[idx]] -// CHECK: OpStore [[ac]] %uint_0 +// CHECK: [[ac:%[0-9]+]] = OpAccessChain %_ptr_Output_bool [[var]] [[idx]] +// CHECK: OpStore [[ac]] %false primitives[svGroupIndex].m_cullPrimitive = false; } diff --git a/tools/clang/test/CodeGenSPIRV/sm6_6.descriptorheap.acbuffer.hlsl b/tools/clang/test/CodeGenSPIRV/sm6_6.descriptorheap.acbuffer.hlsl index 053a32869a..359ac0cb70 100644 --- a/tools/clang/test/CodeGenSPIRV/sm6_6.descriptorheap.acbuffer.hlsl +++ b/tools/clang/test/CodeGenSPIRV/sm6_6.descriptorheap.acbuffer.hlsl @@ -9,6 +9,9 @@ // CHECK: OpDecorate [[resource_heap_abuffer]] Binding 0 // CHECK: OpDecorate [[resource_heap_abuffer_counter:%[_a-zA-Z0-9]+]] DescriptorSet 0 // CHECK: OpDecorate [[resource_heap_abuffer_counter]] Binding 1 +// CHECK-NOT: OpDecorate %_runtimearr_type_ACSBuffer_counter ArrayStride +// CHECK: OpDecorate %type_ACSBuffer_counter BufferBlock +// CHECK-NOT: OpDecorate %_runtimearr_type_ACSBuffer_counter ArrayStride // CHECK-DAG: [[ra_uint_t:%[_a-zA-Z0-9]+]] = OpTypeRuntimeArray %uint diff --git a/tools/clang/test/CodeGenSPIRV/type.rwstructured-buffer.array.counter.hlsl b/tools/clang/test/CodeGenSPIRV/type.rwstructured-buffer.array.counter.hlsl index cdf7991a12..4e56f62c1c 100644 --- a/tools/clang/test/CodeGenSPIRV/type.rwstructured-buffer.array.counter.hlsl +++ b/tools/clang/test/CodeGenSPIRV/type.rwstructured-buffer.array.counter.hlsl @@ -7,8 +7,12 @@ struct PSInput // CHECK: OpDecorate %g_rwbuffer DescriptorSet 2 // CHECK: OpDecorate %g_rwbuffer Binding 0 +// CHECK-NOT: OpDecorate %_arr_type_ACSBuffer_counter_uint_5 ArrayStride // CHECK: OpDecorate %counter_var_g_rwbuffer DescriptorSet 2 // CHECK: OpDecorate %counter_var_g_rwbuffer Binding 1 +// CHECK-NOT: OpDecorate %_arr_type_ACSBuffer_counter_uint_5 ArrayStride +// CHECK: OpDecorate %type_ACSBuffer_counter BufferBlock +// CHECK-NOT: OpDecorate %_runtimearr_type_ACSBuffer_counter ArrayStride // CHECK: %g_rwbuffer = OpVariable %_ptr_Uniform__arr_type_RWStructuredBuffer_uint_uint_5 Uniform // CHECK: %counter_var_g_rwbuffer = OpVariable %_ptr_Uniform__arr_type_ACSBuffer_counter_uint_5 Uniform diff --git a/tools/clang/test/CodeGenSPIRV/type.rwstructured-buffer.array.unbounded.counter.hlsl b/tools/clang/test/CodeGenSPIRV/type.rwstructured-buffer.array.unbounded.counter.hlsl index 535d79053d..1c0ea8520f 100644 --- a/tools/clang/test/CodeGenSPIRV/type.rwstructured-buffer.array.unbounded.counter.hlsl +++ b/tools/clang/test/CodeGenSPIRV/type.rwstructured-buffer.array.unbounded.counter.hlsl @@ -7,8 +7,12 @@ struct PSInput // CHECK: OpDecorate %g_rwbuffer DescriptorSet 2 // CHECK: OpDecorate %g_rwbuffer Binding 0 +// CHECK-NOT: OpDecorate %_runtimearr_type_ACSBuffer_counter ArrayStride // CHECK: OpDecorate %counter_var_g_rwbuffer DescriptorSet 2 // CHECK: OpDecorate %counter_var_g_rwbuffer Binding 1 +// CHECK-NOT: OpDecorate %_runtimearr_type_ACSBuffer_counter ArrayStride +// CHECK: OpDecorate %type_ACSBuffer_counter BufferBlock +// CHECK-NOT: OpDecorate %_runtimearr_type_ACSBuffer_counter ArrayStride // CHECK: %g_rwbuffer = OpVariable %_ptr_Uniform__runtimearr_type_RWStructuredBuffer_uint Uniform // CHECK: %counter_var_g_rwbuffer = OpVariable %_ptr_Uniform__runtimearr_type_ACSBuffer_counter Uniform From f8ef48362027241b3400b116bb8f2b55d3efabfd Mon Sep 17 00:00:00 2001 From: Joshua Batista Date: Thu, 6 Feb 2025 17:03:16 -0800 Subject: [PATCH 16/20] [NFC][Docs] Update release notes / release template process (#7100) Some update to clarify the release procedure and to update the current release notes with PR links. The readme should also now mention the inclusion of dxv.exe in each release. --- .github/ISSUE_TEMPLATE/release.md | 6 +++--- README.md | 2 ++ docs/ReleaseNotes.md | 8 ++++---- 3 files changed, 9 insertions(+), 7 deletions(-) diff --git a/.github/ISSUE_TEMPLATE/release.md b/.github/ISSUE_TEMPLATE/release.md index c8eccd90f8..70f9760cf5 100644 --- a/.github/ISSUE_TEMPLATE/release.md +++ b/.github/ISSUE_TEMPLATE/release.md @@ -13,7 +13,7 @@ assignees: '' - At this point, changes must be cherry-picked into the release branch in order for them to be included in the release. - [ ] MM/DD/YYYY - Release Candidate 1 (begin Ask Mode[^1] for release branch). - - At this point, changes must be approved by @microsoft/hlsl-release + - At this point, cherry-picked changes must be approved by @microsoft/hlsl-release - [ ] MM/DD/YYYY - Final Release Candidate - [ ] MM/DD/YYYY - Target Release Date @@ -26,7 +26,7 @@ This part of the release process is to 'prime the pump' - that is to make sure that all the various parts of the engineering system are set into place so that we are confident we can generate builds for the new branch -- [ ] Update version number +- [ ] Update version numbers in utils/version/latest-release.json and utils/version/version.inc - [ ] Create the release branch from `main` - The release branch is kept into sync with main via regular fast-forward merges. @@ -39,7 +39,7 @@ we are confident we can generate builds for the new branch ## After Fork -- [ ] Update README.md +- [ ] Update README.md if necessary - [ ] Create draft of Release post on GitHub ## Quality Sign Off diff --git a/README.md b/README.md index a06ee5ab97..35c0132068 100644 --- a/README.md +++ b/README.md @@ -18,6 +18,8 @@ At the moment, the DirectX HLSL Compiler provides the following components: - dxilconv.dll, a DLL providing a converter from DXBC (older shader bytecode format) +- dxv.exe, a command-line tool that validates DXIL IR (compiled HLSL programs). + - various other tools based on the above components The Microsoft Windows SDK releases include a supported version of the compiler and validator. diff --git a/docs/ReleaseNotes.md b/docs/ReleaseNotes.md index 04ae484f34..bd0b6d17e2 100644 --- a/docs/ReleaseNotes.md +++ b/docs/ReleaseNotes.md @@ -23,10 +23,10 @@ Place release notes for the upcoming release below this line and remove this lin ### Version 1.8.2502 -- The incomplete WaveMatrix implementation has been removed. -- DXIL Validator Hash is open sourced. -- DXIL container validation for PSV0 part allows any content ordering inside string and semantic index tables. -- The and() and or() intrinsics will now accept non-integer parameters by casting them to bools. +- The incomplete WaveMatrix implementation has been removed. [#6807](https://github.com/microsoft/DirectXShaderCompiler/pull/6807) +- DXIL Validator Hash is open sourced. [#6846](https://github.com/microsoft/DirectXShaderCompiler/pull/6846) +- DXIL container validation for PSV0 part allows any content ordering inside string and semantic index tables. [#6859](https://github.com/microsoft/DirectXShaderCompiler/pull/6859) +- The and() and or() intrinsics will now accept non-integer parameters by casting them to bools. [#7060](https://github.com/microsoft/DirectXShaderCompiler/pull/7060) ### Version 1.8.2407 From 20950d662f4d24850d7edfe1e647f6749d5c13b3 Mon Sep 17 00:00:00 2001 From: Joshua Batista Date: Fri, 7 Feb 2025 11:11:46 -0800 Subject: [PATCH 17/20] [NFC] Remove unused functions in ExecutionTest.cpp and assert.cpp (#7129) Some functions need to be removed / changed to unblock some internal pipelines. They are emitting warnings that they are unused. Specifically, `CompareOutputWithExpectedValueFloat`, `CompareOutputWithExpectedValueHalf`, and `CompareOutputWithExpectedValueFloat` need to be removed because they are unused. Additionally, there is a case in assert.cpp where `llvm_assert_trap` is left unused depending on what's been defined and how the preprocessor directives execute. This needs to be remedied so that the function is only defined when it is used. --- lib/Support/assert.cpp | 3 ++ .../unittests/HLSLExec/ExecutionTest.cpp | 54 ------------------- 2 files changed, 3 insertions(+), 54 deletions(-) diff --git a/lib/Support/assert.cpp b/lib/Support/assert.cpp index 991ae01857..75111ea405 100644 --- a/lib/Support/assert.cpp +++ b/lib/Support/assert.cpp @@ -10,6 +10,8 @@ #include "assert.h" #include "llvm/Support/Compiler.h" #include "llvm/Support/raw_ostream.h" + +#if defined(LLVM_ASSERTIONS_TRAP) || !defined(WIN32) namespace { void llvm_assert_trap(const char *_Message, const char *_File, unsigned _Line, const char *_Function) { @@ -18,6 +20,7 @@ void llvm_assert_trap(const char *_Message, const char *_File, unsigned _Line, LLVM_BUILTIN_TRAP; } } // namespace +#endif #ifdef _WIN32 #include "dxc/Support/Global.h" diff --git a/tools/clang/unittests/HLSLExec/ExecutionTest.cpp b/tools/clang/unittests/HLSLExec/ExecutionTest.cpp index 79a28d26f1..7066247883 100644 --- a/tools/clang/unittests/HLSLExec/ExecutionTest.cpp +++ b/tools/clang/unittests/HLSLExec/ExecutionTest.cpp @@ -6880,22 +6880,6 @@ ToleranceType ToleranceStringToEnum(LPCWSTR toleranceType) { } } -static bool CompareOutputWithExpectedValueFloat( - float output, float ref, ToleranceType type, double tolerance, - hlsl::DXIL::Float32DenormMode mode = hlsl::DXIL::Float32DenormMode::Any) { - if (type == ToleranceType::RELATIVE_EPSILON) { - return CompareFloatRelativeEpsilon(output, ref, (int)tolerance, mode); - } else if (type == ToleranceType::EPSILON) { - return CompareFloatEpsilon(output, ref, (float)tolerance, mode); - } else if (type == ToleranceType::ULP) { - return CompareFloatULP(output, ref, (int)tolerance, mode); - } else { - LogErrorFmt(L"Failed to read comparison type %S", type); - } - - return false; -} - static bool CompareOutputWithExpectedValueFloat( float output, float ref, LPCWSTR type, double tolerance, hlsl::DXIL::Float32DenormMode mode = hlsl::DXIL::Float32DenormMode::Any) { @@ -6919,21 +6903,6 @@ static bool VerifyOutputWithExpectedValueFloat( CompareOutputWithExpectedValueFloat(output, ref, type, tolerance, mode)); } -static bool CompareOutputWithExpectedValueHalf(uint16_t output, uint16_t ref, - ToleranceType type, - double tolerance) { - if (type == ToleranceType::RELATIVE_EPSILON) { - return CompareHalfRelativeEpsilon(output, ref, (int)tolerance); - } else if (type == ToleranceType::EPSILON) { - return CompareHalfEpsilon(output, ref, (float)tolerance); - } else if (type == ToleranceType::ULP) { - return CompareHalfULP(output, ref, (float)tolerance); - } else { - LogErrorFmt(L"Failed to read comparison type %S", type); - return false; - } -} - static bool CompareOutputWithExpectedValueHalf(uint16_t output, uint16_t ref, LPCWSTR type, double tolerance) { if (_wcsicmp(type, L"Relative") == 0) { @@ -6954,29 +6923,6 @@ static bool VerifyOutputWithExpectedValueHalf(uint16_t output, uint16_t ref, CompareOutputWithExpectedValueHalf(output, ref, type, tolerance)); } -template -static bool CompareOutputWithExpectedValue(T output, T ref, - ToleranceType toleranceType, - double tolerance) { - if (std::is_same::value) { // uint16 treated - // as half - return CompareOutputWithExpectedValueHalf((uint16_t)output, (uint16_t)ref, - toleranceType, tolerance); - } else if (std::is_integral::value && - std::is_signed::value) { // signed ints - return CompareOutputWithExpectedValueInt((int)output, (int)ref, - (int)tolerance); - } else if (std::is_integral::value) { // unsigned ints - return CompareOutputWithExpectedValueUInt((uint32_t)output, (uint32_t)ref, - (uint32_t)tolerance); - } else if (std::is_floating_point::value) { // floating point - return CompareOutputWithExpectedValueFloat((float)output, (float)ref, - toleranceType, tolerance); - } - - DXASSERT_NOMSG("Invalid Parameter Type"); -} - template static bool CompareOutputWithExpectedValue(T output, T ref, LPCWSTR toleranceType, From 6b68f3b5832cb57dc69b2902b9ab520c9506f77a Mon Sep 17 00:00:00 2001 From: Steven Perron Date: Tue, 18 Feb 2025 05:31:37 -0500 Subject: [PATCH 18/20] [SPIRV] Update link to inline spirv wiki (#7140) Fixes #6412 --- docs/SPIR-V.rst | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/docs/SPIR-V.rst b/docs/SPIR-V.rst index 5093dd0870..c30286e4e6 100644 --- a/docs/SPIR-V.rst +++ b/docs/SPIR-V.rst @@ -4145,7 +4145,7 @@ GL_EXT_spirv_intrinsics is an extension of GLSL that allows users to embed arbitrary SPIR-V instructions in the GLSL code similar to the concept of inline assembly in the C code. We support the HLSL version of GL_EXT_spirv_intrinsics. See -`wiki `_ +`wiki `_ for the details. Supported Command-line Options From d2f2aaeb14bc75748fe3541610c2236828b17de7 Mon Sep 17 00:00:00 2001 From: Steven Perron Date: Tue, 18 Feb 2025 06:52:52 -0500 Subject: [PATCH 19/20] [SPIRV] Stop running ReplaceInvalidOpcPass (#7138) That pass was intended to remove derivative instructions used in non-fragment shaders. Now they are available in more places because of new extensions. In general, it is better practice to issue an error if they are used in an place they are not allowed. Fixes https://github.com/microsoft/DirectXShaderCompiler/issues/7086 --- tools/clang/lib/SPIRV/SpirvEmitter.cpp | 1 - .../CodeGenSPIRV/texture.calculate.lod.compute.linear.hlsl | 3 ++- .../CodeGenSPIRV/texture.calculate.lod.compute.quad.hlsl | 3 ++- .../texture.calculate.lod.unclamped.compute.linear.hlsl | 3 ++- .../texture.calculate.lod.unclamped.compute.quad.hlsl | 3 ++- .../test/CodeGenSPIRV/texture.sample.compute.linear.hlsl | 3 ++- .../test/CodeGenSPIRV/texture.sample.compute.quad.hlsl | 3 ++- .../CodeGenSPIRV/texture.samplebias.compute.linear.hlsl | 7 ++++--- .../test/CodeGenSPIRV/texture.samplebias.compute.quad.hlsl | 3 ++- .../CodeGenSPIRV/texture.samplecmp.compute.linear.hlsl | 3 ++- .../test/CodeGenSPIRV/texture.samplecmp.compute.quad.hlsl | 3 ++- 11 files changed, 22 insertions(+), 13 deletions(-) diff --git a/tools/clang/lib/SPIRV/SpirvEmitter.cpp b/tools/clang/lib/SPIRV/SpirvEmitter.cpp index 9addb6d22c..f9c54fea45 100644 --- a/tools/clang/lib/SPIRV/SpirvEmitter.cpp +++ b/tools/clang/lib/SPIRV/SpirvEmitter.cpp @@ -15219,7 +15219,6 @@ bool SpirvEmitter::spirvToolsLegalize(std::vector *mod, optimizer.RegisterPass( spvtools::CreateAggressiveDCEPass(spirvOptions.preserveInterface)); } - optimizer.RegisterPass(spvtools::CreateReplaceInvalidOpcodePass()); optimizer.RegisterPass(spvtools::CreateCompactIdsPass()); optimizer.RegisterPass(spvtools::CreateSpreadVolatileSemanticsPass()); if (spirvOptions.fixFuncCallArguments) { diff --git a/tools/clang/test/CodeGenSPIRV/texture.calculate.lod.compute.linear.hlsl b/tools/clang/test/CodeGenSPIRV/texture.calculate.lod.compute.linear.hlsl index cb2e8c5916..31d3b2e2cb 100644 --- a/tools/clang/test/CodeGenSPIRV/texture.calculate.lod.compute.linear.hlsl +++ b/tools/clang/test/CodeGenSPIRV/texture.calculate.lod.compute.linear.hlsl @@ -1,4 +1,5 @@ // RUN: %dxc -T cs_6_6 -E main -fspv-extension=SPV_NV_compute_shader_derivatives -fcgl %s -spirv 2>&1 | FileCheck %s --check-prefix=CHECK +// RUN: %dxc -T cs_6_6 -E main -fspv-extension=SPV_NV_compute_shader_derivatives %s -spirv 2>&1 | FileCheck %s --check-prefix=CHECK // CHECK: OpCapability ComputeDerivativeGroupLinearKHR // CHECK: OpExtension "SPV_NV_compute_shader_derivatives" @@ -19,4 +20,4 @@ void main(uint3 id : SV_GroupThreadID) //CHECK-NEXT: [[query1:%[0-9]+]] = OpImageQueryLod %v2float [[si1]] %float_0_5 //CHECK-NEXT: {{%[0-9]+}} = OpCompositeExtract %float [[query1]] 0 o[0] = t1.CalculateLevelOfDetail(ss, 0.5); -} \ No newline at end of file +} diff --git a/tools/clang/test/CodeGenSPIRV/texture.calculate.lod.compute.quad.hlsl b/tools/clang/test/CodeGenSPIRV/texture.calculate.lod.compute.quad.hlsl index 84cdcdf22a..8b4e41b768 100644 --- a/tools/clang/test/CodeGenSPIRV/texture.calculate.lod.compute.quad.hlsl +++ b/tools/clang/test/CodeGenSPIRV/texture.calculate.lod.compute.quad.hlsl @@ -1,4 +1,5 @@ // RUN: %dxc -T cs_6_6 -E main -fspv-extension=SPV_NV_compute_shader_derivatives -fcgl %s -spirv 2>&1 | FileCheck %s +// RUN: %dxc -T cs_6_6 -E main -fspv-extension=SPV_NV_compute_shader_derivatives %s -spirv 2>&1 | FileCheck %s // CHECK: OpCapability ComputeDerivativeGroupQuadsKHR // CHECK: OpExtension "SPV_NV_compute_shader_derivatives" @@ -19,4 +20,4 @@ void main(uint3 id : SV_GroupThreadID) //CHECK-NEXT: [[query1:%[0-9]+]] = OpImageQueryLod %v2float [[si1]] %float_0_5 //CHECK-NEXT: {{%[0-9]+}} = OpCompositeExtract %float [[query1]] 0 o[0] = t1.CalculateLevelOfDetail(ss, 0.5); -} \ No newline at end of file +} diff --git a/tools/clang/test/CodeGenSPIRV/texture.calculate.lod.unclamped.compute.linear.hlsl b/tools/clang/test/CodeGenSPIRV/texture.calculate.lod.unclamped.compute.linear.hlsl index efbb0d82a5..12990fbf2f 100644 --- a/tools/clang/test/CodeGenSPIRV/texture.calculate.lod.unclamped.compute.linear.hlsl +++ b/tools/clang/test/CodeGenSPIRV/texture.calculate.lod.unclamped.compute.linear.hlsl @@ -1,4 +1,5 @@ // RUN: %dxc -T cs_6_6 -E main -fspv-extension=SPV_NV_compute_shader_derivatives -fcgl %s -spirv 2>&1 | FileCheck %s +// RUN: %dxc -T cs_6_6 -E main -fspv-extension=SPV_NV_compute_shader_derivatives %s -spirv 2>&1 | FileCheck %s // CHECK: OpCapability ComputeDerivativeGroupLinearKHR // CHECK: OpExtension "SPV_NV_compute_shader_derivatives" @@ -19,4 +20,4 @@ void main(uint3 id : SV_GroupThreadID) //CHECK-NEXT: [[query1:%[0-9]+]] = OpImageQueryLod %v2float [[si1]] %float_0_5 //CHECK-NEXT: {{%[0-9]+}} = OpCompositeExtract %float [[query1]] 1 o[0] = t1.CalculateLevelOfDetailUnclamped(ss, 0.5); -} \ No newline at end of file +} diff --git a/tools/clang/test/CodeGenSPIRV/texture.calculate.lod.unclamped.compute.quad.hlsl b/tools/clang/test/CodeGenSPIRV/texture.calculate.lod.unclamped.compute.quad.hlsl index 10de7c2583..13f3060818 100644 --- a/tools/clang/test/CodeGenSPIRV/texture.calculate.lod.unclamped.compute.quad.hlsl +++ b/tools/clang/test/CodeGenSPIRV/texture.calculate.lod.unclamped.compute.quad.hlsl @@ -1,4 +1,5 @@ // RUN: %dxc -T cs_6_6 -E main -fspv-extension=SPV_NV_compute_shader_derivatives -fcgl %s -spirv 2>&1 | FileCheck %s +// RUN: %dxc -T cs_6_6 -E main -fspv-extension=SPV_NV_compute_shader_derivatives %s -spirv 2>&1 | FileCheck %s // CHECK: OpCapability ComputeDerivativeGroupQuadsKHR // CHECK: OpExtension "SPV_NV_compute_shader_derivatives" @@ -19,4 +20,4 @@ void main(uint3 id : SV_GroupThreadID) //CHECK-NEXT: [[query1:%[0-9]+]] = OpImageQueryLod %v2float [[si1]] %float_0_5 //CHECK-NEXT: {{%[0-9]+}} = OpCompositeExtract %float [[query1]] 1 o[0] = t1.CalculateLevelOfDetailUnclamped(ss, 0.5); -} \ No newline at end of file +} diff --git a/tools/clang/test/CodeGenSPIRV/texture.sample.compute.linear.hlsl b/tools/clang/test/CodeGenSPIRV/texture.sample.compute.linear.hlsl index 0d56a3a904..d15fcda2e9 100644 --- a/tools/clang/test/CodeGenSPIRV/texture.sample.compute.linear.hlsl +++ b/tools/clang/test/CodeGenSPIRV/texture.sample.compute.linear.hlsl @@ -1,4 +1,5 @@ // RUN: %dxc -T cs_6_6 -E main -fspv-extension=SPV_NV_compute_shader_derivatives -fcgl %s -spirv 2>&1 | FileCheck %s +// RUN: %dxc -T cs_6_6 -E main -fspv-extension=SPV_NV_compute_shader_derivatives %s -spirv 2>&1 | FileCheck %s // CHECK: OpCapability ComputeDerivativeGroupLinearKHR // CHECK: OpExtension "SPV_NV_compute_shader_derivatives" @@ -18,4 +19,4 @@ void main(uint3 id : SV_GroupThreadID) // CHECK-NEXT: [[sampledImg:%[0-9]+]] = OpSampledImage %type_sampled_image [[t1]] [[ss]] // CHECK-NEXT: {{%[0-9]+}} = OpImageSampleImplicitLod %v4float [[sampledImg]] %float_1 None o[0] = t1.Sample(ss, 1); -} \ No newline at end of file +} diff --git a/tools/clang/test/CodeGenSPIRV/texture.sample.compute.quad.hlsl b/tools/clang/test/CodeGenSPIRV/texture.sample.compute.quad.hlsl index bbedad9c66..51bcf9c850 100644 --- a/tools/clang/test/CodeGenSPIRV/texture.sample.compute.quad.hlsl +++ b/tools/clang/test/CodeGenSPIRV/texture.sample.compute.quad.hlsl @@ -1,4 +1,5 @@ // RUN: %dxc -T cs_6_6 -E main -fspv-extension=SPV_NV_compute_shader_derivatives -fcgl %s -spirv 2>&1 | FileCheck %s +// RUN: %dxc -T cs_6_6 -E main -fspv-extension=SPV_NV_compute_shader_derivatives %s -spirv 2>&1 | FileCheck %s // CHECK: OpCapability ComputeDerivativeGroupQuadsKHR // CHECK: OpExtension "SPV_NV_compute_shader_derivatives" @@ -18,4 +19,4 @@ void main(uint3 id : SV_GroupThreadID) // CHECK-NEXT: [[sampledImg:%[0-9]+]] = OpSampledImage %type_sampled_image [[t1]] [[ss]] // CHECK-NEXT: {{%[0-9]+}} = OpImageSampleImplicitLod %v4float [[sampledImg]] %float_1 None o[0] = t1.Sample(ss, 1); -} \ No newline at end of file +} diff --git a/tools/clang/test/CodeGenSPIRV/texture.samplebias.compute.linear.hlsl b/tools/clang/test/CodeGenSPIRV/texture.samplebias.compute.linear.hlsl index 2ce02bb9e4..1cd9965a89 100644 --- a/tools/clang/test/CodeGenSPIRV/texture.samplebias.compute.linear.hlsl +++ b/tools/clang/test/CodeGenSPIRV/texture.samplebias.compute.linear.hlsl @@ -1,4 +1,4 @@ -// RUN: %dxc -T cs_6_6 -E main -fspv-extension=SPV_NV_compute_shader_derivatives -fcgl %s -spirv 2>&1 | FileCheck %s +// RUN: %dxc -T cs_6_6 -E main -fspv-extension=SPV_NV_compute_shader_derivatives %s -spirv 2>&1 | FileCheck %s // CHECK: OpCapability ComputeDerivativeGroupLinearKHR // CHECK: OpExtension "SPV_NV_compute_shader_derivatives" @@ -13,9 +13,10 @@ Texture1D t1; [numthreads(8,1,1)] void main(uint3 id : SV_GroupThreadID) { + Texture1D local_texture = t1; // CHECK: [[t1:%[0-9]+]] = OpLoad %type_1d_image %t1 // CHECK-NEXT: [[ss:%[0-9]+]] = OpLoad %type_sampler %ss // CHECK-NEXT: [[sampledImg:%[0-9]+]] = OpSampledImage %type_sampled_image [[t1]] [[ss]] // CHECK-NEXT: {{%[0-9]+}} = OpImageSampleImplicitLod %v4float [[sampledImg]] %float_1 Bias %float_0_5 - o[0] = t1.SampleBias(ss, 1, 0.5); -} \ No newline at end of file + o[0] = local_texture.SampleBias(ss, 1, 0.5); +} diff --git a/tools/clang/test/CodeGenSPIRV/texture.samplebias.compute.quad.hlsl b/tools/clang/test/CodeGenSPIRV/texture.samplebias.compute.quad.hlsl index a64e6f84db..aaa8d281ea 100644 --- a/tools/clang/test/CodeGenSPIRV/texture.samplebias.compute.quad.hlsl +++ b/tools/clang/test/CodeGenSPIRV/texture.samplebias.compute.quad.hlsl @@ -1,4 +1,5 @@ // RUN: %dxc -T cs_6_6 -E main -fspv-extension=SPV_NV_compute_shader_derivatives -fcgl %s -spirv 2>&1 | FileCheck %s +// RUN: %dxc -T cs_6_6 -E main -fspv-extension=SPV_NV_compute_shader_derivatives %s -spirv 2>&1 | FileCheck %s // CHECK: OpCapability ComputeDerivativeGroupQuadsKHR // CHECK: OpExtension "SPV_NV_compute_shader_derivatives" @@ -18,4 +19,4 @@ void main(uint3 id : SV_GroupThreadID) // CHECK-NEXT: [[sampledImg:%[0-9]+]] = OpSampledImage %type_sampled_image [[t1]] [[ss]] // CHECK-NEXT: {{%[0-9]+}} = OpImageSampleImplicitLod %v4float [[sampledImg]] %float_1 Bias %float_0_5 o[0] = t1.SampleBias(ss, 1, 0.5); -} \ No newline at end of file +} diff --git a/tools/clang/test/CodeGenSPIRV/texture.samplecmp.compute.linear.hlsl b/tools/clang/test/CodeGenSPIRV/texture.samplecmp.compute.linear.hlsl index a5fca1ada7..367a84db45 100644 --- a/tools/clang/test/CodeGenSPIRV/texture.samplecmp.compute.linear.hlsl +++ b/tools/clang/test/CodeGenSPIRV/texture.samplecmp.compute.linear.hlsl @@ -1,4 +1,5 @@ // RUN: %dxc -T cs_6_6 -E main -fspv-extension=SPV_NV_compute_shader_derivatives -fcgl %s -spirv 2>&1 | FileCheck %s +// RUN: %dxc -T cs_6_6 -E main -fspv-extension=SPV_NV_compute_shader_derivatives %s -spirv 2>&1 | FileCheck %s // CHECK: OpCapability ComputeDerivativeGroupLinearKHR // CHECK: OpExtension "SPV_NV_compute_shader_derivatives" @@ -18,4 +19,4 @@ void main(uint3 id : SV_GroupThreadID) // CHECK-NEXT: [[sampledImg:%[0-9]+]] = OpSampledImage %type_sampled_image [[t1]] [[scs]] // CHECK-NEXT: {{%[0-9]+}} = OpImageSampleDrefImplicitLod %float [[sampledImg]] %float_1 %float_0_5 o[0] = t1.SampleCmp(scs, 1, 0.5); -} \ No newline at end of file +} diff --git a/tools/clang/test/CodeGenSPIRV/texture.samplecmp.compute.quad.hlsl b/tools/clang/test/CodeGenSPIRV/texture.samplecmp.compute.quad.hlsl index 550c80f413..96b0c356e5 100644 --- a/tools/clang/test/CodeGenSPIRV/texture.samplecmp.compute.quad.hlsl +++ b/tools/clang/test/CodeGenSPIRV/texture.samplecmp.compute.quad.hlsl @@ -1,4 +1,5 @@ // RUN: %dxc -T cs_6_6 -E main -fspv-extension=SPV_NV_compute_shader_derivatives -fcgl %s -spirv 2>&1 | FileCheck %s +// RUN: %dxc -T cs_6_6 -E main -fspv-extension=SPV_NV_compute_shader_derivatives %s -spirv 2>&1 | FileCheck %s // CHECK: OpCapability ComputeDerivativeGroupQuadsKHR // CHECK: OpExtension "SPV_NV_compute_shader_derivatives" @@ -18,4 +19,4 @@ void main(uint3 id : SV_GroupThreadID) // CHECK-NEXT: [[sampledImg:%[0-9]+]] = OpSampledImage %type_sampled_image [[t1]] [[scs]] // CHECK-NEXT: {{%[0-9]+}} = OpImageSampleDrefImplicitLod %float [[sampledImg]] %float_1 %float_0_5 o[0] = t1.SampleCmp(scs, 1, 0.5); -} \ No newline at end of file +} From fcbd2a18429447502ecf91fa4f684d23142a8b57 Mon Sep 17 00:00:00 2001 From: Steven Perron Date: Tue, 18 Feb 2025 11:56:14 -0500 Subject: [PATCH 20/20] [SPIRV] Add the vk::image_format to runtime arrays (#7128) A case was missed when adding the `vk::image_format` to the image types. We did not handle runtime arrays. We fix that oversight. Fixes #6880 --- tools/clang/lib/SPIRV/LowerTypeVisitor.cpp | 11 +++++++++++ .../test/CodeGenSPIRV/vk.attribute.image-format.hlsl | 6 ++++++ 2 files changed, 17 insertions(+) diff --git a/tools/clang/lib/SPIRV/LowerTypeVisitor.cpp b/tools/clang/lib/SPIRV/LowerTypeVisitor.cpp index 0eb2243216..9999621d31 100644 --- a/tools/clang/lib/SPIRV/LowerTypeVisitor.cpp +++ b/tools/clang/lib/SPIRV/LowerTypeVisitor.cpp @@ -214,6 +214,17 @@ bool LowerTypeVisitor::visitInstruction(SpirvInstruction *instr) { arrayType->getStride()); instr->setResultType(resultType); } + } else if (const auto *runtimeArrayType = + dyn_cast(resultType)) { + if (const auto *imageType = + dyn_cast(runtimeArrayType->getElementType())) { + auto newImgType = spvContext.getImageType( + imageType, + vkImgFeatures.format.value_or(spv::ImageFormat::Unknown)); + resultType = spvContext.getRuntimeArrayType( + newImgType, runtimeArrayType->getStride()); + instr->setResultType(resultType); + } } } } diff --git a/tools/clang/test/CodeGenSPIRV/vk.attribute.image-format.hlsl b/tools/clang/test/CodeGenSPIRV/vk.attribute.image-format.hlsl index 4d10dc446b..12b03fffda 100644 --- a/tools/clang/test/CodeGenSPIRV/vk.attribute.image-format.hlsl +++ b/tools/clang/test/CodeGenSPIRV/vk.attribute.image-format.hlsl @@ -60,6 +60,12 @@ RWBuffer Buf_r64i; [[vk::image_format("r64ui")]] RWBuffer Buf_r64ui; +[[vk::image_format("r16f")]] +// CHECK: [[ImgType:%[0-9a-zA-Z_]+]] = OpTypeImage %float 2D 2 0 0 2 R16f +// CHECK: [[ArrayType:%[0-9a-zA-Z_]+]] = OpTypeRuntimeArray [[ImgType]] +// CHECK: [[PtrType:%[0-9a-zA-Z_]+]] = OpTypePointer UniformConstant [[ArrayType]] +RWTexture2D Buf_r16f_bindless[]; + struct S { RWBuffer b; };