tryout fsst decompress speedup
#6998Comparing
ji/make-view
(a05fcab
) withdevelop
(50868db
)+22%
- Improvement
- 21
- Untouched
- 988
- New
- 20
- Skipped
- 1515
Benchmarks
2544 total
compress_fsst[(1000, 256, 8)]
encodings/fsst/benches/fsst_compress.rsN/A
N/A3.5 ms
compress_fsst[(1000, 256, 4)]
encodings/fsst/benches/fsst_compress.rsN/A
N/A2.9 ms
compress_fsst[(10000, 256, 4)]
encodings/fsst/benches/fsst_compress.rsN/A
N/A28.7 ms
compress_fsst[(10000, 256, 8)]
encodings/fsst/benches/fsst_compress.rsN/A
N/A34.1 ms
decompress_fsst[(1000, 256, 4)]
encodings/fsst/benches/fsst_compress.rsN/A
N/A859.2 µs
decompress_fsst[(1000, 256, 8)]
encodings/fsst/benches/fsst_compress.rsN/A
N/A1 ms
decompress_fsst[(10000, 256, 4)]
encodings/fsst/benches/fsst_compress.rsN/A
N/A8.3 ms
pushdown_compare[(1000, 256, 8)]
encodings/fsst/benches/fsst_compress.rsN/A
N/A464.4 µs
pushdown_compare[(1000, 256, 4)]
encodings/fsst/benches/fsst_compress.rsN/A
N/A455 µs
pushdown_compare[(10000, 256, 4)]
encodings/fsst/benches/fsst_compress.rsN/A
N/A614.6 µs
pushdown_compare[(10000, 256, 8)]
encodings/fsst/benches/fsst_compress.rsN/A
N/A625.1 µs
decompress_fsst[(10000, 256, 8)]
encodings/fsst/benches/fsst_compress.rsN/A
N/A10.1 ms
train_compressor[(1000, 256, 8)]
encodings/fsst/benches/fsst_compress.rsN/A
N/A4.3 ms
train_compressor[(1000, 256, 4)]
encodings/fsst/benches/fsst_compress.rsN/A
N/A4.2 ms
train_compressor[(10000, 256, 4)]
encodings/fsst/benches/fsst_compress.rsN/A
N/A4.7 ms
train_compressor[(10000, 256, 8)]
encodings/fsst/benches/fsst_compress.rsN/A
N/A4.8 ms
chunked_canonicalize_into[(1000, 50, 8, 16)]
encodings/fsst/benches/fsst_compress.rs+10%
8.9 ms8.1 ms
chunked_into_canonical[(1000, 50, 8, 16)]
encodings/fsst/benches/fsst_compress.rs+9%
9.1 ms8.4 ms
chunked_canonicalize_into[(1000, 100, 16, 16)]
encodings/fsst/benches/fsst_compress.rs+9%
20.5 ms18.9 ms
chunked_into_canonical[(1000, 100, 16, 16)]
encodings/fsst/benches/fsst_compress.rs+8%
20.8 ms19.1 ms
canonicalize_compare[(1000, 4, 8)]
encodings/fsst/benches/fsst_compress.rs+8%
202.8 µs187.3 µs
canonicalize_compare[(1000, 4, 4)]
encodings/fsst/benches/fsst_compress.rs+8%
199.4 µs184.2 µs
chunked_canonicalize_into[(1000, 50, 8, 64)]
encodings/fsst/benches/fsst_compress.rs+8%
10.5 ms9.7 ms
chunked_into_canonical[(1000, 50, 8, 64)]
encodings/fsst/benches/fsst_compress.rs+8%
10.8 ms10 ms
canonicalize_compare[(1000, 16, 4)]
encodings/fsst/benches/fsst_compress.rs+8%
224.9 µs209.1 µs
Commits
Click on a commit to change the comparison rangeBase
develop
50868db
+0.13%
fix
5de1a3d
22 hours ago
by joseph-isaacs+11.35%
fix
046b1b6
22 hours ago
by joseph-isaacs+9.15%
fix
fde3497
21 hours ago
by joseph-isaacs+1.87%
fix
a05fcab
20 hours ago
by joseph-isaacs