-
Notifications
You must be signed in to change notification settings - Fork 13.2k
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.
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
[DO NOT MERGE] default x64 linux target to x64-v2 for perf test #95302
Conversation
@bors try @rust-timer queue |
Awaiting bors try build completion. @rustbot label: +S-waiting-on-perf |
⌛ Trying commit 5c078dc3ac0f6dd9a136d2241fcc4ba24238c565 with merge c7e2a3ab68a1894d6a01a41078454ee7c52e61e3... |
☀️ Try build successful - checks-actions |
Queued c7e2a3ab68a1894d6a01a41078454ee7c52e61e3 with parent 8a0c550, future comparison URL. |
Finished benchmarking commit (c7e2a3ab68a1894d6a01a41078454ee7c52e61e3): comparison url. Summary: This benchmark run shows 164 relevant improvements 🎉 to instruction counts.
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 |
Even if we somehow were able to bump to v2 minimum or distribute 2 x64 versions, it would not super interesting by itself. It could help with using SIMD in LEB128 encoding/decoding, or parsing (via pshufb), etc but there are no worthwhile perf gains due to the higher micro arch level itself (which makes sense, they should come mostly from autovectorization in this context, and that would start having noticeable improvements with the higher lane counts of AVX). Might as well have up-to-date x64-v3 results, on the new benchmarks and new LLVM, so: |
Awaiting bors try build completion. @rustbot label: +S-waiting-on-perf |
⌛ Trying commit 3598f3bc2a09fbbb5272d0a4f51b42723ab7e7ae with merge 0c6a6f5061902626d38e28f946782927a6cc3897... |
☀️ Try build successful - checks-actions |
Queued 0c6a6f5061902626d38e28f946782927a6cc3897 with parent 185a3f0, future comparison URL. |
Finished benchmarking commit (0c6a6f5061902626d38e28f946782927a6cc3897): comparison url. Summary: This benchmark run shows 448 relevant improvements 🎉 to instruction counts.
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 |
cc @nnethercote |
So v2 has lots of 0.3% to 0.5% wins, while v3 has lots of 1%-3.5% wins. |
We'd like to have up-to-date numbers for this after the recent LTO changes, so I'll reopen this for a new perf run soon. |
@bors try @rust-timer queue |
Awaiting bors try build completion. @rustbot label: +S-waiting-on-perf |
⌛ Trying commit 39e3c52 with merge 18256ff35e7885ae108986756552f14a16d240d0... |
☀️ Try build successful - checks-actions |
Queued 18256ff35e7885ae108986756552f14a16d240d0 with parent 2efb0cd, future comparison URL. |
Finished benchmarking commit (18256ff35e7885ae108986756552f14a16d240d0): comparison URL. Overall result: ✅ improvements - no action neededBenchmarking 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 may lead to changes in compiler perf. @bors rollup=never Instruction countThis is a highly reliable metric that was used to determine the overall result at the top of this comment.
Max RSS (memory usage)ResultsThis is a less reliable metric that may be of interest but was not used to determine the overall result at the top of this comment.
CyclesResultsThis is a less reliable metric that may be of interest but was not used to determine the overall result at the top of this comment.
|
We should have the data we need, so I'll close this one again. |
Another perf test like #90440, but on x64-v2.
r? @ghost