Mamba SSM dt projection softplus clamp backward
Backward pass for Mamba SSM dt projection with softplus and clamp. Computes gradients through clamp, softplus, and bias addition operations.
Reported evidence · last observed 2026-08-15. Reported by source; not independently reproduced. No public source. License unknown.
Current records
Workloadsuite of 16 cases · mean latencynum_heads = 4016 cases
Not measured on H100 for this workload. Challenges →
15.84µs86% SOL1.00×Reported · license unknown · no source2026-08-15
Measured exactly what you asked. Reported by source; not independently reproduced.
25.86µs86% SOL1.00×Reported · license unknown · no source2026-07-16
Measured exactly what you asked. Reported by source; not independently reproduced.
36.22µs85% SOL1.06×Reported · license unknown · no source2026-07-10
Measured exactly what you asked. Reported by source; not independently reproduced.
46.52µs84% SOL1.11×Reported · license unknown · no source2026-08-24
Measured exactly what you asked. Reported by source; not independently reproduced.
56.54µs84% SOL1.12×Reported · license unknown · no source2026-07-10
Measured exactly what you asked. Reported by source; not independently reproduced.
66.66µs84% SOL1.14×Reported · license unknown · no source2026-08-06
Measured exactly what you asked. Reported by source; not independently reproduced.
76.92µs83% SOL1.18×Reported · license unknown · no source2026-08-23
Measured exactly what you asked. Reported by source; not independently reproduced.
87.09µs83% SOL1.21×Reported · license unknown · no source2026-07-10
Measured exactly what you asked. Reported by source; not independently reproduced.
97.30µs83% SOL1.25×Reported · license unknown · no source2026-08-03
Measured exactly what you asked. Reported by source; not independently reproduced.
107.51µs82% SOL1.29×Reported · license unknown · no source2026-07-10
Measured exactly what you asked. Reported by source; not independently reproduced.
117.53µs82% SOL1.29×Reported · license unknown · no source2026-08-23
Measured exactly what you asked. Reported by source; not independently reproduced.
127.56µs82% SOL1.29×Reported · license unknown · no source2026-08-24
Measured exactly what you asked. Reported by source; not independently reproduced.
137.71µs81% SOL1.32×Reported · license unknown · no source2026-07-10
Measured exactly what you asked. Reported by source; not independently reproduced.
148.56µs80% SOL1.47×Reported · license unknown · no source2026-08-12
Measured exactly what you asked. Reported by source; not independently reproduced.
159.01µs79% SOL1.54×Reported · license unknown · no source2026-08-10
Measured exactly what you asked. Reported by source; not independently reproduced.
1611.3µs74% SOL1.93×Reported · license unknown · no source2026-07-23
Measured exactly what you asked. Reported by source; not independently reproduced.
1712.3µs72% SOL2.11×Reported · license unknown · no source2026-07-10
Measured exactly what you asked. Reported by source; not independently reproduced.