[WIP] MIR move elimination pass

[Amanieu]

View all comments

This is very much a work in progress, and will sometimes generate broken/unsound MIR.

[Amanieu]

@bors try @rust-timer queue

[rust-bors]

💔 Test for 884fdb2 failed: CI. Failed job:

• try - dist-x86_64-linux-quick (web logs, enhanced plaintext logs)

[Amanieu]

@bors try

[rust-bors]

💔 Test for b42852b failed: CI. Failed job:

• try - dist-x86_64-linux-quick (web logs, enhanced plaintext logs)

[Amanieu]

@bors try

[rust-bors]

💔 Test for 8165e10 failed: CI. Failed job:

• try - dist-x86_64-linux-quick (web logs, enhanced plaintext logs)

[Amanieu]

@bors try

[rust-bors]

☀️ Try build successful (CI)
Build commit: 6c8a539 (6c8a539b5dff4f4f443a5c0c4ed57749516e3395, parent: ad3a598ca4bc7c68bcbbce3e0d3be9a7618df190)

[rust-bors]

💔 Test for f4741d9 failed: CI. Failed job:

• try - dist-x86_64-linux-quick (web logs, enhanced plaintext logs)

[Amanieu]

@bors try

[rust-bors]

☀️ Try build successful (CI)
Build commit: 4eca0f3 (4eca0f31a0f5e100354734a70c8023dd16a08796, parent: 9128490ab32b5ab37a4bb9f6667a9ce5d32924c0)

[rust-timer]

Finished benchmarking commit (4eca0f3): comparison URL.

Overall result: ❌✅ regressions and improvements - please read:

Benchmarking means the PR may be perf-sensitive. It's automatically marked not fit for rolling up. Overriding is possible but disadvised: it risks changing compiler perf.

Next, please: If you can, justify the regressions found in this try perf run in writing along with @rustbot label: +perf-regression-triaged. If not, fix the regressions and do another perf run. Neutral or positive results will clear the label automatically.

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

Instruction count

Our most reliable metric. Used to determine the overall result above. However, even this metric can be noisy.

	mean	range	count
Regressions ❌ (primary)	2.5%	[0.2%, 10.9%]	24
Regressions ❌ (secondary)	14.9%	[0.0%, 762.2%]	59
Improvements ✅ (primary)	-1.0%	[-6.0%, -0.2%]	214
Improvements ✅ (secondary)	-1.2%	[-13.3%, -0.1%]	183
All ❌✅ (primary)	-0.7%	[-6.0%, 10.9%]	238

Max RSS (memory usage)

Results (primary 0.8%, secondary -2.5%)

A less reliable metric. May be of interest, but not used to determine the overall result above.

	mean	range	count
Regressions ❌ (primary)	2.5%	[2.2%, 2.9%]	5
Regressions ❌ (secondary)	0.4%	[0.4%, 0.4%]	1
Improvements ✅ (primary)	-3.5%	[-6.5%, -0.4%]	2
Improvements ✅ (secondary)	-2.8%	[-3.7%, -1.8%]	12
All ❌✅ (primary)	0.8%	[-6.5%, 2.9%]	7

Cycles

Results (primary -2.5%, secondary 0.1%)

A less reliable metric. May be of interest, but not used to determine the overall result above.

	mean	range	count
Regressions ❌ (primary)	3.9%	[1.7%, 7.0%]	5
Regressions ❌ (secondary)	64.2%	[5.5%, 282.4%]	5
Improvements ✅ (primary)	-3.0%	[-5.6%, -1.7%]	62
Improvements ✅ (secondary)	-5.0%	[-20.4%, -1.5%]	63
All ❌✅ (primary)	-2.5%	[-5.6%, 7.0%]	67

Binary size

Results (primary 0.4%, secondary 1.4%)

A less reliable metric. May be of interest, but not used to determine the overall result above.

	mean	range	count
Regressions ❌ (primary)	1.5%	[0.0%, 6.6%]	82
Regressions ❌ (secondary)	2.2%	[0.0%, 6.6%]	94
Improvements ✅ (primary)	-1.0%	[-3.5%, -0.1%]	60
Improvements ✅ (secondary)	-1.1%	[-5.6%, -0.1%]	28
All ❌✅ (primary)	0.4%	[-3.5%, 6.6%]	142

Bootstrap: 517.203s -> 494.695s (-4.35%)
Artifact size: 400.60 MiB -> 397.22 MiB (-0.84%)

[Amanieu]

@bors try @rust-timer queue

[rust-log-analyzer]

The job x86_64-gnu-gcc-core-tests failed! Check out the build log: (web) (plain enhanced) (plain)

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

For more information how to resolve CI failures of this job, visit this link.

[rust-bors]

☀️ Try build successful (CI)
Build commit: 9d211b2 (9d211b2625f34105f9580690ab55479dbcb1b9d0, parent: 6368fd52cb9f230dfb156097625993e7a8891800)

[rust-timer]

Finished benchmarking commit (9d211b2): comparison URL.

Overall result: ❌✅ regressions and improvements - please read:

Benchmarking means the PR may be perf-sensitive. It's automatically marked not fit for rolling up. Overriding is possible but disadvised: it risks changing compiler perf.

Next, please: If you can, justify the regressions found in this try perf run in writing along with @rustbot label: +perf-regression-triaged. If not, fix the regressions and do another perf run. Neutral or positive results will clear the label automatically.

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

Instruction count

Our most reliable metric. Used to determine the overall result above. However, even this metric can be noisy.

	mean	range	count
Regressions ❌ (primary)	2.0%	[0.2%, 13.8%]	23
Regressions ❌ (secondary)	15.4%	[0.0%, 761.3%]	56
Improvements ✅ (primary)	-1.0%	[-6.0%, -0.2%]	218
Improvements ✅ (secondary)	-1.2%	[-12.7%, -0.0%]	185
All ❌✅ (primary)	-0.7%	[-6.0%, 13.8%]	241

Max RSS (memory usage)

Results (primary 0.3%, secondary -2.0%)

A less reliable metric. May be of interest, but not used to determine the overall result above.

	mean	range	count
Regressions ❌ (primary)	3.0%	[1.6%, 4.1%]	4
Regressions ❌ (secondary)	1.0%	[1.0%, 1.0%]	1
Improvements ✅ (primary)	-1.3%	[-1.9%, -1.0%]	7
Improvements ✅ (secondary)	-2.4%	[-3.2%, -2.1%]	9
All ❌✅ (primary)	0.3%	[-1.9%, 4.1%]	11

Cycles

Results (primary -2.7%, secondary 0.3%)

A less reliable metric. May be of interest, but not used to determine the overall result above.

	mean	range	count
Regressions ❌ (primary)	5.3%	[1.8%, 9.6%]	4
Regressions ❌ (secondary)	32.4%	[2.1%, 284.8%]	11
Improvements ✅ (primary)	-3.3%	[-6.4%, -1.8%]	56
Improvements ✅ (secondary)	-5.1%	[-20.1%, -1.1%]	65
All ❌✅ (primary)	-2.7%	[-6.4%, 9.6%]	60

Binary size

Results (primary 0.5%, secondary 1.6%)

A less reliable metric. May be of interest, but not used to determine the overall result above.

	mean	range	count
Regressions ❌ (primary)	1.7%	[0.1%, 6.7%]	74
Regressions ❌ (secondary)	2.3%	[0.0%, 6.7%]	90
Improvements ✅ (primary)	-1.0%	[-2.7%, -0.2%]	55
Improvements ✅ (secondary)	-1.2%	[-5.6%, -0.2%]	25
All ❌✅ (primary)	0.5%	[-2.7%, 6.7%]	129

Bootstrap: 511.654s -> 500.763s (-2.13%)
Artifact size: 400.70 MiB -> 397.37 MiB (-0.83%)