encodings/fsst: add OptimizedDecompressor for throughput
#7485Comparing
ji/fsst-optimized-decompressor
(c1f032a
) withdevelop
(12f63a4
)+20%
- Improvement
- 9
- Untouched
- 1154
- Skipped
- 1457
Benchmarks
2620 total
take_10k_first_chunk_only
encodings/fastlanes/benches/bitpacking_take.rs+20%
270.6 µs225.7 µs
take_map[(0.1, 1.0)]
vortex-array/benches/take_patches.rs+19%
2 ms1.7 ms
take_10k_dispersed
encodings/fastlanes/benches/bitpacking_take.rs+19%
284.4 µs239.5 µs
take_map[(0.1, 0.5)]
vortex-array/benches/take_patches.rs+17%
1,154.5 µs990.4 µs
patched_take_10k_contiguous_patches
encodings/fastlanes/benches/bitpacking_take.rs+13%
258.1 µs227.7 µs
patched_take_10k_contiguous_not_patches
encodings/fastlanes/benches/bitpacking_take.rs+13%
258.4 µs228.1 µs
patched_take_10k_random
encodings/fastlanes/benches/bitpacking_take.rs+13%
270.3 µs240 µs
patched_take_10k_first_chunk_only
encodings/fastlanes/benches/bitpacking_take.rs+11%
302 µs271.8 µs
patched_take_10k_dispersed
encodings/fastlanes/benches/bitpacking_take.rs+11%
316 µs285.8 µs
decompress_fsst[(10000, 64, 8)]
encodings/fsst/benches/fsst_compress.rs+10%
1.7 ms1.5 ms
chunked_canonicalize_into[(10, 10000, 64, 4)]
encodings/fsst/benches/fsst_compress.rs+9%
14.5 ms13.3 ms
chunked_into_canonical[(10, 10000, 64, 4)]
encodings/fsst/benches/fsst_compress.rs+9%
14.5 ms13.3 ms
canonicalize_compare[(10000, 64, 8)]
encodings/fsst/benches/fsst_compress.rs+9%
1.8 ms1.7 ms
decompress_fsst[(10000, 64, 4)]
encodings/fsst/benches/fsst_compress.rs+9%
1.4 ms1.3 ms
take_map[(0.1, 0.1)]
vortex-array/benches/take_patches.rs+9%
483.4 µs445 µs
eq_canonicalize_low_match
encodings/fsst/benches/fsst_url_compare.rs+8%
13.7 ms12.7 ms
eq_canonicalize_high_match
encodings/fsst/benches/fsst_url_compare.rs+8%
13.8 ms12.8 ms
canonicalize_compare[(10000, 64, 4)]
encodings/fsst/benches/fsst_compress.rs+8%
1.6 ms1.5 ms
decompress_fsst[(1000, 64, 8)]
encodings/fsst/benches/fsst_compress.rs+7%
189.3 µs177 µs
decompress_fsst[(1000, 64, 4)]
encodings/fsst/benches/fsst_compress.rs+6%
166.8 µs157.9 µs
decompress_fsst[(10000, 16, 8)]
encodings/fsst/benches/fsst_compress.rs+6%
654 µs619.5 µs
chunked_dict_fsst_canonical_into[(1000, 10, 100)]
encodings/fsst/benches/chunked_dict_fsst_builder.rs+5%
15.4 ms14.6 ms
chunked_dict_fsst_into_canonical[(1000, 10, 100)]
encodings/fsst/benches/chunked_dict_fsst_builder.rs+5%
15.4 ms14.6 ms
chunked_dict_fsst_into_canonical[(1000, 100, 100)]
encodings/fsst/benches/chunked_dict_fsst_builder.rs+5%
15.6 ms14.8 ms
chunked_dict_fsst_into_canonical[(1000, 10, 10)]
encodings/fsst/benches/chunked_dict_fsst_builder.rs+5%
1.6 ms1.5 ms
Commits
Click on a commit to change the comparison rangeBase
develop
12f63a4
+19.89%
encodings/fsst: add OptimizedDecompressor for throughput
c1f032a
10 hours ago
by joseph-isaacs