Skip to content

[AArch64] Math.log is 10% slower than StrictMath.log on aarch64 #28306

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Draft
wants to merge 2 commits into
base: master
Choose a base branch
from
+429 −4
Draft

[AArch64] Math.log is 10% slower than StrictMath.log on aarch64 #28306

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
2b46003
[AArch64] Math.log is 10% slower than StrictMath.log on macosx-aarch64
dhanalla Nov 13, 2025
5d5518d
[AArch64] Math.log is 10% slower than StrictMath.log on macosx-aarch64
dhanalla Nov 13, 2025
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
12 changes: 12 additions & 0 deletions src/hotspot/cpu/aarch64/assembler_aarch64.hpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -2962,6 +2962,18 @@ template<typename R, typename... Rx>
rf(Vn, 5), rf(Vd, 0);
}

// frecps - Floating-point Reciprocal Step
void frecps(FloatRegister Vd, FloatRegister Vn, FloatRegister Vm, SIMD_RegVariant type) {
assert(type == D || type == S, "Wrong type for frecps");
starti;
f(0b010111100, 31, 23),
f(type == D ? 1 : 0, 22);
f(1, 21);
rf(Vm, 16);
f(0b111111, 15, 10);
rf(Vn, 5), rf(Vd, 0);
}

// (long) {a, b} -> (a + b)
void addpd(FloatRegister Vd, FloatRegister Vn) {
starti;
Expand Down
14 changes: 10 additions & 4 deletions src/hotspot/cpu/aarch64/c1_LIRGenerator_aarch64.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -832,12 +832,18 @@ void LIRGenerator::do_LibmIntrinsic(Intrinsic* x) {
}
break;
case vmIntrinsics::_dlog:
// Math.log intrinsic is not implemented on AArch64 (see JDK-8210858),
// but we can still call the shared runtime.
__ call_runtime_leaf(CAST_FROM_FN_PTR(address, SharedRuntime::dlog), getThreadTemp(), result_reg, cc->args());
if (StubRoutines::dlog() != nullptr) {
__ call_runtime_leaf(StubRoutines::dlog(), getThreadTemp(), result_reg, cc->args());
} else {
__ call_runtime_leaf(CAST_FROM_FN_PTR(address, SharedRuntime::dlog), getThreadTemp(), result_reg, cc->args());
}
break;
case vmIntrinsics::_dlog10:
__ call_runtime_leaf(CAST_FROM_FN_PTR(address, SharedRuntime::dlog10), getThreadTemp(), result_reg, cc->args());
if (StubRoutines::dlog10() != nullptr) {
__ call_runtime_leaf(StubRoutines::dlog10(), getThreadTemp(), result_reg, cc->args());
} else {
__ call_runtime_leaf(CAST_FROM_FN_PTR(address, SharedRuntime::dlog10), getThreadTemp(), result_reg, cc->args());
}
break;
case vmIntrinsics::_dpow:
if (StubRoutines::dpow() != nullptr) {
Expand Down
6 changes: 6 additions & 0 deletions src/hotspot/cpu/aarch64/macroAssembler_aarch64.hpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -1542,6 +1542,12 @@ class MacroAssembler: public Assembler {
FloatRegister vtmp2, FloatRegister vtmp3,
FloatRegister vtmp4, FloatRegister vtmp5);

void fast_log(FloatRegister vtmp0, FloatRegister vtmp1, FloatRegister vtmp2,
FloatRegister vtmp3, FloatRegister vtmp4, FloatRegister vtmp5,
FloatRegister tmpC1, FloatRegister tmpC2, FloatRegister tmpC3,
FloatRegister tmpC4, Register tmp1, Register tmp2,
Register tmp3, Register tmp4, Register tmp5);

void generate_dsin_dcos(bool isCos, address npio2_hw, address two_over_pi,
address pio2, address dsin_coef, address dcos_coef);
private:
Expand Down
Loading