FP8 MoE expert linear
FP8-quantized MoE expert computation with fused gate-up projection, SiLU activation, and down projection. Uses blockwise FP8 quantization with BlockWise1x128 for activations and BlockWise128x128 for weights.
Reported evidence · last observed 2026-08-23. Reported by source; not independently reproduced. No public source. License unknown.
Current records
Workloadsuite of 16 cases · mean latencyhidden_size = 3584 · hidden_blocks = 28 · intermediate_size = 2048 · intermediate_blocks = 16 · double_intermediate_size = 4096 · double_intermediate_blocks = 3216 cases
Not measured on H100 for this workload. Challenges →
193.6µs85% SOL1.00×Reported · license unknown · no source2026-08-23
Measured exactly what you asked. Reported by source; not independently reproduced.
296.7µs84% SOL1.03×Reported · license unknown · no source2026-08-14
Measured exactly what you asked. Reported by source; not independently reproduced.
3105.3µs83% SOL1.13×Reported · license unknown · no source2026-08-01
Measured exactly what you asked. Reported by source; not independently reproduced.
4108.7µs82% SOL1.16×Reported · license unknown · no source2026-08-03
Measured exactly what you asked. Reported by source; not independently reproduced.
5110.4µs82% SOL1.18×Reported · license unknown · no source2026-08-18
Measured exactly what you asked. Reported by source; not independently reproduced.
6113.2µs81% SOL1.21×Reported · license unknown · no source2026-08-20
Measured exactly what you asked. Reported by source; not independently reproduced.
7113.4µs81% SOL1.21×Reported · license unknown · no source2026-08-12
Measured exactly what you asked. Reported by source; not independently reproduced.
8113.5µs81% SOL1.21×Reported · license unknown · no source2026-08-17
Measured exactly what you asked. Reported by source; not independently reproduced.
9113.6µs81% SOL1.21×Reported · license unknown · no source2026-08-18
Measured exactly what you asked. Reported by source; not independently reproduced.
10116.9µs81% SOL1.25×Reported · license unknown · no source2026-07-31
Measured exactly what you asked. Reported by source; not independently reproduced.
11120.7µs80% SOL1.29×Reported · license unknown · no source2026-07-10
Measured exactly what you asked. Reported by source; not independently reproduced.
12121.6µs80% SOL1.30×Reported · license unknown · no source2026-08-19
Measured exactly what you asked. Reported by source; not independently reproduced.
13135.0µs78% SOL1.44×Reported · license unknown · no source2026-07-10
Measured exactly what you asked. Reported by source; not independently reproduced.
14138.7µs77% SOL1.48×Reported · license unknown · no source2026-08-07
Measured exactly what you asked. Reported by source; not independently reproduced.
15149.4µs76% SOL1.60×Reported · license unknown · no source2026-07-19
Measured exactly what you asked. Reported by source; not independently reproduced.
16150.7µs76% SOL1.61×Reported · license unknown · no source2026-07-10
Measured exactly what you asked. Reported by source; not independently reproduced.
17156.1µs75% SOL1.67×Reported · license unknown · no source2026-07-10
Measured exactly what you asked. Reported by source; not independently reproduced.
18167.3µs73% SOL1.79×Reported · license unknown · no source2026-08-03
Measured exactly what you asked. Reported by source; not independently reproduced.
19177.4µs72% SOL1.90×Reported · license unknown · no source2026-08-05
Measured exactly what you asked. Reported by source; not independently reproduced.
20212.8µs68% SOL2.27×Reported · license unknown · no source2026-08-02
Measured exactly what you asked. Reported by source; not independently reproduced.
21608.8µs41% SOL6.51×Reported · license unknown · no source2026-07-10
Measured exactly what you asked. Reported by source; not independently reproduced.
222.50ms14% SOL26.75×Reported · license unknown · no source2026-07-10
Measured exactly what you asked. Reported by source; not independently reproduced.