RMS norm
Fused residual addition followed by RMSNorm. Computes hidden_states = residual + hidden_states, then applies RMSNorm: output = weight * (x / sqrt(mean(x^2) + eps)). This pattern occurs twice per decoder layer in transformer models.
Reported evidence · last observed 2026-06-29. Reported by source; not independently reproduced. No public source. License unknown.
Current records
Workloadsuite of 16 cases · mean latencyhidden_size = 819216 cases
Not measured on H100 for this workload. Challenges →
116.3µs80% SOL1.00×Reported · license unknown · no source2026-06-29
Measured exactly what you asked. Reported by source; not independently reproduced.
216.4µs80% SOL1.00×Reported · license unknown · no source2026-07-07
Measured exactly what you asked. Reported by source; not independently reproduced.
316.7µs78% SOL1.02×Reported · license unknown · no source2026-06-30
Measured exactly what you asked. Reported by source; not independently reproduced.
416.9µs78% SOL1.04×Reported · license unknown · no source2026-07-01
Measured exactly what you asked. Reported by source; not independently reproduced.
517.2µs79% SOL1.06×Reported · license unknown · no source2026-06-07
Measured exactly what you asked. Reported by source; not independently reproduced.
618.2µs74% SOL1.12×Reported · license unknown · no source2026-05-20
Measured exactly what you asked. Reported by source; not independently reproduced.
718.6µs73% SOL1.14×Reported · license unknown · no source2026-05-15
Measured exactly what you asked. Reported by source; not independently reproduced.
818.7µs73% SOL1.15×Reported · license unknown · no source2026-04-23
Measured exactly what you asked. Reported by source; not independently reproduced.
918.7µs73% SOL1.15×Reported · license unknown · no source2026-04-15
Measured exactly what you asked. Reported by source; not independently reproduced.
1019.0µs72% SOL1.16×Reported · license unknown · no source2026-04-14
Measured exactly what you asked. Reported by source; not independently reproduced.
1119.1µs71% SOL1.17×Reported · license unknown · no source2026-04-14
Measured exactly what you asked. Reported by source; not independently reproduced.
1219.1µs71% SOL1.17×Reported · license unknown · no source2026-04-14
Measured exactly what you asked. Reported by source; not independently reproduced.
1319.1µs72% SOL1.17×Reported · license unknown · no source2026-04-15
Measured exactly what you asked. Reported by source; not independently reproduced.
1419.3µs71% SOL1.18×Reported · license unknown · no source2026-04-15
Measured exactly what you asked. Reported by source; not independently reproduced.
1519.3µs70% SOL1.19×Reported · license unknown · no source2026-07-06
Measured exactly what you asked. Reported by source; not independently reproduced.
1620.2µs68% SOL1.24×Reported · license unknown · no source2026-04-14
Measured exactly what you asked. Reported by source; not independently reproduced.
1721.9µs65% SOL1.34×Reported · license unknown · no source2026-04-15
Measured exactly what you asked. Reported by source; not independently reproduced.
1822.6µs63% SOL1.39×Reported · license unknown · no source2026-05-08
Measured exactly what you asked. Reported by source; not independently reproduced.
1923.6µs60% SOL1.45×Reported · license unknown · no source2026-05-06
Measured exactly what you asked. Reported by source; not independently reproduced.
2024.0µs59% SOL1.47×Reported · license unknown · no source2026-05-05
Measured exactly what you asked. Reported by source; not independently reproduced.
2124.0µs58% SOL1.47×Reported · license unknown · no source2026-05-11
Measured exactly what you asked. Reported by source; not independently reproduced.
2224.1µs58% SOL1.48×Reported · license unknown · no source2026-05-11
Measured exactly what you asked. Reported by source; not independently reproduced.
2325.7µs57% SOL1.57×Reported · license unknown · no source2026-05-06
Measured exactly what you asked. Reported by source; not independently reproduced.
2428.3µs52% SOL1.73×Reported · license unknown · no source2026-04-15
Measured exactly what you asked. Reported by source; not independently reproduced.
2535.7µs44% SOL2.19×Reported · license unknown · no source2026-05-06
Measured exactly what you asked. Reported by source; not independently reproduced.
2642.1µs40% SOL2.58×Reported · license unknown · no source2026-04-15
Measured exactly what you asked. Reported by source; not independently reproduced.
27239.6µs10% SOL14.70×Reported · license unknown · no source2026-04-14
Measured exactly what you asked. Reported by source; not independently reproduced.