Try and improve allocation-happy code paths
#6337Comparing
adamg/reduce-allocations-some-cases
(8172fcb
) withdevelop
(a0cf8ae
)+36%
- Improvement
- 6
- Untouched
- 1132
- Skipped
- 1265
Benchmarks
2403 total
chunked_opt_bool_into_canonical[(10, 1000)]
vortex-array/benches/chunk_array_builder.rs+36%
1.9 ms1.4 ms
chunked_bool_into_canonical[(10, 1000)]
vortex-array/benches/chunk_array_builder.rs+28%
1.8 ms1.4 ms
chunked_bool_into_canonical[(100, 100)]
vortex-array/benches/chunk_array_builder.rs+24%
228.1 µs183.2 µs
chunked_varbinview_into_canonical[(10, 1000)]
vortex-array/benches/chunk_array_builder.rs+21%
3 ms2.5 ms
chunked_opt_bool_into_canonical[(100, 100)]
vortex-array/benches/chunk_array_builder.rs+16%
444.8 µs384.4 µs
chunked_bool_into_canonical[(1000, 10)]
vortex-array/benches/chunk_array_builder.rs+15%
87.5 µs75.8 µs
chunked_opt_bool_into_canonical[(1000, 10)]
vortex-array/benches/chunk_array_builder.rs+9%
113.3 µs103.5 µs
chunked_varbinview_opt_into_canonical[(10, 1000)]
vortex-array/benches/chunk_array_builder.rs+9%
7.4 ms6.7 ms
scalar_at_struct_wide
vortex-array/benches/scalar_at_struct.rs+8%
3.2 ms2.9 ms
chunked_dict_primitive_into_canonical[u32, (1000, 100, 10)]
vortex-array/benches/chunked_dict_builder.rs+8%
110.4 µs102.5 µs
chunked_into_canonical[(1000, 50, 8, 4)]
encodings/fsst/benches/fsst_compress.rs+8%
8 ms7.4 ms
decode_primitives[u8, (1000, 2)]
vortex-array/benches/dict_compress.rs+8%
24.8 µs23 µs
chunked_dict_primitive_into_canonical[f32, (1000, 100, 10)]
vortex-array/benches/chunked_dict_builder.rs+7%
111.3 µs103.6 µs
scalar_at_struct_simple
vortex-array/benches/scalar_at_struct.rs+7%
836.4 µs781.5 µs
chunked_dict_primitive_into_canonical[u32, (1000, 100, 100)]
vortex-array/benches/chunked_dict_builder.rs+7%
837.5 µs783.7 µs
chunked_dict_primitive_into_canonical[f32, (1000, 10, 10)]
vortex-array/benches/chunked_dict_builder.rs+7%
109.4 µs102.4 µs
chunked_dict_primitive_into_canonical[u32, (1000, 10, 100)]
vortex-array/benches/chunked_dict_builder.rs+7%
820.4 µs768.5 µs
chunked_dict_primitive_into_canonical[u32, (1000, 1000, 10)]
vortex-array/benches/chunked_dict_builder.rs+6%
127 µs119.3 µs
decode_primitives[f32, (1000, 8)]
vortex-array/benches/dict_compress.rs+6%
24.7 µs23.2 µs
chunked_into_canonical[(1000, 50, 8, 16)]
encodings/fsst/benches/fsst_compress.rs+6%
9.1 ms8.5 ms
decode_primitives[u8, (1000, 128)]
vortex-array/benches/dict_compress.rs+6%
22.3 µs21 µs
compress[(1000, 256)]
encodings/runend/benches/run_end_compress.rs+6%
52.5 µs49.5 µs
decode_primitives[u8, (1000, 8)]
vortex-array/benches/dict_compress.rs+6%
22.2 µs21 µs
decode_primitives[u8, (1000, 32)]
vortex-array/benches/dict_compress.rs+6%
22.3 µs21 µs
decode_primitives[u8, (1000, 512)]
vortex-array/benches/dict_compress.rs+6%
22.3 µs21.1 µs
Commits
Click on a commit to change the comparison rangeBase
develop
a0cf8ae
+36.4%
why not
015b5d7
1 day ago
by AdamGS-0.01%
Revert filter change
8172fcb
1 day ago
by AdamGS