Repository navigation
perf(cuda): release closed mappings off the decode thread - #10718
Performance Gate Passed
3 warnings, 1 note
⚠️ Walltime results may be unreliableSome benchmarks used the Walltime instrument outside CodSpeed Macro Runners, which can make the results inconsistent.
⚠️ Unable to measure the execution of 2 benchmarksThese benchmarks measured no execution time, so their results are not comparable and count as unchanged. This usually happens when the compiler removes the code under test.
⚠️ 24 benchmarks are dominated by system callsSystem calls cannot be simulated consistently and are left out of the measure. The reported value understates the real cost. Switch to the Walltime instrument to measure them.
ℹ️ 40 benchmarks have no flamegraph
The flamegraph wasn't generated: the symbols of the profiling trace could not be resolved, because no information about the language frames was collected.
⚡ 3 improved benchmarks
✅ 2405 untouched benchmarks
⏩ 12 skipped benchmarks1
🗄️ 43 archived benchmarks run2
Performance Changes
| Mode | Benchmark | BASE |
HEAD |
Efficiency | |
|---|---|---|---|---|---|
| ⚡ | CPU Simulation | bench_compare_primitive[(10000, 32)] |
97 µs | 72.9 µs | +33.09% |
| ⚡ | CPU Simulation | chunked_dict_primitive_canonical_into[f64, (1000, 1000, 10)] |
191.8 µs | 151.5 µs | +26.64% |
| ⚡ | Walltime | starts_with_avx512[8, VarBin] |
46.1 µs | 39.6 µs | +16.45% |
Tip
Curious why performance improved? Comment @codspeedbot explain why performance improved on this PR, or directly use the CodSpeed MCP with your agent.
Comparing ad/cuda-mapped-release (5fadeca) with develop (6d44b35)
Footnotes
-
12 benchmarks were skipped, so the baseline results were used instead. If they were deleted from the codebase, click here and archive them to remove them from the performance reports. ↩
-
43 benchmarks were run, but are now archived. If they were deleted in another branch, consider rebasing to remove them from the report. Instead if they were added back, click here to restore them. ↩