claude/blissful-mccarthy-6wV2r - Branch - spiraldb/fastlanes

Refactor bit-transpose benchmarks with macro and add walltime CI

#144

Comparing

claude/blissful-mccarthy-6wV2r

(

fd689c3

) with

develop

(

938100e

)

+39%

Improvement: 4

Untouched: 139

New: 14

Skipped: 134

Unknown Walltime execution environment detected

Using the Walltime instrument on standard Hosted Runners will lead to inconsistent data. For the most accurate results, we recommend using CodSpeed Macro Runners: bare-metal machines fine-tuned for performance measurement consistency.

Different runtime environments detected

Some benchmarks with significant performance changes were compared across different runtime environments, which may affect the accuracy of the results. Learn more about regression causes.

Benchmarks

291 total

for_pack_16_to_3_stack

benches/ffor.rs

+44%

1.8 µs1.3 µs

unpack_16_from_3_stack

benches/bitpacking.rs

+42%

2 µs1.4 µs

unchecked_unpack_16_from_3_stack

benches/bitpacking.rs

+38%

2.1 µs1.5 µs

pack_16_to_3_stack

benches/bitpacking.rs

+33%

2.4 µs1.8 µs

scalar_transpose

benches/bit_transpose.rs

N/A

N/A55.3 µs

scalar_untranspose[u8]

benches/bit_transpose.rs

N/A

N/A45.3 µs

dispatch_untranspose[u16]

benches/bit_transpose.rs

N/A

N/A51 µs

dispatch_untranspose[u32]

benches/bit_transpose.rs

N/A

N/A61.8 µs

scalar_untranspose[u16]

benches/bit_transpose.rs

N/A

N/A51.3 µs

dispatch_untranspose[u64]

benches/bit_transpose.rs

N/A

N/A5.6 µs

dispatch_transpose

benches/bit_transpose.rs

N/A

N/A6 µs

scalar_untranspose[u64]

benches/bit_transpose.rs

N/A

N/A60.6 µs

scalar_untranspose[u32]

benches/bit_transpose.rs

N/A

N/A62 µs

dispatch_untranspose[u8]

benches/bit_transpose.rs

N/A

N/A45.8 µs

vbmi_untranspose

benches/bit_transpose.rs::x86

N/A

N/A4.9 µs

vbmi_transpose

benches/bit_transpose.rs::x86

N/A

N/A4.7 µs

bmi2_untranspose

benches/bit_transpose.rs::x86

N/A

N/A33.8 µs

bmi2_transpose

benches/bit_transpose.rs::x86

N/A

N/A33.7 µs

delta_u16_fused

benches/delta.rs

-7%

3.3 µs3.6 µs

delta_u16_unfused

benches/delta.rs

-6%

4.9 µs5.2 µs

cmp_fused[u64, 47]

benches/bitpacking_cmp.rs::bench

+6%

10.9 µs10.3 µs

cmp_fused[u64, 51]

benches/bitpacking_cmp.rs::bench

+4%

11.3 µs10.8 µs

cmp_fused[u64, 30]

benches/bitpacking_cmp.rs::bench

+3%

8.7 µs8.4 µs

unfor_pack_16_from_3_stack

benches/ffor.rs

-3%

1.6 µs1.7 µs

cmp_fused[u64, 44]

benches/bitpacking_cmp.rs::bench

+3%

9.6 µs9.3 µs

Commits

Click on a commit to change the comparison range

Base

develop

938100e

+39.11%

Add walltime benchmarks for bit-transpose Intel feature tiers

cd39769

25 days ago

by claude

-2.89%

Fold bit-transpose walltime bench into the codspeed matrix

c19aedd

25 days ago

by claude

-5.97%

Run walltime benches on newest-gen Intel (c8i)

d1ae805

25 days ago

by claude

-7.64%

Spell out bit-transpose benches, shrink the macro

ca32bdb

25 days ago

by claude

+7.75%

Run all codspeed benches on c8i

882df4b

25 days ago

by claude

-30.37%

Gate bit-transpose benches per Intel feature tier (single runner each)

a90b1b4

25 days ago

by claude

+38.94%

Add #[bench(tier)] proc-macro; one runner per tier

fd689c3

25 days ago

by claude

Home Terms Privacy Docs