feat[vortex-cuda]: GPU FSST decompression kernel
#7776Comparing
asubiotto/fsst-cuda
(46670e8
) withdevelop
(903ee6c
)+10%
- Improvement
- 1
- Untouched
- 1168
- Skipped
- 138
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.Benchmarks
1307 total
cuda/bitpacked_u16/unpack/5bw[100M]
vortex-cuda/benches/bitpacked_cuda.rs::benches::benchmark_bitunpack::cuda+10%
613.2 µs557.2 µs
cuda/bitpacked_u32/unpack/6bw[100M]
vortex-cuda/benches/bitpacked_cuda.rs::benches::benchmark_bitunpack::cuda+6%
1.1 ms1 ms
take_10_contiguous
encodings/fastlanes/benches/bitpacking_take.rs+4%
25.2 µs24.2 µs
iter_vortex_buffer[128]
vortex-buffer/benches/vortex_bitbuffer.rs+3%
981.1 ns951.9 ns
patched_take_10_contiguous
encodings/fastlanes/benches/bitpacking_take.rs+3%
30.7 µs29.9 µs
set_indices_vortex_buffer[128]
vortex-buffer/benches/vortex_bitbuffer.rs+3%
1 µs1 µs
encode_varbin[(1000, 4)]
vortex-array/benches/dict_compress.rs+3%
167.3 µs162.8 µs
encode_varbin[(1000, 2)]
vortex-array/benches/dict_compress.rs+2%
166.3 µs162.6 µs
append_arrow_buffer[128]
vortex-buffer/benches/vortex_bitbuffer.rs+2%
1.4 µs1.4 µs
decode_varbin[(1000, 512)]
vortex-array/benches/dict_compress.rs+2%
30 µs29.4 µs
append_n_arrow_buffer[128]
vortex-buffer/benches/vortex_bitbuffer.rs+2%
1.8 µs1.8 µs
transpose_scalar
encodings/fastlanes/benches/bit_transpose.rs+2%
1.8 µs1.8 µs
decode_varbin[(1000, 4)]
vortex-array/benches/dict_compress.rs+2%
26.4 µs25.9 µs
decode_varbin[(1000, 32)]
vortex-array/benches/dict_compress.rs+2%
26.7 µs26.3 µs
map_each[Buffer<i32>, 128]
vortex-buffer/benches/vortex_buffer.rs+1%
2.1 µs2.1 µs
chunked_canonicalize_into[(10, 10000, 4, 4)]
encodings/fsst/benches/fsst_compress.rs+1%
4.2 ms4.1 ms
decode_varbinview[(1000, 32)]
vortex-array/benches/dict_compress.rs+1%
27.1 µs26.8 µs
chunked_dict_primitive_canonical_into[f32, (1000, 100, 10)]
vortex-array/benches/chunked_dict_builder.rs+1%
84.7 µs83.6 µs
encode_varbinview[(1000, 512)]
vortex-array/benches/dict_compress.rs+1%
225.2 µs222.4 µs
train_compressor[(10000, 4, 4)]
encodings/fsst/benches/fsst_compress.rs+1%
3.7 ms3.6 ms
chunked_dict_primitive_into_canonical[f32, (1000, 1000, 100)]
vortex-array/benches/chunked_dict_builder.rs+1%
881 µs871.1 µs
append_n_arrow_buffer[1024]
vortex-buffer/benches/vortex_bitbuffer.rs+1%
2.6 µs2.6 µs
append_n_arrow_buffer[2048]
vortex-buffer/benches/vortex_bitbuffer.rs+1%
2.6 µs2.6 µs
chunked_dict_primitive_into_canonical[f64, (1000, 10, 10)]
vortex-array/benches/chunked_dict_builder.rs+1%
122.8 µs121.7 µs
encode_primitives[i64, (1000, 8)]
vortex-array/benches/dict_compress.rs+1%
92.6 µs91.7 µs
Commits
Click on a commit to change the comparison rangeBase
develop
903ee6c
+10.03%
feat[vortex-cuda]: GPU FSST decompression kernel
46670e8
5 days ago
by asubiotto