Avatar for the zarr-developers user
zarr-developers
cast-value.rs
BlogDocsChangelog

perf: new simd kernels for improved performance

#22Merged
Comparing
d-v-b:perf/beat-numpy
(
276b0da
) with
main
(
48edf3a
)
CodSpeed Performance Gauge
+61%
Improvement
3
Untouched
42
New
6
Skipped
21

Benchmarks

72 total
f64_to_f32/nearest_even[64000]
core/benches/conversions.rs::benches::bench_f64_to_f32::conversions
CodSpeed Performance Gauge
+61%
688.3 µs427.7 µs
f64_to_f32/nearest_even[1024]
core/benches/conversions.rs::benches::bench_f64_to_f32::conversions
CodSpeed Performance Gauge
+49%
11.5 µs7.7 µs
f64_to_f32/nearest_even[1000000]
core/benches/conversions.rs::benches::bench_f64_to_f32::conversions
CodSpeed Performance Gauge
+45%
10.7 ms7.4 ms
f64_to_f32/towards_zero[1024]
core/benches/conversions.rs::benches::bench_f64_to_f32::conversions
CodSpeed Performance Gauge
N/A
N/A20.2 µs
f64_to_i32/no_oor[64000]
core/benches/conversions.rs::benches::bench_f64_to_i32::conversions
CodSpeed Performance Gauge
N/A
N/A497.9 µs
f64_to_i32/no_oor[1024]
core/benches/conversions.rs::benches::bench_f64_to_i32::conversions
CodSpeed Performance Gauge
N/A
N/A9.2 µs
f64_to_f32/towards_zero[1000000]
core/benches/conversions.rs::benches::bench_f64_to_f32::conversions
CodSpeed Performance Gauge
N/A
N/A18.8 ms
f64_to_i32/no_oor[1000000]
core/benches/conversions.rs::benches::bench_f64_to_i32::conversions
CodSpeed Performance Gauge
N/A
N/A7.8 ms
f64_to_f32/towards_zero[64000]
core/benches/conversions.rs::benches::bench_f64_to_f32::conversions
CodSpeed Performance Gauge
N/A
N/A1.2 ms
f64_to_f32/clamp[1000000]
core/benches/conversions.rs::benches::bench_f64_to_f32::conversions
CodSpeed Performance Gauge
+8%
6.3 ms5.8 ms
f64_to_f32/clamp[64000]
core/benches/conversions.rs::benches::bench_f64_to_f32::conversions
CodSpeed Performance Gauge
+8%
403.8 µs373.1 µs
f64_to_f32/clamp[1024]
core/benches/conversions.rs::benches::bench_f64_to_f32::conversions
CodSpeed Performance Gauge
+2%
6.9 µs6.8 µs
f64_to_i32/clamp[1024]
core/benches/conversions.rs::benches::bench_f64_to_i32::conversions
CodSpeed Performance Gauge
+1%
8.6 µs8.5 µs
f64_to_u8/clamp+TowardsPositive[1024]
core/benches/conversions.rs::benches::bench_f64_to_u8::conversions
CodSpeed Performance Gauge
+1%
6.3 µs6.2 µs
f64_to_u8/clamp+TowardsZero[1024]
core/benches/conversions.rs::benches::bench_f64_to_u8::conversions
CodSpeed Performance Gauge
+1%
6.3 µs6.3 µs
f64_to_u8/clamp+TowardsNegative[1024]
core/benches/conversions.rs::benches::bench_f64_to_u8::conversions
CodSpeed Performance Gauge
+1%
6.3 µs6.3 µs
f64_to_u8/clamp[1024]
core/benches/conversions.rs::benches::bench_f64_to_u8::conversions
CodSpeed Performance Gauge
+1%
6.3 µs6.3 µs
i32_to_u8/clamp[1024]
core/benches/conversions.rs::benches::bench_i32_to_u8::conversions
CodSpeed Performance Gauge
+1%
3.4 µs3.4 µs
i64_to_f32/nearest_even[1024]
core/benches/conversions.rs::benches::bench_i64_to_f32::conversions
CodSpeed Performance Gauge
+1%
7.5 µs7.5 µs
i32_to_u8/clamp+scalar_map[1024]
core/benches/conversions.rs::benches::bench_i32_to_u8::conversions
CodSpeed Performance Gauge
0%
8.4 µs8.3 µs
i64_to_f32/towards_zero[1024]
core/benches/conversions.rs::benches::bench_i64_to_f32::conversions
CodSpeed Performance Gauge
0%
11.5 µs11.5 µs
f64_to_u8/clamp+scalar_map[1024]
core/benches/conversions.rs::benches::bench_f64_to_u8::conversions
CodSpeed Performance Gauge
0%
25.8 µs25.8 µs
f64_to_i32/clamp[64000]
core/benches/conversions.rs::benches::bench_f64_to_i32::conversions
CodSpeed Performance Gauge
0%
466.7 µs466.6 µs
f64_to_u8/clamp+TowardsPositive[64000]
core/benches/conversions.rs::benches::bench_f64_to_u8::conversions
CodSpeed Performance Gauge
0%
321.4 µs321.4 µs
f64_to_u8/clamp+TowardsZero[64000]
core/benches/conversions.rs::benches::bench_f64_to_u8::conversions
CodSpeed Performance Gauge
0%
322.6 µs322.5 µs

Commits

Click on a commit to change the comparison range
Base
main
48edf3a
+60.94%
perf: more architectures
380c284
2 hours ago
by d-v-b
0%
fix: add lifetimes
276b0da
2 hours ago
by d-v-b
© 2026 CodSpeed Technology
Home Terms Privacy Docs