Wild experiments

[mati865]

Looks like there is quite a lot of a publicity here, so I probably at least provide minimal info if somebody wants to try it out.

You can try it locally out by adding -Clinker=wild to RUSTFLAGS or with cargo +<toolchain_name> rustc -- -C linker=wild and verify whether it works by running readelf -p .comment target/debug/<binary/shared object name> | rg -i linker - it should say Wild ....
This experiment only covers x86_64-unknown-linux-gnu host for now.

[rust-bors]

☀️ Try build successful (CI)
Build commit: 8d02507 (8d02507f77424e2db34b91eaac187b532c4f3244, parent: d1ed52b1f5b78bf66127b670af813b84d57aeedb)

[rust-timer]

Finished benchmarking commit (8d02507): comparison URL.

Overall result: ✅ improvements - no action needed

Benchmarking this pull request means it may be perf-sensitive – we'll automatically label it not fit for rolling up. You can override this, but we strongly advise not to, due to possible changes in compiler perf.

@bors rollup=never
@rustbot label: -S-waiting-on-perf -perf-regression

Instruction count

Our most reliable metric. Used to determine the overall result above. However, even this metric can be noisy.

	mean	range	count
Regressions ❌ (primary)	-	-	0
Regressions ❌ (secondary)	-	-	0
Improvements ✅ (primary)	-3.2%	[-8.9%, -0.4%]	29
Improvements ✅ (secondary)	-1.9%	[-7.6%, -0.1%]	71
All ❌✅ (primary)	-3.2%	[-8.9%, -0.4%]	29

Max RSS (memory usage)

Results (primary -5.7%)

A less reliable metric. May be of interest, but not used to determine the overall result above.

	mean	range	count
Regressions ❌ (primary)	-	-	0
Regressions ❌ (secondary)	-	-	0
Improvements ✅ (primary)	-5.7%	[-12.4%, -0.6%]	3
Improvements ✅ (secondary)	-	-	0
All ❌✅ (primary)	-5.7%	[-12.4%, -0.6%]	3

Cycles

Results (primary 9.0%, secondary 10.3%)

A less reliable metric. May be of interest, but not used to determine the overall result above.

	mean	range	count
Regressions ❌ (primary)	11.0%	[2.3%, 29.2%]	11
Regressions ❌ (secondary)	11.2%	[2.2%, 31.7%]	28
Improvements ✅ (primary)	-2.2%	[-2.2%, -2.2%]	2
Improvements ✅ (secondary)	-1.9%	[-2.3%, -1.4%]	2
All ❌✅ (primary)	9.0%	[-2.2%, 29.2%]	13

Binary size

Results (primary -1.1%, secondary -1.7%)

A less reliable metric. May be of interest, but not used to determine the overall result above.

	mean	range	count
Regressions ❌ (primary)	0.9%	[0.9%, 0.9%]	4
Regressions ❌ (secondary)	1.3%	[0.8%, 1.4%]	4
Improvements ✅ (primary)	-1.4%	[-3.3%, -0.4%]	31
Improvements ✅ (secondary)	-1.8%	[-4.9%, -0.0%]	85
All ❌✅ (primary)	-1.1%	[-3.3%, 0.9%]	35

Bootstrap: 473.872s -> 517.604s (9.23%)
Artifact size: 390.16 MiB -> 390.17 MiB (0.00%)

[mati865]

Summary of Wild compared to LLD:

• without forking (Unix syscall) the linker: https://perf.rust-lang.org/compare.html?start=637b50be01093962ac6f4432d6881ab41d6d90b4&end=3a5439fb15a90156a0db2aab1a7daa0be6cdf1c7&stat=wall-time
• with fork: https://perf.rust-lang.org/compare.html?start=02c7b1a7ac1d739663878030510508372e46f254&end=cee02659a5c76c78bcae425edbfe4c97d767f214&stat=wall-time
• with fork, ThinLTO and @lqd's PR to skip excluded sections: https://perf.rust-lang.org/compare.html?start=ddaf12390d3ffb7d5ba74491a48f3cd528e5d777&end=87248e14656411e3502b645bd0deff132c90db6c&stat=wall-time
• with fork, ThinLTO, @lqd's PR to skip excluded sections, and capping Wild's threads to 8: https://perf.rust-lang.org/compare.html?start=d1ed52b1f5b78bf66127b670af813b84d57aeedb&end=8d02507f77424e2db34b91eaac187b532c4f3244&stat=wall-time&nonRelevant=true

Keep in mind those are wall-times which are prone to be noisy, so focus on the general tendencies, rather than specific cases.

[hkBst]

I was puzzled by bootstrap taking longer, but that is because it now also includes the time to build wild, right?

[mati865]

Yes, I made it a part of rustc to make the benchmarking simple. In the proper setup, it'd be an optional component that can be enabled or disabled.

[mati865]

@bors try @rust-timer queue

0.7 was released recently.

[rust-log-analyzer]

The job aarch64-gnu-llvm-20-2 failed! Check out the build log: (web) (plain enhanced) (plain)

Click to see the possible cause of the failure (guessed by this bot)

fmt: checked 6531 files
tidy check
tidy [rustdoc_json (src)]: `rustdoc-json-types` modified, checking format version
tidy: Skipping binary file check, read-only filesystem
tidy [style (compiler)]: /checkout/compiler/rustc_target/src/spec/mod.rs:131: TODO is used for tasks that should be done before merging a PR; If you want to leave a message in the codebase use FIXME
tidy [style (compiler)]: FAIL
tidy: The following check failed: style (compiler)
Command `/checkout/obj/build/aarch64-unknown-linux-gnu/stage1-tools-bin/rust-tidy /checkout /checkout/obj/build/aarch64-unknown-linux-gnu/stage0/bin/cargo /checkout/obj/build 4 npm` failed with exit code 1
Created at: src/bootstrap/src/core/build_steps/tool.rs:1650:23
Executed at: src/bootstrap/src/core/build_steps/test.rs:1279:29

Command has failed. Rerun with -v to see more details.
Bootstrap failed while executing `--stage 2 test --skip tests --skip coverage-map --skip coverage-run --skip library --skip tidyselftest`
Build completed unsuccessfully in 0:00:50
  local time: Sat Nov  8 00:30:12 UTC 2025
  network time: Sat, 08 Nov 2025 00:30:12 GMT
##[error]Process completed with exit code 1.

[rust-bors]

☀️ Try build successful (CI)
Build commit: 357fa00 (357fa000ef4c410dafcae618e021d93c13be546d, parent: 843f8ce2ebc01d35a30484eadc8a84cdc6130844)

[rust-timer]

Finished benchmarking commit (357fa00): comparison URL.

Overall result: ✅ improvements - no action needed

Benchmarking this pull request means it may be perf-sensitive – we'll automatically label it not fit for rolling up. You can override this, but we strongly advise not to, due to possible changes in compiler perf.

@bors rollup=never
@rustbot label: -S-waiting-on-perf -perf-regression

Instruction count

Our most reliable metric. Used to determine the overall result above. However, even this metric can be noisy.

	mean	range	count
Regressions ❌ (primary)	-	-	0
Regressions ❌ (secondary)	-	-	0
Improvements ✅ (primary)	-4.1%	[-11.7%, -0.4%]	29
Improvements ✅ (secondary)	-2.9%	[-10.4%, -0.2%]	69
All ❌✅ (primary)	-4.1%	[-11.7%, -0.4%]	29

Max RSS (memory usage)

Results (primary -9.2%)

A less reliable metric. May be of interest, but not used to determine the overall result above.

	mean	range	count
Regressions ❌ (primary)	-	-	0
Regressions ❌ (secondary)	-	-	0
Improvements ✅ (primary)	-9.2%	[-14.5%, -3.8%]	2
Improvements ✅ (secondary)	-	-	0
All ❌✅ (primary)	-9.2%	[-14.5%, -3.8%]	2

Cycles

Results (primary -6.2%, secondary -4.8%)

A less reliable metric. May be of interest, but not used to determine the overall result above.

	mean	range	count
Regressions ❌ (primary)	-	-	0
Regressions ❌ (secondary)	3.5%	[3.5%, 3.5%]	1
Improvements ✅ (primary)	-6.2%	[-10.4%, -2.2%]	14
Improvements ✅ (secondary)	-5.0%	[-10.2%, -2.2%]	35
All ❌✅ (primary)	-6.2%	[-10.4%, -2.2%]	14

Binary size

Results (primary -1.1%, secondary -1.6%)

A less reliable metric. May be of interest, but not used to determine the overall result above.

	mean	range	count
Regressions ❌ (primary)	0.9%	[0.9%, 0.9%]	4
Regressions ❌ (secondary)	1.3%	[0.8%, 1.4%]	4
Improvements ✅ (primary)	-1.4%	[-3.4%, -0.4%]	32
Improvements ✅ (secondary)	-1.8%	[-4.9%, -0.0%]	87
All ❌✅ (primary)	-1.1%	[-3.4%, 0.9%]	36

Bootstrap: 483.325s -> 518.997s (7.38%)
Artifact size: 390.81 MiB -> 390.83 MiB (0.01%)

[mati865]

From last time's 7% wall-time improvement to 9.7% wall-time improvement. Getting better.