adamg/no-iter-rle-loop - Branch - spiraldb/fastlanes

Iterate directly over known range instead of `iter`

#124Closed

Comparing

adamg/no-iter-rle-loop

(

23f98a2

) with

develop

(

7c3e7e5

)

-15%

Improvement: 1

Regression: 2

Untouched: 24

Archived: 25

Benchmarks

52 total

unpack_16_from_3_stack

benches/bitpacking.rs

-15%

2.4 µs2.9 µs

unchecked_unpack_16_from_3_stack

benches/bitpacking.rs

-14%

2.6 µs3 µs

for_pack_16_to_3_stack

benches/ffor.rs

+29%

2.3 µs1.8 µs

unfor_pack_16_from_3_stack

benches/ffor.rs

+1%

3 µs2.9 µs

unchecked_unfor_pack_16_from_3_stack

benches/ffor.rs

+1%

3.2 µs3.1 µs

unpack_then_add_reference_16_from_3_stack

benches/ffor.rs

+1%

3.7 µs3.7 µs

unchecked_unpack_then_add_reference_16_from_3_stack

benches/ffor.rs

+1%

3.9 µs3.9 µs

delta_u16_fused

benches/delta.rs

+1%

3.6 µs3.6 µs

rle_encode_u32

benches/rle.rs

+1%

5.7 µs5.6 µs

delta_u16_unfused

benches/delta.rs

+1%

5.1 µs5.1 µs

transpose_u16

benches/transpose.rs

10.8 µs10.8 µs

delta_throughput_compress

benches/delta.rs

804.7 µs804.7 µs

throughput_compress

benches/ffor.rs

824.1 µs824 µs

throughput_decompress

benches/bitpacking.rs

1.2 ms1.2 ms

for_pack_16_to_3_heap

benches/ffor.rs

1.8 µs1.8 µs

pack_16_to_3_stack

benches/bitpacking.rs

2.2 µs2.2 µs

unpack_single_16_from_3

benches/bitpacking.rs

14.2 µs14.2 µs

throughput_decompress

benches/ffor.rs

1.2 ms1.2 ms

throughput_decompress_separate_reference

benches/ffor.rs

1.4 ms1.4 ms

rle_decode_u32

benches/rle.rs

4.5 µs4.5 µs

rle_throughput_decode_32

benches/rle.rs

2.4 ms2.4 ms

delta_throughput_decompress

benches/delta.rs

1.2 ms1.2 ms

rle_throughput_encode_32

benches/rle.rs

4.5 ms4.6 ms

throughput_decompress_unchecked_separate_reference

benches/ffor.rs

1.4 ms1.4 ms

throughput_decompress_unchecked

benches/ffor.rs

1.3 ms1.3 ms

Commits

Click on a commit to change the comparison range

Base

develop

7c3e7e5

-15.25%

How much does this abstraction cost

23f98a2

1 month ago

by AdamGS

Home Terms Privacy Docs