[do not merge] Hashbrown code size optimization #87233
Conversation
rust-highfive
added
the
S-waiting-on-review
|
@bors try @rust-timer queue |
|
Awaiting bors try build completion. @rustbot label: +S-waiting-on-perf |
rustbot
added
the
S-waiting-on-perf
|
⌛ Trying commit f081421 with merge 228e3a77bfa70ec4f62f8a025c9063885b0c329a... |
|
☀️ Try build successful - checks-actions |
|
Queued 228e3a77bfa70ec4f62f8a025c9063885b0c329a with parent c7331d6, future comparison URL. |
|
Finished benchmarking try commit (228e3a77bfa70ec4f62f8a025c9063885b0c329a): comparison url. Summary: This change led to significant mixed results 🤷 in compiler performance.
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. Next Steps: If you can justify the regressions found in this perf run, please indicate this with @bors rollup=never |
rustbot
added
perf-regression
Inline small functions This adds `#[inline]` to small functions which should be beneficial to inline. rustc compilation performance (code size of `rustc_driver` up by 0.09%): ``` clap:check 1.9486s 1.9416s -0.36% hashmap-instances:check 0.0629s 0.0626s -0.52% helloworld:check 0.0443s 0.0439s -0.69% hyper:check 0.3011s 0.3000s -0.36% regex:check 1.1505s 1.1468s -0.33% syn:check 1.6989s 1.6904s -0.50% syntex_syntax:check 6.8479s 6.8288s -0.28% winapi:check 8.3437s 8.2967s -0.56% Total 20.3979s 20.3108s -0.43% Summary 4.0000s 3.9820s -0.45% ``` This is the effect on compile time this has on my [HashMap compile time benchmark](#277 (comment)): ``` hashmap-instances:check 0.0635s 0.0632s -0.33% hashmap-instances:release 32.0928s 32.4440s +1.09% hashmap-instances:debug 7.2193s 7.2800s +0.84% Total 39.3756s 39.7873s +1.05% Summary 1.5000s 1.5080s +0.54% ``` We saw a 1.6% improvement in rustc's build time for a -20% improvement on `hashmap-instances:release` on rust-lang/rust#87233. So I would expect around a 0.08% regression for rustc's build time from this PR.
Let's see if the 20% reduction in compile time for the hashbrown test crate with my
size-optbranch translates into any improvements to rustc's compile time.r? @Amanieu