Comparing
ji/take-inline-compute
(bf042cb
) withdevelop
(68ea858
)-16%
- Improvement
- 45
- Regression
- 1
- Untouched
- 816
- Archived
- 1541
Benchmarks
2403 total
decompress[u8, (1000, 4)]
encodings/runend/benches/run_end_compress.rs-16%
26.2 µs31.2 µs
take_fsl_nullable_random[4096, 1000]
vortex-array/benches/take_fsl.rs×4,800
35,266.2 µs7.3 µs
take_fsl_random[4096, 1000]
vortex-array/benches/take_fsl.rs×4,800
35,462.8 µs7.4 µs
take_fsl_nullable_random[1024, 1000]
vortex-array/benches/take_fsl.rs×1,200
8,884.5 µs7.3 µs
take_fsl_random[1024, 1000]
vortex-array/benches/take_fsl.rs×1,200
8,896.1 µs7.4 µs
take_fsl_nullable_random[4096, 100]
vortex-array/benches/take_fsl.rs×490
3,577.9 µs7.3 µs
take_fsl_random[4096, 100]
vortex-array/benches/take_fsl.rs×480
3,569.1 µs7.5 µs
take_fsl_random[256, 1000]
vortex-array/benches/take_fsl.rs×300
2,253.6 µs7.5 µs
take_fsl_nullable_random[256, 1000]
vortex-array/benches/take_fsl.rs×290
2,162.2 µs7.3 µs
patched_take_10k_contiguous_patches
encodings/fastlanes/benches/bitpacking_take.rs×270
2,074.5 µs7.8 µs
patched_take_10k_random
encodings/fastlanes/benches/bitpacking_take.rs×230
2,160.2 µs9.3 µs
patched_take_10k_adversarial
encodings/fastlanes/benches/bitpacking_take.rs×230
2,056.1 µs8.9 µs
take_indices[(100000, 4)]
encodings/runend/benches/run_end_compress.rs×190
1,643 µs8.6 µs
patched_take_10k_contiguous_not_patches
encodings/fastlanes/benches/bitpacking_take.rs×160
1,245.5 µs7.8 µs
take_fsl_nullable_random[1024, 100]
vortex-array/benches/take_fsl.rs×130
928.8 µs7.3 µs
take_10k_random
encodings/fastlanes/benches/bitpacking_take.rs×130
970.2 µs7.6 µs
take_fsl_random[1024, 100]
vortex-array/benches/take_fsl.rs×120
912.8 µs7.4 µs
take_indices[(100000, 16)]
encodings/runend/benches/run_end_compress.rs×100
738.8 µs7.2 µs
take_fsl_nullable_random[64, 1000]
vortex-array/benches/take_fsl.rs×83
609.5 µs7.3 µs
take_fsl_random[64, 1000]
vortex-array/benches/take_fsl.rs×80
593.4 µs7.4 µs
take_fsl_nullable_random[256, 100]
vortex-array/benches/take_fsl.rs×36
261.2 µs7.3 µs
take_fsl_random[256, 100]
vortex-array/benches/take_fsl.rs×33
248.8 µs7.5 µs
take_10k_contiguous
encodings/fastlanes/benches/bitpacking_take.rs×33
254.2 µs7.7 µs
take_indices[(100000, 1024)]
encodings/runend/benches/run_end_compress.rs×33
237.9 µs7.3 µs
take_indices[(100000, 4096)]
encodings/runend/benches/run_end_compress.rs×32
235.8 µs7.3 µs
Commits
Click on a commit to change the comparison rangeBase
develop
68ea858
-16.09%
chore[array]: inline take compute function
bf042cb
3 months ago
by joseph-isaacs