Skip to content

reduction subgroup

Sums 65,536 bounded u32 inputs into 256 independent 256-element segment sums. Uses subgroupAdd with one workgroup atomic addition per subgroup; requires subgroups, assumes no subgroup width. Reports input elements/s; all variants produce complete segment sums.

Measured in element/s. Higher values mean better performance. Benchmark your device.

Performance distribution

 
4.98 Melement/s565 Melement/s64.1 Gelement/s

Performance thresholds

99% of results achieve at least 6.51 Melement/s

95% of results achieve at least 231 Melement/s

90% of results achieve at least 756 Melement/s

Median (50th percentile) 17.6 Gelement/s

Best performer

Highest recorded performance: 64.1 Gelement/s. View benchmark profile.

Recorded 10/3/2026, 4:01:21 PM

WebGPU device
Vendornvidia
Device—
Architectureblackwell
Description—
WebGL device
VendorGoogle Inc. (NVIDIA)
RendererANGLE (NVIDIA, NVIDIA GeForce RTX 5080 (0x00002C02) Direct3D11 vs_5_0 ps_5_0, D3D11)
Browser
NameEdge
Version154.0.0.0
Operating system
NameWindows
Version10