Optimize Eq and Hash for Path/PathBuf

[the8472]

# new

test path::tests::bench_hash_path_long                            ... bench:          86 ns/iter (+/- 1)
test path::tests::bench_hash_path_short                           ... bench:          13 ns/iter (+/- 1)
test path::tests::bench_path_hashset                              ... bench:         197 ns/iter (+/- 6)
test path::tests::bench_path_hashset_miss                         ... bench:          94 ns/iter (+/- 4)

# old

test path::tests::bench_hash_path_long                            ... bench:         192 ns/iter (+/- 2)
test path::tests::bench_hash_path_short                           ... bench:          33 ns/iter (+/- 1)
test path::tests::bench_path_hashset                              ... bench:       1,121 ns/iter (+/- 24)
test path::tests::bench_path_hashset_miss                         ... bench:         273 ns/iter (+/- 6)

[rust-highfive]

r? @Mark-Simulacrum

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

[the8472]

@bors try @rust-timer queue

[rust-timer]

Awaiting bors try build completion.

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

[bors]

⌛ Trying commit c3a8b341d5d494891381ed81db3527b0362549c4 with merge eb2afd0294bbfeb7aa214aeacedd15e1c1351dad...

[bors]

☀️ Try build successful - checks-actions
Build commit: eb2afd0294bbfeb7aa214aeacedd15e1c1351dad (eb2afd0294bbfeb7aa214aeacedd15e1c1351dad)

[rust-timer]

Queued eb2afd0294bbfeb7aa214aeacedd15e1c1351dad with parent d22dd65, future comparison URL.

[rust-timer]

Finished benchmarking commit (eb2afd0294bbfeb7aa214aeacedd15e1c1351dad): comparison url.

Summary: This change led to large relevant mixed results 🤷 in compiler performance.

• Large improvement in instruction counts (up to -2.3% on incr-unchanged builds of helloworld)
• Large regression in instruction counts (up to 4.8% on incr-patched: println builds of webrender-wrench)

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 try perf run, please indicate this with @rustbot label: +perf-regression-triaged along with sufficient written justification. If you cannot justify the regressions please fix the regressions and do another perf run. If the next run shows neutral or positive results, the label will be automatically removed.

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

[the8472]

Almost entirely improvements except for webrender-wrench - opt - incr-patched: println which spends more time in llvm. I guess that it makes use of a lot of paths and has to spend more time to optimize it, likely trading compile time for runtime in that case. The patched file does make use of PathBuf.

[joshtriplett]

This seems like a reasonable set of performance numbers to me; it's paying compile time to get more runtime performance.

[Mark-Simulacrum]

OK, this looks good to me. I think the commit history can be cleaned up a little -- 8 commits seems a little too many :)

But otherwise r=me.

[bors]

⌛ Testing commit a6e0aa2 with merge 2ec3ee22b5808ab9a312c46fff596a7789f58752...

[bors]

💔 Test failed - checks-actions

[the8472]

Apparently the drive prefixes are the only thing where we do case-insensitive comparison on paths. 😮‍💨
Not sure yet how to solve that efficiently.

[ChrisDenton]

Apparently the drive prefixes are the only thing where we do case-insensitive comparison on paths.

They are the only place that are guaranteed to be ASCII case-insensitive. Other parts of the path could be case-sensitive.

[the8472]

@Mark-Simulacrum ready for another review, the latest commit should fix the windows failure.

[Mark-Simulacrum]

@bors r=Mark-Simulacrum rollup=never

[bors]

📌 Commit c1ea7bd has been approved by Mark-Simulacrum

[bors]

⌛ Testing commit c1ea7bd with merge c8e9497...

[bors]

☀️ Test successful - checks-actions
Approved by: Mark-Simulacrum
Pushing c8e9497 to master...

[rust-timer]

Finished benchmarking commit (c8e9497): comparison url.

Summary: This change led to very large relevant mixed results 🤷 in compiler performance.

• Very large improvement in instruction counts (up to -6.0% on incr-unchanged builds of coercions)
• Very large regression in instruction counts (up to 12.9% on incr-unchanged builds of tuple-stress)

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

Next Steps: If you can justify the regressions found in this perf run, please indicate this with @rustbot label: +perf-regression-triaged along with sufficient written justification. If you cannot justify the regressions please open an issue or create a new PR that fixes the regressions, add a comment linking to the newly created issue or PR, and then add the perf-regression-triaged label to this PR.

@rustbot label: +perf-regression

[the8472]

Comparing the call graphs of perf profiles before and after this PR for tuple-stress shows a big difference in the callees of rustc_query_system::query::plumbing::try_load_from_disk_and_cache_in_memory::<rustc_query_impl::plumbing::QueryCtxt, rustc_span::def_id::DefId, &rustc_middle::mir::Body>.

Before the PR:

• <rustc_query_impl::on_disk_cache::OnDiskCache>::try_load_query_result::<&rustc_middle::mir::Body>
  (7.8e9 instructions)

After the PR:

• <rustc_query_impl::on_disk_cache::OnDiskCache>::try_load_query_result::<&rustc_middle::mir::Body>
  (7.8e9 instructions)
• rustc_query_system::query::plumbing::incremental_verify_ich::<rustc_query_impl::plumbing::QueryCtxt, rustc_span::def_id::DefId, &rustc_middle::mir::Body>
  (6.3e9 instructions)

In other words changes to hashing triggered probabilistic incremental verification (#90361)