> graphs, compare, dashboard, bootstrap, status, help.

Comparing cpu-clock between #129115 and #128982

#129115 2024-08-18 (23f762d3)
#128982 2024-08-18 (79f5c160)
🔎 compare commits
Do another comparison ▶
Commits
Metric
Compile-time
RangeMean
[ -2.22%,  +0.56%]
-0.34%
Runtime
No results
Bootstrap
749.938 -> 750.626
0.7s (0.092%)
Artifact size
339.18 MiB -> 339.15 MiB
-28.69 KiB (-0.008%)
Select metric:
Instructions
Cycles
Wall time
Max RSS
Binary size
Filters ▶
Filter
Profiles
? The different compilation profiles (check, debug, opt, doc).
  • ? Check build that does not generate any code.
  • ? Debug build that produces unoptimized code.
  • ? Release build that produces as optimized code as possible.
  • ? Documentation build that produces HTML documentation site produced by `rustdoc`.
Scenarios
? The different scenarios based on their incremental compilation cache state.
  • ? A non-incremental full build starting with empty cache.
  • ? An incremental build starting with empty cache.
  • ? An incremental build starting with complete cache, and unchanged source directory -- the "perfect" scenario for incremental.
  • ? An incremental build starting with complete cache, and an altered source directory. The typical variant of this is "println" which represents the addition of a `println!` macro somewhere in the source code.
Backends
? The different codegen backends used to generate executable code.
  • ? The default LLVM backend.
  • ? Alternative Cranelift backend, used primarily for faster debug builds.
Categories
? Select benchmarks based on their category (primary or secondary).
  • ? Real-world benchmarks.
  • ? Artificial benchmarks and stress-tests.
Artifacts
? Select benchmarks based on the artifact that they produce (either a binary or a library).
Show non-relevant results
? Whether to show test case results that are not relevant (i.e., not significant or have a large enough magnitude). You can see here how relevance is calculated.
Display raw data
? Whether to display or not raw data columns.
RangeMeanCount
❌
[ +0.31%,  +0.56%]
+0.41%4 (4)
✅
[ -2.22%,  -0.31%]
-0.50%19 (8)
❌,✅
[ -2.22%,  +0.56%]
-0.34%23 (11)
? The table shows summaries of regressions, improvements and all changes calculated from data that is currently visible (data that passes the active filters).
Aggregations ▶
Profile
check
No regressions
[ -0.54%,  -0.31%]
-0.43%11 (6)
[ -0.54%,  -0.31%]
-0.43%11 (6)
debug
[ +0.31%,  +0.56%]
+0.41%4 (4)
[ -2.22%,  -0.35%]
-0.74%5 (5)
[ -2.22%,  +0.56%]
-0.23%9 (8)
opt
No regressions
[ -0.39%,  -0.32%]
-0.35%3 (3)
[ -0.39%,  -0.32%]
-0.35%3 (3)
doc
No results
Scenario
full
No regressions
[ -0.54%,  -0.39%]
-0.46%2 (2)
[ -0.54%,  -0.39%]
-0.46%2 (2)
incr-full
[ +0.38%,  +0.41%]
+0.39%2 (2)
[ -0.49%,  -0.32%]
-0.38%7 (3)
[ -0.49%,  +0.41%]
-0.21%9 (5)
incr-unchanged
No regressions
[ -0.52%,  -0.33%]
-0.41%6 (4)
[ -0.52%,  -0.33%]
-0.41%6 (4)
incr-patched
[ +0.31%,  +0.56%]
+0.43%2 (2)
[ -2.22%,  -0.31%]
-0.86%4 (4)
[ -2.22%,  +0.56%]
-0.43%6 (5)
Category
primary
[ +0.31%,  +0.31%]
+0.31%1 (1)
[ -0.54%,  -0.32%]
-0.42%10 (3)
[ -0.54%,  +0.31%]
-0.35%11 (4)
secondary
[ +0.38%,  +0.56%]
+0.45%3 (3)
[ -2.22%,  -0.31%]
-0.59%9 (5)
[ -2.22%,  +0.56%]
-0.33%12 (7)
Primary benchmarks ⮯
BenchmarkProfileScenarioBackend% Change Significance Threshold
? The minimum % change that is considered significant. The higher the significance threshold, the noisier a test case is. You can see here how the significance threshold is calculated.
Significance Factor
? How much a particular result is over the significance threshold. A factor of 2.50x means the result is 2.5 times over the significance threshold.
▶helloworldcheckfullllvm
-0.54%
0.20%
2.70x
▶helloworldcheckincr-unchangedllvm
-0.52%
0.20%
2.58x
▶helloworldcheckincr-fullllvm
-0.49%
0.20%
2.43x
▶helloworldcheckincr-patched: printlnllvm
-0.48%
0.20%
2.38x
▶bitmaps-3.1.0checkincr-fullllvm
-0.41%
0.20%
2.05x
▶bitmaps-3.1.0debugincr-fullllvm
-0.38%
0.20%
1.90x
▶typenum-1.17.0checkincr-fullllvm
-0.35%
0.20%
1.77x
▶typenum-1.17.0debugincr-fullllvm
-0.35%
0.20%
1.75x
▶typenum-1.17.0optincr-fullllvm
-0.35%
0.20%
1.73x
▶bitmaps-3.1.0optincr-fullllvm
-0.32%
0.20%
1.60x
▶regex-1.5.5debugincr-patched: Jobllvm
0.31%
0.20%
1.53x
Secondary benchmarks ⮭
BenchmarkProfileScenarioBackend% Change Significance Threshold
? The minimum % change that is considered significant. The higher the significance threshold, the noisier a test case is. You can see here how the significance threshold is calculated.
Significance Factor
? How much a particular result is over the significance threshold. A factor of 2.50x means the result is 2.5 times over the significance threshold.
▶deep-vectordebugincr-patched: printlnllvm
-2.22%
0.20%
11.10x
▶deep-vectordebugincr-patched: add vec itemllvm
0.56%
0.20%
2.80x
▶unify-linearlycheckincr-unchangedllvm
-0.46%
0.20%
2.28x
▶unify-linearlycheckincr-patched: dummy fnllvm
-0.43%
0.20%
2.14x
▶regression-31157debugincr-fullllvm
0.41%
0.20%
2.05x
▶await-call-treecheckincr-unchangedllvm
-0.39%
0.20%
1.95x
▶await-call-treeoptincr-unchangedllvm
-0.39%
0.20%
1.94x
▶deeply-nested-multidebugfullllvm
-0.39%
0.20%
1.93x
▶await-call-treedebugincr-unchangedllvm
-0.38%
0.20%
1.91x
▶coercionsdebugincr-fullllvm
0.38%
0.20%
1.89x
▶issue-46449checkincr-unchangedllvm
-0.33%
0.20%
1.67x
▶issue-46449checkincr-patched: u8 3072llvm
-0.31%
0.20%
1.56x


Updated as of: 4/12/2025, 2:56:18 AM
Contribute on GitHub