MoE expert MLP with load balancing
Complete MoE expert computation including gate routing, top-k selection, expert MLPs with SwiGLU activation, and weighted aggregation. Processes tokens through 64 experts with top-8 selection. Each expert is a 3-layer MLP (gate_proj, up_proj, down_proj) with SwiGLU activation.
Reported evidence · last observed 2026-08-16. Reported by source; not independently reproduced. No public source. License unknown.
Current records
Workloadsuite of 16 cases · mean latencyhidden_size = 4096 · num_experts = 64 · num_experts_per_tok = 8 · moe_intermediate_size = 256016 cases
Not measured on H100 for this workload. Challenges →
1713.5µs90% SOL1.00×Reported · license unknown · no source2026-08-16
Measured exactly what you asked. Reported by source; not independently reproduced.
2768.5µs96% SOL1.08×Reported · license unknown · no source2026-07-25
Measured exactly what you asked. Reported by source; not independently reproduced.
3804.9µs93% SOL1.13×Reported · license unknown · no source2026-08-21
Measured exactly what you asked. Reported by source; not independently reproduced.
4978.1µs85% SOL1.37×Reported · license unknown · no source2026-07-17
Measured exactly what you asked. Reported by source; not independently reproduced.
5992.0µs80% SOL1.39×Reported · license unknown · no source2026-08-08
Measured exactly what you asked. Reported by source; not independently reproduced.
61.00ms82% SOL1.40×Reported · license unknown · no source2026-07-16
Measured exactly what you asked. Reported by source; not independently reproduced.
71.00ms82% SOL1.41×Reported · license unknown · no source2026-08-15
Measured exactly what you asked. Reported by source; not independently reproduced.
81.00ms82% SOL1.41×Reported · license unknown · no source2026-08-15
Measured exactly what you asked. Reported by source; not independently reproduced.
91.00ms82% SOL1.41×Reported · license unknown · no source2026-08-15
Measured exactly what you asked. Reported by source; not independently reproduced.
101.01ms79% SOL1.42×Reported · license unknown · no source2026-07-26
Measured exactly what you asked. Reported by source; not independently reproduced.
111.02ms82% SOL1.42×Reported · license unknown · no source2026-07-10
Measured exactly what you asked. Reported by source; not independently reproduced.
121.15ms76% SOL1.61×Reported · license unknown · no source2026-07-10
Measured exactly what you asked. Reported by source; not independently reproduced.
131.24ms72% SOL1.74×Reported · license unknown · no source2026-07-19
Measured exactly what you asked. Reported by source; not independently reproduced.
141.33ms68% SOL1.86×Reported · license unknown · no source2026-07-12
Measured exactly what you asked. Reported by source; not independently reproduced.
151.38ms67% SOL1.94×Reported · license unknown · no source2026-07-27
Measured exactly what you asked. Reported by source; not independently reproduced.
161.65ms58% SOL2.31×Reported · license unknown · no source2026-07-12
Measured exactly what you asked. Reported by source; not independently reproduced.
171.97ms51% SOL2.76×Reported · license unknown · no source2026-07-23
Measured exactly what you asked. Reported by source; not independently reproduced.
182.30ms45% SOL3.22×Reported · license unknown · no source2026-07-10
Measured exactly what you asked. Reported by source; not independently reproduced.
192.32ms45% SOL3.25×Reported · license unknown · no source2026-07-10
Measured exactly what you asked. Reported by source; not independently reproduced.
202.75ms39% SOL3.85×Reported · license unknown · no source2026-07-23
Measured exactly what you asked. Reported by source; not independently reproduced.