chore(l2): build SP1 with recommended settings

[JereSalo]

Motivation

• Improve perf with SP1

Description

Based on SP1 Docs we should compile with lto thin, true or fat and codegen-units 1.

Ethrex Replay Build Command: cargo build --release --profile sp1,profiling

Branch	Empty	1 Transfer	100 Transfers	23897700 (30M)	23897687 (33M)	23897646 (55M)	23897709 (60M)
main (Gas)	6,395,981	7,562,712	45,607,202	992,588,338	1,636,706,753	2,403,112,555	1,592,397,865
this branch (Gas)	5,780,387	7,342,824	43,493,602	967,103,357	1,605,060,723	2,355,045,830	1,548,978,533
% improvement	9.62%	2.91%	4.63%	2.57%	1.93%	2.00%	2.73%
main (Cycles)	1,117,654	1,402,868	20,150,510	723,222,391	1,325,458,588	1,852,238,963	1,209,370,294
this branch (Cycles)	1,019,876	1,296,104	17,718,398	692,676,359	1,289,227,624	1,796,213,076	1,158,792,828
% improvement (Cycles)	8.75%	7.61%	12.07%	4.22%	2.73%	3.02%	4.18%

[github-actions]

Lines of code report

Total lines added: 4
Total lines removed: 0
Total lines changed: 4

Detailed view

+----------------------------------------------------+-------+------+
| File                                               | Lines | Diff |
+----------------------------------------------------+-------+------+
| ethrex/crates/l2/prover/src/guest_program/build.rs | 197   | +4   |
+----------------------------------------------------+-------+------+

[Copilot]

Pull request overview

This PR introduces performance optimizations for zkVM builds by configuring the Cargo release profile with recommended compiler settings from the SP1 documentation. The changes set lto = "thin" and codegen-units = 1 to improve execution performance, resulting in 2-10% gas and cycle improvements across various test scenarios.

• Added a helper function to configure Cargo profile environment variables for LTO and codegen optimization
• Applied these settings consistently across all four zkVM backends (RISC0, SP1, Zisk, and OpenVM)

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

You can also share your feedback on Copilot code review for a chance to win a $100 gift card. Take the survey.

[github-actions]

Benchmark Block Execution Results Comparison Against Main

Command	Mean [s]	Min [s]	Max [s]	Relative
base	61.183 ± 0.338	60.908	61.972	1.01 ± 0.01
head	60.837 ± 0.124	60.692	61.044	1.00