Skip to content

Extend the RunEnd filter benchmark - #10978

Merged
robert3005 merged 1 commit into
developfrom
rk/runend-filter-bench
Oct 29, 2026
Merged

robert3005 merged 1 commit into
developfrom
rk/runend-filter-bench

Conversation

@robert3005

@robert3005 robert3005 commented Oct 29, 2026 •

Copy link
Copy Markdown
Contributor

No description provided.

@robert3005
robert3005 added this pull request to stack #10980 October 29, 2026 16:28
@robert3005 robert3005 added the changelog/chore A trivial change label Oct 29, 2026
@robert3005
robert3005 marked this pull request as ready for review October 29, 2026 16:30
Add cases with runs of 4 and 256 rows at half density, so the filter
benchmark covers both ends of the run length range. Allocate the filtered
run ends through the allocator of a session execution context, as the
filter kernel does, instead of the static allocator.

Signed-off-by: Robert Kruszewski <github@robertk.io>
@robert3005
robert3005 force-pushed the rk/runend-filter-bench branch from 7315c6c to 1351c90 Compare October 29, 2026 16:32
@robert3005
robert3005 removed this pull request from stack #10980 October 29, 2026 16:32
@robert3005
robert3005 added this pull request to stack #10981 October 29, 2026 16:32
@robert3005 robert3005 changed the title Add short-run and long-run cases to the RunEnd filter benchmark Extend the RunEnd filter benchmark Oct 29, 2026
@codspeed

codspeed Bot commented Oct 29, 2026 •

Copy link
Copy Markdown

Merging this PR will degrade performance by 19.43%

1 warning

⚠️ Walltime results may be unreliable

Some benchmarks used the Walltime instrument outside CodSpeed Macro Runners, which can make the results inconsistent.

Learn more about Macro Runners

See affected benchmarks

⚡ 1 improved benchmark
❌ 4 regressed benchmarks
✅ 2299 untouched benchmarks
🆕 2 new benchmarks

Warning

Please fix the performance issues or acknowledge them on CodSpeed.

Performance Changes

Mode Benchmark BASE HEAD Efficiency
❌ Walltime threshold_neon[64] 2.6 µs 3.7 µs -29.93%
❌ Walltime filter_neon[80] 2.6 µs 3.6 µs -27.77%
❌ Walltime filter_neon[64] 1.9 µs 2.5 µs -26.24%
❌ Walltime threshold_neon[80] 3.5 µs 4.2 µs -17.53%
⚡ Walltime threshold_avx2[80] 3 µs 2.7 µs +10.28%
🆕 CPU Simulation filter_run_end[len=65536_run=256_density=0.5] N/A 65.2 µs N/A
🆕 CPU Simulation filter_run_end[len=65536_run=4_density=0.5] N/A 2.6 ms N/A

Tip

Investigate this regression by commenting @codspeedbot fix this regression on this PR, or directly use the CodSpeed MCP with your agent.


Comparing rk/runend-filter-bench (1351c90) with develop (1c2f6eb)

Open in CodSpeed

@robert3005
robert3005 merged commit faa6473 into develop Oct 29, 2026
85 of 87 checks passed
@robert3005
robert3005 deleted the rk/runend-filter-bench branch October 29, 2026 16:45
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

changelog/chore A trivial change

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants