GQA paged decode h32 kv4 d128 ps1
Batched Grouped Query Attention decode with a paged KV cache. Captured from Qwen3-30B-A3B.
Reported evidence · last observed 2026-08-21. Reported by source; not independently reproduced. No public source. License unknown.
Current records
Workloadsuite of 48 cases · mean latency96 cases
Not measured on H100 for this workload. Challenges →
112.9µs94% SOL1.00×Reported · license unknown · no source2026-08-21
Measured exactly what you asked. Reported by source; not independently reproduced.
213.0µs94% SOL1.00×Reported · license unknown · no source2026-08-13
Measured exactly what you asked. Reported by source; not independently reproduced.
313.7µs93% SOL1.06×Reported · license unknown · no source2026-08-09
Measured exactly what you asked. Reported by source; not independently reproduced.
413.8µs93% SOL1.07×Reported · license unknown · no source2026-08-03
Measured exactly what you asked. Reported by source; not independently reproduced.
514.7µs93% SOL1.14×Reported · license unknown · no source2026-07-26
Measured exactly what you asked. Reported by source; not independently reproduced.
615.3µs93% SOL1.18×Reported · license unknown · no source2026-07-10
Measured exactly what you asked. Reported by source; not independently reproduced.
715.7µs92% SOL1.22×Reported · license unknown · no source2026-07-29
Measured exactly what you asked. Reported by source; not independently reproduced.
816.2µs92% SOL1.25×Reported · license unknown · no source2026-07-10
Measured exactly what you asked. Reported by source; not independently reproduced.
916.3µs92% SOL1.26×Reported · license unknown · no source2026-07-10
Measured exactly what you asked. Reported by source; not independently reproduced.
1016.5µs92% SOL1.28×Reported · license unknown · no source2026-07-30
Measured exactly what you asked. Reported by source; not independently reproduced.
1117.0µs91% SOL1.31×Reported · license unknown · no source2026-07-21
Measured exactly what you asked. Reported by source; not independently reproduced.
1217.4µs90% SOL1.35×Reported · license unknown · no source2026-08-02
Measured exactly what you asked. Reported by source; not independently reproduced.
1317.6µs90% SOL1.36×Reported · license unknown · no source2026-08-01
Measured exactly what you asked. Reported by source; not independently reproduced.
1418.4µs90% SOL1.42×Reported · license unknown · no source2026-08-04
Measured exactly what you asked. Reported by source; not independently reproduced.
1518.5µs90% SOL1.43×Reported · license unknown · no source2026-07-18
Measured exactly what you asked. Reported by source; not independently reproduced.
1618.5µs91% SOL1.43×Reported · license unknown · no source2026-08-21
Measured exactly what you asked. Reported by source; not independently reproduced.
1718.6µs90% SOL1.44×Reported · license unknown · no source2026-07-31
Measured exactly what you asked. Reported by source; not independently reproduced.
1818.7µs90% SOL1.44×Reported · license unknown · no source2026-07-10
Measured exactly what you asked. Reported by source; not independently reproduced.
1918.8µs89% SOL1.45×Reported · license unknown · no source2026-08-02
Measured exactly what you asked. Reported by source; not independently reproduced.
2018.8µs89% SOL1.45×Reported · license unknown · no source2026-08-04
Measured exactly what you asked. Reported by source; not independently reproduced.
2118.9µs89% SOL1.46×Reported · license unknown · no source2026-07-31
Measured exactly what you asked. Reported by source; not independently reproduced.
2218.9µs89% SOL1.46×Reported · license unknown · no source2026-07-31
Measured exactly what you asked. Reported by source; not independently reproduced.
2319.0µs90% SOL1.47×Reported · license unknown · no source2026-07-10
Measured exactly what you asked. Reported by source; not independently reproduced.
2419.2µs91% SOL1.48×Reported · license unknown · no source2026-07-16
Measured exactly what you asked. Reported by source; not independently reproduced.
2519.3µs89% SOL1.49×Reported · license unknown · no source2026-07-30
Measured exactly what you asked. Reported by source; not independently reproduced.
2619.5µs89% SOL1.51×Reported · license unknown · no source2026-08-01
Measured exactly what you asked. Reported by source; not independently reproduced.
2719.6µs89% SOL1.52×Reported · license unknown · no source2026-07-30
Measured exactly what you asked. Reported by source; not independently reproduced.
2819.7µs89% SOL1.52×Reported · license unknown · no source2026-07-29
Measured exactly what you asked. Reported by source; not independently reproduced.
2926.9µs85% SOL2.08×Reported · license unknown · no source2026-07-29
Measured exactly what you asked. Reported by source; not independently reproduced.
3028.8µs83% SOL2.23×Reported · license unknown · no source2026-07-10
Measured exactly what you asked. Reported by source; not independently reproduced.
3138.3µs79% SOL2.96×Reported · license unknown · no source2026-07-30
Measured exactly what you asked. Reported by source; not independently reproduced.
3239.9µs79% SOL3.08×Reported · license unknown · no source2026-07-21
Measured exactly what you asked. Reported by source; not independently reproduced.
3342.7µs77% SOL3.30×Reported · license unknown · no source2026-07-19
Measured exactly what you asked. Reported by source; not independently reproduced.
3442.9µs77% SOL3.31×Reported · license unknown · no source2026-07-15
Measured exactly what you asked. Reported by source; not independently reproduced.
3544.8µs76% SOL3.46×Reported · license unknown · no source2026-07-10
Measured exactly what you asked. Reported by source; not independently reproduced.
36114.1µs58% SOL8.81×Reported · license unknown · no source2026-07-10
Measured exactly what you asked. Reported by source; not independently reproduced.
37281.1µs42% SOL21.71×Reported · license unknown · no source2026-08-21
Measured exactly what you asked. Reported by source; not independently reproduced.
381.93ms12% SOL149.26×Reported · license unknown · no source2026-07-30
Measured exactly what you asked. Reported by source; not independently reproduced.
Implementations
Show all 62 implementations ›Showing all 62 implementations ⌄
Semantics
Also indexed as 012-gqa-paged-decode-h32-kv4-d128-ps1 · reviewed as the same computation; every definition's cohorts are shown here, ranked separately.