[perftest] Update bootstrap LLVM

[Kobzol]

This PR updates the LLVM used to build dist artifacts on CI to version 14. The goal is to see how does this change affect perf (if at all) to isolate the effect of using BOLT.

Could someone schedule a perf. run please? Thanks.

[rust-highfive]

r? @pietroalbini

(rust-highfive has picked a reviewer for you, use r? to override)

[Mark-Simulacrum]

@bors try @rust-timer queue

[rust-timer]

Awaiting bors try build completion.

@rustbot label: +S-waiting-on-perf

[bors]

⌛ Trying commit ba7b6d242c346b7d79168cdac7824e50cba3e39c with merge e9fcc3de4de81bcdce8ff4ab01ae8ad97fc23684...

[bors]

☀️ Try build successful - checks-actions
Build commit: e9fcc3de4de81bcdce8ff4ab01ae8ad97fc23684 (e9fcc3de4de81bcdce8ff4ab01ae8ad97fc23684)

[rust-timer]

Queued e9fcc3de4de81bcdce8ff4ab01ae8ad97fc23684 with parent a64180f, future comparison URL.

[rust-timer]

Finished benchmarking commit (e9fcc3de4de81bcdce8ff4ab01ae8ad97fc23684): comparison url.

Summary: This benchmark run did not return any relevant results. 34 results were found to be statistically significant but too small to be relevant.

If you disagree with this performance assessment, please file an issue in rust-lang/rustc-perf.

Benchmarking this pull request likely means that it is perf-sensitive, so we're automatically marking it as not fit for rolling up. While you can manually mark this PR as fit for rollup, we strongly recommend not doing so since this PR led to changes in compiler perf.

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

[mati865]

Looks like half of max-rss improvement comes from upgrading bootstrap LLVM, other metrics are basically a noise.

[Mark-Simulacrum]

From this PR, but this PR also includes a gcc version bump, and I think we might be using the gcc standard library for something, so it might be due to that. (IOW, I am not sure it is down to LLVM 13 vs. 14 difference).

[Kobzol]

Well the newer GCC is needed anyway to build LLVM 14. Could there be some issues with upgrading GCC?

It looks like BOLT has a nontrivial effect on its own, but it's good to know that LLVM/GCC upgrade also helps.

[Kobzol]

This PR is not supposed to be merged, so I'll use it for more LLVM CI experiments. Let's see if gathering more profiling data for PGO for LLVM helps.

[lqd]

@bors try @rust-timer queue

[rust-timer]

Awaiting bors try build completion.

@rustbot label: +S-waiting-on-perf

[bors]

⌛ Trying commit 03faa8a with merge abf097afa10dde1aa3d8a1d422724a46aab79bf3...

[bors]

☀️ Try build successful - checks-actions
Build commit: abf097afa10dde1aa3d8a1d422724a46aab79bf3 (abf097afa10dde1aa3d8a1d422724a46aab79bf3)

[rust-timer]

Queued abf097afa10dde1aa3d8a1d422724a46aab79bf3 with parent 2631aee, future comparison URL.

[rust-timer]

Finished benchmarking commit (abf097afa10dde1aa3d8a1d422724a46aab79bf3): comparison url.

Summary: This benchmark run shows 75 relevant improvements 🎉 to instruction counts.

• Arithmetic mean of relevant improvements: -2.6%
• Largest improvement in instruction counts: -5.0% on incr-full builds of webrender debug

If you disagree with this performance assessment, please file an issue in rust-lang/rustc-perf.

Benchmarking this pull request likely means that it is perf-sensitive, so we're automatically marking it as not fit for rolling up. While you can manually mark this PR as fit for rollup, we strongly recommend not doing so since this PR led to changes in compiler perf.

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

[Kobzol]

Closing in favour of #95171.