BF16 Matmul Throughput Explorer

Interactive view of BF16 TFLOPs for matrix multiply shapes (m × k · k × n)

Kernel 6.17.3-3-cachyos-deckify torch 2.10.0a0+rocm7.10.0a20251011 hip 7.1.25406-c11ca311d7 Device AMD Radeon Graphics (gfx1151 · 62.2 GB · 20 CUs)

Loading benchmark data…

Click a data point to inspect its TFLOPs.