tile shared padded
The same 3x3 box stencil on 256 independent 16x16 tiles, wrapping within each tile. Cooperatively stages the tile in workgroup memory with row stride 17; includes loading and synchronization. Reports outputs/s. A fixed reuse pattern, not a reuse sweep.
Measured in output/s. Higher values mean better performance. Benchmark your device.
Performance distribution
649 koutput/s204 Moutput/s63.9 Goutput/s
649 koutput/s204 Moutput/s63.9 Goutput/s
Performance thresholds
99% of results achieve at least 3.96 Moutput/s
95% of results achieve at least 148 Moutput/s
90% of results achieve at least 414 Moutput/s
Median (50th percentile) 5.19 Goutput/s
Best performer
Highest recorded performance: 63.9 Goutput/s. View benchmark profile.
Recorded 9/23/2026, 4:10:52 AM
WebGPU device
Vendornvidia
Device—
Architectureblackwell
Description—
WebGL device
VendorGoogle Inc. (NVIDIA)
RendererANGLE (NVIDIA, NVIDIA GeForce RTX 5070 Ti (0x00002C05) Direct3D11 vs_5_0 ps_5_0, D3D11)
Browser
NameChrome
Version153.0.0.0
Operating system
NameWindows
Version10