WIP: vendor rustc-hash and override str/length_prefix methods

[Mark-Simulacrum]

Locally avoiding length and str sentinel hashing adds 2,611,200 cases of hashes with just one add_to_hash call (out of roughly 43,063,296 total cases). Unclear what kind of effect this will have on runtime performance at this time, but seems like it's worth trying!

cc #56308

[rustbot]

@Mark-Simulacrum: no appropriate reviewer found, use r? to override

[rustbot]

These commits modify the Cargo.lock file. Unintentional changes to Cargo.lock can be introduced when switching branches and rebasing PRs.

If this was unintentional then you should revert the changes before this PR is merged.
Otherwise, you can ignore this comment.

[Mark-Simulacrum]

@bors try @rust-timer queue

cc #96762 -- probably worthwhile data for that tracking issue, regardless of whether positive or negative.

[bors]

⌛ Trying commit c9e83b6 with merge 5f68de8...

[bors]

💔 Test failed - checks-actions

[Mark-Simulacrum]

@bors try

[bors]

⌛ Trying commit b24c62d with merge 256f15b...

[bors]

☀️ Try build successful - checks-actions
Build commit: 256f15b (256f15b530c3f5b8e4ef8e236d07432bdda9e3fc)

[rust-timer]

Finished benchmarking commit (256f15b): comparison URL.

Overall result: ❌ regressions - ACTION NEEDED

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 may lead 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-perf +perf-regression

Instruction count

This is a highly reliable metric that was used to determine the overall result at the top of this comment.

	mean	range	count
Regressions ❌ (primary)	5.2%	[1.5%, 9.9%]	13
Regressions ❌ (secondary)	14.3%	[0.5%, 25.8%]	15
Improvements ✅ (primary)	-	-	0
Improvements ✅ (secondary)	-	-	0
All ❌✅ (primary)	5.2%	[1.5%, 9.9%]	13

Max RSS (memory usage)

This benchmark run did not return any relevant results for this metric.

Cycles

Results

This 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.

	mean	range	count
Regressions ❌ (primary)	5.9%	[1.8%, 12.5%]	13
Regressions ❌ (secondary)	16.0%	[2.1%, 30.7%]	16
Improvements ✅ (primary)	-	-	0
Improvements ✅ (secondary)	-	-	0
All ❌✅ (primary)	5.9%	[1.8%, 12.5%]	13

Binary size

This benchmark run did not return any relevant results for this metric.

Bootstrap: 663.454s -> 664.246s (0.12%)
Artifact size: 308.30 MiB -> 308.30 MiB (-0.00%)

[rust-log-analyzer]

The job mingw-check failed! Check out the build log: (web) (plain)

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

GITHUB_ENV=/home/runner/work/_temp/_runner_file_commands/set_env_58a9cf6e-7d5b-44ad-bb9f-7b1a2addf92c
GITHUB_EVENT_NAME=pull_request
GITHUB_EVENT_PATH=/home/runner/work/_temp/_github_workflow/event.json
GITHUB_GRAPHQL_URL=https://api.github.com/graphql
GITHUB_HEAD_REF=skip-hashers
GITHUB_JOB=pr
GITHUB_PATH=/home/runner/work/_temp/_runner_file_commands/add_path_58a9cf6e-7d5b-44ad-bb9f-7b1a2addf92c
GITHUB_REF=refs/pull/120078/merge
GITHUB_REF_NAME=120078/merge
GITHUB_REF_PROTECTED=false
---
* Read errors: 0
* Files with copyright information: 179930 / 179937
* Files with license information: 179929 / 179937

Unfortunately, your project is not compliant with version 3.0 of the REUSE Specification :-(
  network time: Sat, 20 Jan 2024 01:20:16 GMT
##[error]Process completed with exit code 1.
Post job cleanup.

[Mark-Simulacrum]

@bors try @rust-timer queue

It looks like we have pretty terrible hash quality for strings -- looking locally, e.g., libcore has ~70k Symbols so roughly speaking the ~17 lower bits of the hash function are getting used. For simplicity if we take the lower 16 bits we can see there are 41.7k unique values when hashing in 0xFF and 39.9k unique values when just hashing the bytes of the string. My guess is that final 0xFF hash helps smear/distribute short tails which otherwise dominate the hash's value too much.

For now I'm going to re-run with 0xFF added back in but I think we should consider whether hashing long byte strings is actually working well enough -- and maybe a different hash function would be better for maps where we have non (roughly) u32 integers as keys. The two benchmarks that regressed here have similarly not-great performance for the hash function, which is why they're regressing so much.

[bors]

⌛ Trying commit 6ca3c44 with merge f7f7717...

[bors]

☀️ Try build successful - checks-actions
Build commit: f7f7717 (f7f7717d4e9135238fdf8784d6487d1380c6d91d)

[rust-timer]

Finished benchmarking commit (f7f7717): comparison URL.

Overall result: no relevant changes - no action needed

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 may lead to changes in compiler perf.

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

Instruction count

This benchmark run did not return any relevant results for this metric.

Max RSS (memory usage)

Results

This 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.

	mean	range	count
Regressions ❌ (primary)	1.6%	[1.6%, 1.6%]	1
Regressions ❌ (secondary)	2.9%	[2.9%, 2.9%]	2
Improvements ✅ (primary)	-	-	0
Improvements ✅ (secondary)	-4.6%	[-4.6%, -4.6%]	1
All ❌✅ (primary)	1.6%	[1.6%, 1.6%]	1

Cycles

Results

This 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.

	mean	range	count
Regressions ❌ (primary)	1.5%	[1.5%, 1.5%]	1
Regressions ❌ (secondary)	-	-	0
Improvements ✅ (primary)	-1.2%	[-1.2%, -1.2%]	1
Improvements ✅ (secondary)	-	-	0
All ❌✅ (primary)	0.2%	[-1.2%, 1.5%]	2

Binary size

Results

This 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.

	mean	range	count
Regressions ❌ (primary)	0.5%	[0.5%, 0.5%]	1
Regressions ❌ (secondary)	-	-	0
Improvements ✅ (primary)	-	-	0
Improvements ✅ (secondary)	-	-	0
All ❌✅ (primary)	0.5%	[0.5%, 0.5%]	1

Bootstrap: 663.584s -> 666.051s (0.37%)
Artifact size: 308.31 MiB -> 308.33 MiB (0.01%)

[Mark-Simulacrum]

Seems like this didn't really pan out directly.