feat(cuda): single-pass dispatch plan builder
#7197Comparing
ad/cuda-next
(b4adfcd
) withdevelop
(d0ed3fc
)-16%
- Regression
- 2
- Untouched
- 1104
- Skipped
- 1522
Benchmarks
2628 total
bitwise_not_vortex_buffer_mut[128]
vortex-buffer/benches/vortex_bitbuffer.rs-16%
317.8 ns376.1 ns
bitwise_not_vortex_buffer_mut[1024]
vortex-buffer/benches/vortex_bitbuffer.rs-11%
477.2 ns535.6 ns
binary_search_std
vortex-array/benches/search_sorted.rs+5%
582.8 ns553.6 ns
binary_search_vortex
vortex-array/benches/search_sorted.rs+4%
807.2 ns778.1 ns
iter_arrow_buffer[128]
vortex-buffer/benches/vortex_bitbuffer.rs+3%
897.8 ns868.6 ns
value_arrow_buffer[128]
vortex-buffer/benches/vortex_bitbuffer.rs+3%
1.1 µs1.1 µs
chunked_opt_bool_into_canonical[(1000, 10)]
vortex-array/benches/chunk_array_builder.rs+2%
115.8 µs113 µs
append_arrow_buffer[128]
vortex-buffer/benches/vortex_bitbuffer.rs+2%
1.5 µs1.4 µs
untranspose_bmi2
encodings/fastlanes/benches/bit_transpose.rs::x86+2%
1.8 µs1.8 µs
bitwise_not_vortex_buffer[1024]
vortex-buffer/benches/vortex_bitbuffer.rs+2%
3.9 µs3.8 µs
bitwise_not_vortex_buffer[2048]
vortex-buffer/benches/vortex_bitbuffer.rs+1%
4.1 µs4.1 µs
bitwise_not_vortex_buffer[128]
vortex-buffer/benches/vortex_bitbuffer.rs+1%
4.6 µs4.5 µs
decode_primitives[f32, (1000, 4)]
vortex-array/benches/dict_compress.rs+1%
26.7 µs26.4 µs
map_each[Buffer<i32>, 128]
vortex-buffer/benches/vortex_buffer.rs+1%
2.6 µs2.5 µs
chunked_bool_canonical_into[(1000, 10)]
vortex-array/benches/chunk_array_builder.rs+1%
76.9 µs76.2 µs
append_n_vortex_buffer[128]
vortex-buffer/benches/vortex_bitbuffer.rs+1%
3.2 µs3.2 µs
decode_primitives[f32, (1000, 8)]
vortex-array/benches/dict_compress.rs+1%
26.7 µs26.5 µs
decode_varbin[(1000, 32)]
vortex-array/benches/dict_compress.rs+1%
37.8 µs37.5 µs
decode_primitives[f32, (1000, 2)]
vortex-array/benches/dict_compress.rs+1%
26.5 µs26.3 µs
bitwise_not_vortex_buffer[16384]
vortex-buffer/benches/vortex_bitbuffer.rs+1%
8.4 µs8.3 µs
encode_primitives[u8, (1000, 2)]
vortex-array/benches/dict_compress.rs+1%
96 µs95.3 µs
decode_varbin[(10000, 8)]
vortex-array/benches/dict_compress.rs+1%
152.5 µs151.5 µs
iter_arrow_buffer[1024]
vortex-buffer/benches/vortex_bitbuffer.rs+1%
4.7 µs4.6 µs
decode_primitives[i64, (1000, 32)]
vortex-array/benches/dict_compress.rs+1%
30.5 µs30.3 µs
take_map[(0.005, 1.0)]
vortex-array/benches/take_patches.rs+1%
2.4 ms2.4 ms
Commits
Click on a commit to change the comparison rangeBase
develop
d0ed3fc
-15.51%
docs: adjust comment
76445a0
18 hours ago
by 0ax10%
well actually
ecbec13
17 hours ago
by 0ax10%
prefer `as_` cast
2283116
15 hours ago
by 0ax10%
docs
b4adfcd
6 hours ago
by 0ax1