GQA ragged prefill causal h5 kv1 d128
0 eligible runs
gqa-ragged-attention
Batched Grouped Query Attention prefill with ragged (variable-length) inputs. Causal mask is applied. Captured from Llama 4 Scout 17B-16E with tensor parallel size 8 (40/8=5 q-heads, 8/8=1 kv-head).
Current records
No published measurement for the selected workload.
Implementations
Implementation
Runtime
Best latency
Evidence
Availability
Semantics
Inputs and outputs
qbf16 [total_q, num_qo_heads, head_dim]
kbf16 [total_kv, num_kv_heads, head_dim]
vbf16 [total_kv, num_kv_heads, head_dim]
qo_indptrint32 [len_indptr]
kv_indptrint32 [len_indptr]
sm_scalefp32 scalar
outputbf16 [total_q, num_qo_heads, head_dim]
lsefp32 [total_q, num_qo_heads]
Axes and behavior
total_qvariable
head_dimconstant = 128
total_kvvariable
len_indptrvariable
num_kv_headsconstant = 1
num_qo_headsconstant = 5
determinismunspecified
constraintsNo mutation or aliasing
Identity
No source imports for this operation yet.How records are decidedJSON