Use horde's SyncTable for default query caches and CtxtInterners

[Zoxc]

View all comments

This uses the hash map from my horde crate as the default query cache map and in CtxtInterners. The main benefit of it is that it offers lock-free reads.

This is a proper version of the test in #138419.

This adds a intermediary SyncTable type which uses our dynamic Lock instead of the always-on lock integrated in horde::SyncTable, as this avoids lock overhead for the -Zthreads=1 case.

This integrates the quiescent state based memory reclamation scheme by marking query execution as quiescent with collect and calling release when we're blocking.

[Zoxc]

@bors try @rust-timer queue

[rust-bors]

☀️ Try build successful (CI)
Build commit: 810b08b (810b08be87ff4ade9ca65bf4374cb14ff0468a37, parent: 620e36a8d1fc2f9e5694ce83c2631877651a962c)

[rust-timer]

Finished benchmarking commit (810b08b): comparison URL.

Overall result: ❌✅ regressions and improvements - please read the text below

Benchmarking this pull request means it may be perf-sensitive – we'll automatically label it not fit for rolling up. You can override this, but we strongly advise not to, due to possible changes in compiler perf.

Next Steps: If you can justify the regressions found in this try perf run, please do so in sufficient writing along with @rustbot label: +perf-regression-triaged. If not, please fix the regressions and do another perf run. If its results are neutral or positive, the label will be automatically removed.

@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)	0.4%	[0.1%, 1.0%]	123
Regressions ❌ (secondary)	0.7%	[0.1%, 1.9%]	129
Improvements ✅ (primary)	-	-	0
Improvements ✅ (secondary)	-0.3%	[-0.9%, -0.1%]	12
All ❌✅ (primary)	0.4%	[0.1%, 1.0%]	123

Max RSS (memory usage)

Results (secondary -9.0%)

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

	mean	range	count
Regressions ❌ (primary)	-	-	0
Regressions ❌ (secondary)	1.3%	[0.8%, 2.0%]	4
Improvements ✅ (primary)	-	-	0
Improvements ✅ (secondary)	-11.6%	[-15.2%, -1.0%]	16
All ❌✅ (primary)	-	-	0

Cycles

Results (primary 3.5%, secondary 3.5%)

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

	mean	range	count
Regressions ❌ (primary)	3.5%	[1.4%, 7.0%]	54
Regressions ❌ (secondary)	5.0%	[1.9%, 8.9%]	60
Improvements ✅ (primary)	-	-	0
Improvements ✅ (secondary)	-5.1%	[-7.3%, -2.7%]	11
All ❌✅ (primary)	3.5%	[1.4%, 7.0%]	54

Binary size

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

Bootstrap: 483.729s -> 488.9s (1.07%)
Artifact size: 396.86 MiB -> 395.01 MiB (-0.47%)

[Zoxc]

@bors try @rust-timer queue

[rust-bors]

☀️ Try build successful (CI)
Build commit: a20b8ea (a20b8ea6796c024e8171574fe9e57407f66541d1, parent: e0a8361373c88bd791b5b1335c6cbf83cd68e4d5)

[rust-timer]

Finished benchmarking commit (a20b8ea): comparison URL.

Overall result: ❌✅ regressions and improvements - please read the text below

Benchmarking this pull request means it may be perf-sensitive – we'll automatically label it not fit for rolling up. You can override this, but we strongly advise not to, due to possible changes in compiler perf.

Next Steps: If you can justify the regressions found in this try perf run, please do so in sufficient writing along with @rustbot label: +perf-regression-triaged. If not, please fix the regressions and do another perf run. If its results are neutral or positive, the label will be automatically removed.

@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)	0.3%	[0.1%, 0.7%]	86
Regressions ❌ (secondary)	0.6%	[0.0%, 1.5%]	107
Improvements ✅ (primary)	-0.2%	[-0.3%, -0.1%]	5
Improvements ✅ (secondary)	-0.3%	[-0.9%, -0.0%]	28
All ❌✅ (primary)	0.3%	[-0.3%, 0.7%]	91

Max RSS (memory usage)

Results (primary 2.6%, secondary -9.3%)

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

	mean	range	count
Regressions ❌ (primary)	2.6%	[2.6%, 2.6%]	1
Regressions ❌ (secondary)	0.8%	[0.7%, 1.0%]	3
Improvements ✅ (primary)	-	-	0
Improvements ✅ (secondary)	-11.0%	[-15.3%, -1.0%]	18
All ❌✅ (primary)	2.6%	[2.6%, 2.6%]	1

Cycles

Results (primary 3.4%, secondary 3.0%)

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

	mean	range	count
Regressions ❌ (primary)	3.4%	[2.2%, 6.5%]	36
Regressions ❌ (secondary)	4.7%	[1.6%, 7.3%]	51
Improvements ✅ (primary)	-	-	0
Improvements ✅ (secondary)	-4.5%	[-7.1%, -2.0%]	12
All ❌✅ (primary)	3.4%	[2.2%, 6.5%]	36

Binary size

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

Bootstrap: 480.839s -> 488.485s (1.59%)
Artifact size: 396.90 MiB -> 397.07 MiB (0.05%)

[Zoxc]

@bors try @rust-timer queue

[rust-bors]

☀️ Try build successful (CI)
Build commit: fbe8224 (fbe822436088fa395ebd12435db3f96d3456d1b2, parent: 9e973d8850334be3fa3bcfb8ed2887d21485e92c)

[rust-timer]

Finished benchmarking commit (fbe8224): comparison URL.

Overall result: ❌✅ regressions and improvements - please read the text below

Benchmarking this pull request means it may be perf-sensitive – we'll automatically label it not fit for rolling up. You can override this, but we strongly advise not to, due to possible changes in compiler perf.

Next Steps: If you can justify the regressions found in this try perf run, please do so in sufficient writing along with @rustbot label: +perf-regression-triaged. If not, please fix the regressions and do another perf run. If its results are neutral or positive, the label will be automatically removed.

@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)	0.4%	[0.3%, 0.6%]	6
Regressions ❌ (secondary)	0.3%	[0.0%, 0.4%]	20
Improvements ✅ (primary)	-0.4%	[-0.7%, -0.1%]	10
Improvements ✅ (secondary)	-0.4%	[-0.9%, -0.1%]	36
All ❌✅ (primary)	-0.1%	[-0.7%, 0.6%]	16

Max RSS (memory usage)

Results (primary 0.4%, secondary -0.4%)

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

	mean	range	count
Regressions ❌ (primary)	0.4%	[0.4%, 0.4%]	1
Regressions ❌ (secondary)	1.3%	[0.9%, 1.7%]	2
Improvements ✅ (primary)	-	-	0
Improvements ✅ (secondary)	-3.7%	[-3.7%, -3.7%]	1
All ❌✅ (primary)	0.4%	[0.4%, 0.4%]	1

Cycles

Results (primary 3.0%, secondary 5.2%)

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

	mean	range	count
Regressions ❌ (primary)	3.0%	[2.6%, 3.4%]	9
Regressions ❌ (secondary)	6.4%	[2.2%, 28.6%]	20
Improvements ✅ (primary)	-	-	0
Improvements ✅ (secondary)	-2.7%	[-3.1%, -2.3%]	3
All ❌✅ (primary)	3.0%	[2.6%, 3.4%]	9

Binary size

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

Bootstrap: 484.216s -> 484.742s (0.11%)
Artifact size: 394.84 MiB -> 394.74 MiB (-0.02%)

[Zoxc]

@bors try @rust-timer queue

[rust-bors]

☀️ Try build successful (CI)
Build commit: 3ecda27 (3ecda27879be15e793711696512e5874623c0681, parent: d27207d3d64a1399f90344d6ef3c5743204b3fd3)

[rust-timer]

Finished benchmarking commit (3ecda27): comparison URL.

Overall result: ❌✅ regressions and improvements - please read the text below

Benchmarking this pull request means it may be perf-sensitive – we'll automatically label it not fit for rolling up. You can override this, but we strongly advise not to, due to possible changes in compiler perf.

Next Steps: If you can justify the regressions found in this try perf run, please do so in sufficient writing along with @rustbot label: +perf-regression-triaged. If not, please fix the regressions and do another perf run. If its results are neutral or positive, the label will be automatically removed.

@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)	0.3%	[0.1%, 0.8%]	84
Regressions ❌ (secondary)	0.6%	[0.0%, 1.9%]	108
Improvements ✅ (primary)	-0.2%	[-0.3%, -0.1%]	6
Improvements ✅ (secondary)	-0.3%	[-1.0%, -0.1%]	30
All ❌✅ (primary)	0.3%	[-0.3%, 0.8%]	90

Max RSS (memory usage)

Results (secondary -10.7%)

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

	mean	range	count
Regressions ❌ (primary)	-	-	0
Regressions ❌ (secondary)	0.9%	[0.9%, 0.9%]	1
Improvements ✅ (primary)	-	-	0
Improvements ✅ (secondary)	-11.4%	[-15.1%, -3.8%]	17
All ❌✅ (primary)	-	-	0

Cycles

Results (primary 3.2%, secondary 2.6%)

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

	mean	range	count
Regressions ❌ (primary)	3.2%	[1.4%, 5.5%]	37
Regressions ❌ (secondary)	4.6%	[1.9%, 7.9%]	47
Improvements ✅ (primary)	-	-	0
Improvements ✅ (secondary)	-4.6%	[-7.9%, -2.3%]	13
All ❌✅ (primary)	3.2%	[1.4%, 5.5%]	37

Binary size

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

Bootstrap: 481.726s -> 488.008s (1.30%)
Artifact size: 394.80 MiB -> 397.03 MiB (0.56%)

[Zoxc]

@bors try @rust-timer queue

[rust-bors]

☀️ Try build successful (CI)
Build commit: 8107a7d (8107a7d016fb925dfd99c26c0011c2c4a9039f30, parent: 5b61449ed85a670f1dd3fca6a8c759ee0b451b66)

[rust-timer]

Finished benchmarking commit (8107a7d): comparison URL.

Overall result: ❌✅ regressions and improvements - please read the text below

Benchmarking this pull request means it may be perf-sensitive – we'll automatically label it not fit for rolling up. You can override this, but we strongly advise not to, due to possible changes in compiler perf.

Next Steps: If you can justify the regressions found in this try perf run, please do so in sufficient writing along with @rustbot label: +perf-regression-triaged. If not, please fix the regressions and do another perf run. If its results are neutral or positive, the label will be automatically removed.

@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)	0.4%	[0.2%, 0.7%]	6
Regressions ❌ (secondary)	0.4%	[0.0%, 0.9%]	18
Improvements ✅ (primary)	-0.3%	[-0.8%, -0.1%]	55
Improvements ✅ (secondary)	-0.4%	[-0.8%, -0.1%]	82
All ❌✅ (primary)	-0.2%	[-0.8%, 0.7%]	61

Max RSS (memory usage)

Results (primary 0.5%, secondary -10.6%)

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

	mean	range	count
Regressions ❌ (primary)	0.5%	[0.5%, 0.5%]	1
Regressions ❌ (secondary)	1.1%	[1.1%, 1.1%]	1
Improvements ✅ (primary)	-	-	0
Improvements ✅ (secondary)	-11.3%	[-14.6%, -2.4%]	16
All ❌✅ (primary)	0.5%	[0.5%, 0.5%]	1

Cycles

Results (primary 3.6%, secondary 2.2%)

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

	mean	range	count
Regressions ❌ (primary)	3.6%	[2.2%, 6.4%]	14
Regressions ❌ (secondary)	4.6%	[2.0%, 6.7%]	22
Improvements ✅ (primary)	-	-	0
Improvements ✅ (secondary)	-3.8%	[-5.0%, -2.2%]	9
All ❌✅ (primary)	3.6%	[2.2%, 6.4%]	14

Binary size

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

Bootstrap: 481.08s -> 485.518s (0.92%)
Artifact size: 396.88 MiB -> 395.54 MiB (-0.34%)