fused kernels
#6478Comparing
aduffy/fused-ops
(d94270d
) withaduffy/gpu-scan-measure
(7b61bd6
)+12%
- Improvement
- 1
- Untouched
- 1134
- Skipped
- 1268
Benchmarks
2403 total
old_bp_prim_test_between[i64, 32768]
encodings/fastlanes/benches/compute_between.rs::bitpack+12%
573.3 µs512.1 µs
new_bp_prim_test_between[i32, 32768]
encodings/fastlanes/benches/compute_between.rs::bitpack+9%
159.6 µs146.9 µs
new_bp_prim_test_between[i64, 32768]
encodings/fastlanes/benches/compute_between.rs::bitpack+6%
222.6 µs210.3 µs
true_count_vortex_buffer[128]
vortex-buffer/benches/vortex_bitbuffer.rs+6%
1.1 µs1 µs
true_count_vortex_buffer[1024]
vortex-buffer/benches/vortex_bitbuffer.rs+5%
1.2 µs1.1 µs
set_indices_vortex_buffer[128]
vortex-buffer/benches/vortex_bitbuffer.rs+5%
1.3 µs1.3 µs
true_count_vortex_buffer[2048]
vortex-buffer/benches/vortex_bitbuffer.rs+4%
1.4 µs1.3 µs
take_map[(0.005, 0.05)]
vortex-array/benches/take_patches.rs+4%
174.3 µs167.7 µs
old_bp_prim_test_between[i16, 32768]
encodings/fastlanes/benches/compute_between.rs::bitpack+3%
453.7 µs439.4 µs
old_bp_prim_test_between[i32, 32768]
encodings/fastlanes/benches/compute_between.rs::bitpack+3%
462.2 µs449.2 µs
old_alp_prim_test_between[f64, 32768]
encodings/fastlanes/benches/compute_between.rs::alp+3%
587.8 µs571.4 µs
old_bp_prim_test_between[i32, 16384]
encodings/fastlanes/benches/compute_between.rs::bitpack+3%
304.5 µs296.9 µs
push_n_vortex_buffer[u8, 128]
vortex-buffer/benches/vortex_buffer.rs+2%
2.4 µs2.3 µs
take_map[(0.01, 0.1)]
vortex-array/benches/take_patches.rs+2%
325.7 µs318.1 µs
decode_primitives[f32, (1000, 2)]
vortex-array/benches/dict_compress.rs+2%
24.1 µs23.6 µs
slice_vortex_buffer[65536]
vortex-buffer/benches/vortex_bitbuffer.rs+2%
1.4 µs1.3 µs
slice_vortex_buffer[16384]
vortex-buffer/benches/vortex_bitbuffer.rs+2%
1.4 µs1.3 µs
slice_vortex_buffer[128]
vortex-buffer/benches/vortex_bitbuffer.rs+2%
1.4 µs1.3 µs
slice_vortex_buffer[1024]
vortex-buffer/benches/vortex_bitbuffer.rs+2%
1.4 µs1.3 µs
slice_vortex_buffer[2048]
vortex-buffer/benches/vortex_bitbuffer.rs+2%
1.4 µs1.3 µs
push_n_vortex_buffer[u8, 1024]
vortex-buffer/benches/vortex_buffer.rs+2%
2.8 µs2.7 µs
decode_primitives[u8, (1000, 2)]
vortex-array/benches/dict_compress.rs+2%
23.8 µs23.3 µs
set_indices_vortex_buffer[1024]
vortex-buffer/benches/vortex_bitbuffer.rs+2%
3 µs3 µs
push_n_vortex_buffer[u8, 2048]
vortex-buffer/benches/vortex_buffer.rs+2%
3.2 µs3.1 µs
iter_vortex_buffer[128]
vortex-buffer/benches/vortex_bitbuffer.rs+2%
1.6 µs1.6 µs
Commits
Click on a commit to change the comparison rangeBase
aduffy/gpu-scan-measure
7b61bd6
+11.96%
fused FOR + BP
d94270d
23 hours ago
by a10y