Text decoder layer with self attention and MLP
Complete Mllama text decoder layer combining self-attention with RoPE, RMSNorm, gated MLP, and residual connections. This is the dominant computational pattern in the text model (32 out of 40 layers). Architecture: (1) RMSNorm + self-attention with RoPE and GQA (32 heads, 8 KV heads) + residual, (2) RMSNorm + gated MLP with SiLU + residual.
Reported evidence · last observed 2026-08-14. Reported by source; not independently reproduced. No public source. License unknown.
Current records
Workloadsuite of 16 cases · mean latencyhead_dim = 128 · num_heads = 32 · hidden_size = 4096 · kv_head_dim = 1024 · qkv_head_dim = 4096 · intermediate_size = 14336 · num_key_value_heads = 8 · num_key_value_groups = 416 cases
Not measured on H100 for this workload. Challenges →
1669.2µs72% SOL1.00×Reported · license unknown · no source2026-08-14
Measured exactly what you asked. Reported by source; not independently reproduced.
2674.2µs76% SOL1.01×Reported · license unknown · no source2026-08-23
Measured exactly what you asked. Reported by source; not independently reproduced.
3700.6µs69% SOL1.05×Reported · license unknown · no source2026-08-05
Measured exactly what you asked. Reported by source; not independently reproduced.
4701.4µs75% SOL1.05×Reported · license unknown · no source2026-08-03
Measured exactly what you asked. Reported by source; not independently reproduced.
5704.5µs69% SOL1.05×Reported · license unknown · no source2026-08-23
Measured exactly what you asked. Reported by source; not independently reproduced.
6705.1µs73% SOL1.05×Reported · license unknown · no source2026-08-20
Measured exactly what you asked. Reported by source; not independently reproduced.
7709.9µs70% SOL1.06×Reported · license unknown · no source2026-08-02
Measured exactly what you asked. Reported by source; not independently reproduced.
8711.0µs70% SOL1.06×Reported · license unknown · no source2026-07-10
Measured exactly what you asked. Reported by source; not independently reproduced.
9711.3µs72% SOL1.06×Reported · license unknown · no source2026-08-19
Measured exactly what you asked. Reported by source; not independently reproduced.
10716.2µs68% SOL1.07×Reported · license unknown · no source2026-07-17
Measured exactly what you asked. Reported by source; not independently reproduced.
11717.7µs72% SOL1.07×Reported · license unknown · no source2026-08-18
Measured exactly what you asked. Reported by source; not independently reproduced.
12726.2µs68% SOL1.09×Reported · license unknown · no source2026-08-02
Measured exactly what you asked. Reported by source; not independently reproduced.
13738.6µs67% SOL1.10×Reported · license unknown · no source2026-07-15
Measured exactly what you asked. Reported by source; not independently reproduced.
14754.3µs69% SOL1.13×Reported · license unknown · no source2026-07-10
Measured exactly what you asked. Reported by source; not independently reproduced.
15759.3µs66% SOL1.13×Reported · license unknown · no source2026-08-22
Measured exactly what you asked. Reported by source; not independently reproduced.
16762.1µs66% SOL1.14×Reported · license unknown · no source2026-08-21
Measured exactly what you asked. Reported by source; not independently reproduced.
17769.2µs65% SOL1.15×Reported · license unknown · no source2026-07-11
Measured exactly what you asked. Reported by source; not independently reproduced.
18806.2µs63% SOL1.20×Reported · license unknown · no source2026-08-20
Measured exactly what you asked. Reported by source; not independently reproduced.
19823.4µs62% SOL1.23×Reported · license unknown · no source2026-07-25
Measured exactly what you asked. Reported by source; not independently reproduced.
20823.6µs62% SOL1.23×Reported · license unknown · no source2026-08-17
Measured exactly what you asked. Reported by source; not independently reproduced.
21842.7µs60% SOL1.26×Reported · license unknown · no source2026-07-10
Measured exactly what you asked. Reported by source; not independently reproduced.
22844.5µs62% SOL1.26×Reported · license unknown · no source2026-08-21
Measured exactly what you asked. Reported by source; not independently reproduced.
23855.8µs60% SOL1.28×Reported · license unknown · no source2026-08-17
Measured exactly what you asked. Reported by source; not independently reproduced.
24884.6µs59% SOL1.32×Reported · license unknown · no source2026-08-18
Measured exactly what you asked. Reported by source; not independently reproduced.
25889.8µs58% SOL1.33×Reported · license unknown · no source2026-07-10
Measured exactly what you asked. Reported by source; not independently reproduced.
26896.9µs58% SOL1.34×Reported · license unknown · no source2026-08-18
Measured exactly what you asked. Reported by source; not independently reproduced.
27925.7µs56% SOL1.38×Reported · license unknown · no source2026-08-20
Measured exactly what you asked. Reported by source; not independently reproduced.
28970.6µs55% SOL1.45×Reported · license unknown · no source2026-08-13
Measured exactly what you asked. Reported by source; not independently reproduced.
292.83ms21% SOL4.22×Reported · license unknown · no source2026-08-13
Measured exactly what you asked. Reported by source; not independently reproduced.