Skip to content

Splat run-end bool decode into 64-bit words - #9021

Open
joseph-isaacs wants to merge 5 commits into
claude/runend-bool-decode-benchesfrom
claude/runend-decompress-bools-1t2re2
Open

Splat run-end bool decode into 64-bit words#9021
joseph-isaacs wants to merge 5 commits into
claude/runend-bool-decode-benchesfrom
claude/runend-decompress-bools-1t2re2

Refit both dispatch constants for normal run lengths

13cedf9
Select commit
Loading
Failed to load commit list.
CodSpeed HQ / CodSpeed Performance Analysis succeeded Jul 28, 2026 in 0s

Performance Gate Passed

⚡ 39 improved benchmarks
✅ 1828 untouched benchmarks
🆕 90 new benchmarks
⏩ 46 skipped benchmarks1

Performance Changes

Mode Benchmark BASE HEAD Efficiency
⚡ Simulation decode_bool_nullable[10000_10_mostly_true_mostly_null] 64 µs 36.5 µs +75.45%
⚡ Simulation decode_bool_nullable[10000_10_mostly_true_half_valid] 64.6 µs 38.5 µs +67.67%
⚡ Simulation decode_bool_nullable[10000_10_alternating_half_valid] 62.8 µs 38.6 µs +62.84%
⚡ Simulation decode_bool_corpus[4x100k_run4_even_nullable] 3.7 ms 2.4 ms +54.4%
⚡ Simulation decode_bool_corpus[4x100k_run64_even] 252.9 µs 165.6 µs +52.69%
⚡ Simulation decode_bool_corpus[4x100k_run16_even] 786.8 µs 516.9 µs +52.2%
⚡ Simulation decode_bool_corpus[128x16k_run4_even_nullable] 19.4 ms 12.8 ms +50.83%
⚡ Simulation decode_bool_corpus[128x16k_run16_even] 4.2 ms 2.8 ms +46.8%
⚡ Simulation decode_bool_corpus[4x100k_run4_even] 2.4 ms 1.7 ms +44.51%
⚡ Simulation decode_bool_corpus[128x16k_run4_even] 12.8 ms 9 ms +43.34%
⚡ Simulation decode_bool_corpus[4x100k_run16_even_nullable] 1,178.3 µs 825.2 µs +42.78%
⚡ Simulation decode_bool_nullable[10000_2_alternating_mostly_valid] 185.4 µs 130.6 µs +41.88%
⚡ Simulation decode_bool_corpus[128x16k_run16_even_nullable] 6.2 ms 4.4 ms +41.26%
⚡ Simulation decode_bool_corpus[4x100k_run64_even_nullable] 401.1 µs 285.5 µs +40.51%
⚡ Simulation decode_bool_corpus[128x16k_run64_even] 1.4 ms 1 ms +38.31%
⚡ Simulation decode_bool[10000_2_alternating] 125.5 µs 91.8 µs +36.7%
⚡ Simulation null_count_run_end[(100000, 16, 0.5)] 231.2 µs 170.7 µs +35.42%
⚡ Simulation decode_bool_nullable[10000_10_alternating_mostly_null] 49.7 µs 36.7 µs +35.3%
⚡ Simulation null_count_run_end[(100000, 4, 0.5)] 674.4 µs 512.1 µs +31.71%
⚡ Simulation decode_bool_nullable[10000_10_alternating_mostly_valid] 60.2 µs 45.9 µs +31.25%
... ... ... ... ... ...

ℹ️ Only the first 20 benchmarks are displayed. Go to the app to view all benchmarks.

Tip

Curious why this is faster? Comment @codspeedbot explain why this is faster on this PR, or directly use the CodSpeed MCP with your agent.

Comparing claude/runend-decompress-bools-1t2re2 (13cedf9) with claude/runend-bool-decode-benches (bd674b8)

Open in CodSpeed

Footnotes

  1. 46 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. ↩

View more details on CodSpeed HQ