Comparing
revert-8047-fix-cuda-ptx-gpu-invalidation
(68dc253
) withdevelop
(c54ce7e
)-2%
- Improvement
- 2
- Regression
- 1
- Untouched
- 1234
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
1237 total
fast_lt_out_of_range[4, 65536]
encodings/fastlanes/benches/bitpack_compare.rs-22%
204.3 µs262.3 µs
baseline_eq[16, 65536]
encodings/fastlanes/benches/bitpack_compare.rs+11%
287.6 µs259.6 µs
baseline_lt[16, 65536]
encodings/fastlanes/benches/bitpack_compare.rs+10%
302.7 µs274.7 µs
cuda/alp_for_bp_6bw_f32/dispatch_f32[100M]
vortex-cuda/benches/dynamic_dispatch_cuda.rs::benches::benchmark_dynamic_dispatch::cuda+5%
1.1 ms1 ms
decode_primitives[f32, (1000, 512)]
vortex-array/benches/dict_compress.rs+3%
18.8 µs18.3 µs
chunked_dict_fsst_canonical_into[(1000, 100, 100)]
encodings/fsst/benches/chunked_dict_fsst_builder.rs+3%
14 ms13.6 ms
take_10_contiguous
encodings/fastlanes/benches/bitpacking_take.rs+3%
26.3 µs25.6 µs
chunked_dict_fsst_into_canonical[(1000, 100, 100)]
encodings/fsst/benches/chunked_dict_fsst_builder.rs+3%
14.1 ms13.7 ms
chunked_dict_fsst_canonical_into[(1000, 10, 100)]
encodings/fsst/benches/chunked_dict_fsst_builder.rs+3%
13.9 ms13.6 ms
chunked_dict_fsst_canonical_into[(1000, 1000, 100)]
encodings/fsst/benches/chunked_dict_fsst_builder.rs+3%
14.9 ms14.5 ms
chunked_dict_fsst_into_canonical[(1000, 1000, 100)]
encodings/fsst/benches/chunked_dict_fsst_builder.rs+2%
14.9 ms14.6 ms
baseline_eq[4, 65536]
encodings/fastlanes/benches/bitpack_compare.rs+2%
243.7 µs238.2 µs
encode_varbin[(1000, 8)]
vortex-array/benches/dict_compress.rs+2%
165.8 µs162.1 µs
encode_varbin[(1000, 32)]
vortex-array/benches/dict_compress.rs+2%
171.2 µs167.5 µs
baseline_lt[4, 65536]
encodings/fastlanes/benches/bitpack_compare.rs+2%
257.7 µs252.2 µs
push_n_vortex_buffer[u16, 128]
vortex-buffer/benches/vortex_buffer.rs+2%
1.6 µs1.6 µs
transpose_bmi2
encodings/fastlanes/benches/bit_transpose.rs::x86+2%
1.8 µs1.7 µs
map_each[Buffer<i32>, 128]
vortex-buffer/benches/vortex_buffer.rs+2%
1.9 µs1.8 µs
push_n_vortex_buffer[u64, 128]
vortex-buffer/benches/vortex_buffer.rs+2%
1.9 µs1.9 µs
decode_varbin[(1000, 4)]
vortex-array/benches/dict_compress.rs+1%
27.6 µs27.2 µs
decode_primitives[u8, (1000, 32)]
vortex-array/benches/dict_compress.rs+1%
16.3 µs16.1 µs
decode_primitives[i64, (1000, 512)]
vortex-array/benches/dict_compress.rs+1%
21.7 µs21.4 µs
decode_varbin[(1000, 32)]
vortex-array/benches/dict_compress.rs+1%
27.8 µs27.5 µs
decode_primitives[u8, (1000, 4)]
vortex-array/benches/dict_compress.rs+1%
16.3 µs16.1 µs
push_n_vortex_buffer[u16, 1024]
vortex-buffer/benches/vortex_buffer.rs+1%
3.2 µs3.2 µs
Commits
Click on a commit to change the comparison rangeBase
develop
c54ce7e
-1.68%
Revert "fix: build CUDA kernels as multi-arch fatbin with PTX fallback (#8047)"
68dc253
3 hours ago
by robert3005