Skip to content
KernelIndex
Search⌘K
Sign in

GEMM n4096 k14336

320 eligible runs
gemm

General matrix multiply (GEMM) C = A @ B.T. Captured from Llama 3.1 8B mlp.down_proj.

Fastest reported
38.6µsmean · 59% of SOL
RIAC as wellLicense unknown · unknown

Reported evidence · last observed 2026-08-12. Reported by source; not independently reproduced. No public source. License unknown.

No public sourceRun detail →

Current records

Workloadsuite of 43 cases · mean latency86 cases
m
k
n
dtype
1——fp161143364096fp162——fp162143364096fp164——fp164143364096fp167——fp167143364096fp168——fp168143364096fp1615——fp1615143364096fp1616——fp1616143364096fp1624——fp1624143364096fp1632——fp1632143364096fp1635——fp1635143364096fp1640——fp1640143364096fp1648——fp1648143364096fp1656——fp1656143364096fp1664——fp1664143364096fp1670——fp1670143364096fp1672——fp1672143364096fp1680——fp1680143364096fp1688——fp1688143364096fp1696——fp1696143364096fp16104——fp16104143364096fp16112——fp16112143364096fp16120——fp16120143364096fp16128——fp16128143364096fp16136——fp16136143364096fp16144——fp16144143364096fp16152——fp16152143364096fp16160——fp16160143364096fp16168——fp16168143364096fp16176——fp16176143364096fp16184——fp16184143364096fp16192——fp16192143364096fp16200——fp16200143364096fp16208——fp16208143364096fp16216——fp16216143364096fp16224——fp16224143364096fp16232——fp16232143364096fp16240——fp16240143364096fp16248——fp16248143364096fp16256——fp16256143364096fp16972——fp16972143364096fp162053——fp162053143364096fp162379——fp162379143364096fp168192——fp168192143364096fp16
scrolls · 86 cases total
Aggregatessuite of 43 cases · mean latency
HardwareBest knownImplementationRuns
NVIDIA B200 · env 138.6 µsRIAC as well47NVIDIA B200 · env 239.9 µsRecursive15

Not measured on H100 for this workload. Challenges →

Source-native comparison · GPU NVIDIA B200 · Workload suite of 43 cases · mean latency · Protocol SOL-ExecBench evaluation stack v1.1 · mean · 47 results · last observed 2026-08-23Record history →
#
Implementation
Latency
vs #1
Trust
Observed
1
RIAC as well
38.6µs59% SOL
1.00×
Reported · license unknown · no source
2026-08-12

Measured exactly what you asked. Reported by source; not independently reproduced.

no sourcelicense unknownno install recipeRun detail →
2
Geometric
38.8µs59% SOL
1.00×
Reported · license unknown · no source
2026-08-21

Measured exactly what you asked. Reported by source; not independently reproduced.

no sourcelicense unknownno install recipeRun detail →
3
Geometric
39.0µs59% SOL
1.01×
Reported · license unknown · no source
2026-08-10

Measured exactly what you asked. Reported by source; not independently reproduced.

no sourcelicense unknownno install recipeRun detail →
4
Amir M. Mir | SF Tensor
39.0µs59% SOL
1.01×
Reported · license unknown · no source
2026-08-07

Measured exactly what you asked. Reported by source; not independently reproduced.

no sourcelicense unknownno install recipeRun detail →
5
Amir M. Mir | SF Tensor
39.2µs58% SOL
1.02×
Reported · license unknown · no source
2026-08-02

Measured exactly what you asked. Reported by source; not independently reproduced.

no sourcelicense unknownno install recipeRun detail →
6
Geometric
39.5µs58% SOL
1.02×
Reported · license unknown · no source
2026-08-21

Measured exactly what you asked. Reported by source; not independently reproduced.

no sourcelicense unknownno install recipeRun detail →
7
Geometric
39.6µs58% SOL
1.02×
Reported · license unknown · no source
2026-08-01

Measured exactly what you asked. Reported by source; not independently reproduced.

no sourcelicense unknownno install recipeRun detail →
8
Geometric
39.6µs58% SOL
1.03×
Reported · license unknown · no source
2026-08-23

Measured exactly what you asked. Reported by source; not independently reproduced.

no sourcelicense unknownno install recipeRun detail →
9
Amir M. Mir | SF Tensor
39.6µs58% SOL
1.03×
Reported · license unknown · no source
2026-08-02

Measured exactly what you asked. Reported by source; not independently reproduced.

no sourcelicense unknownno install recipeRun detail →
10
Geometric
39.7µs58% SOL
1.03×
Reported · license unknown · no source
2026-08-04

Measured exactly what you asked. Reported by source; not independently reproduced.

no sourcelicense unknownno install recipeRun detail →
11
Qwen
40.4µs57% SOL
1.04×
Reported · license unknown · no source
2026-08-02

Measured exactly what you asked. Reported by source; not independently reproduced.

no sourcelicense unknownno install recipeRun detail →
12
Geometric
40.4µs57% SOL
1.05×
Reported · license unknown · no source
2026-07-30

Measured exactly what you asked. Reported by source; not independently reproduced.

no sourcelicense unknownno install recipeRun detail →
13
Qwen
40.5µs57% SOL
1.05×
Reported · license unknown · no source
2026-07-30

Measured exactly what you asked. Reported by source; not independently reproduced.

no sourcelicense unknownno install recipeRun detail →
14
Recursive
40.6µs57% SOL
1.05×
Reported · license unknown · no source
2026-07-10

Measured exactly what you asked. Reported by source; not independently reproduced.

no sourcelicense unknownno install recipeRun detail →
15
Geometric
40.6µs57% SOL
1.05×
Reported · license unknown · no source
2026-08-03

Measured exactly what you asked. Reported by source; not independently reproduced.

no sourcelicense unknownno install recipeRun detail →
16
Geometric
40.7µs57% SOL
1.05×
Reported · license unknown · no source
2026-08-01

Measured exactly what you asked. Reported by source; not independently reproduced.

no sourcelicense unknownno install recipeRun detail →
17
Geometric
40.8µs57% SOL
1.06×
Reported · license unknown · no source
2026-08-01

Measured exactly what you asked. Reported by source; not independently reproduced.

no sourcelicense unknownno install recipeRun detail →
18
Geometric
40.8µs56% SOL
1.06×
Reported · license unknown · no source
2026-08-03

Measured exactly what you asked. Reported by source; not independently reproduced.

no sourcelicense unknownno install recipeRun detail →
19
Geometric
40.9µs56% SOL
1.06×
Reported · license unknown · no source
2026-07-30

Measured exactly what you asked. Reported by source; not independently reproduced.

no sourcelicense unknownno install recipeRun detail →
20
Geometric
41.0µs56% SOL
1.06×
Reported · license unknown · no source
2026-07-29

Measured exactly what you asked. Reported by source; not independently reproduced.

no sourcelicense unknownno install recipeRun detail →
21
Geometric
41.0µs56% SOL
1.06×
Reported · license unknown · no source
2026-08-01

Measured exactly what you asked. Reported by source; not independently reproduced.

no sourcelicense unknownno install recipeRun detail →
22
Geometric
41.0µs56% SOL
1.06×
Reported · license unknown · no source
2026-08-01

Measured exactly what you asked. Reported by source; not independently reproduced.

no sourcelicense unknownno install recipeRun detail →
23
ac4k
41.0µs56% SOL
1.06×
Reported · license unknown · no source
2026-08-01

Measured exactly what you asked. Reported by source; not independently reproduced.

no sourcelicense unknownno install recipeRun detail →
24
Geometric
41.0µs56% SOL
1.06×
Reported · license unknown · no source
2026-07-27

Measured exactly what you asked. Reported by source; not independently reproduced.

no sourcelicense unknownno install recipeRun detail →
25
Geometric
41.0µs56% SOL
1.06×
Reported · license unknown · no source
2026-07-24

Measured exactly what you asked. Reported by source; not independently reproduced.

no sourcelicense unknownno install recipeRun detail →
26
ac4k
41.0µs56% SOL
1.06×
Reported · license unknown · no source
2026-07-24

Measured exactly what you asked. Reported by source; not independently reproduced.

no sourcelicense unknownno install recipeRun detail →
27
Geometric
41.0µs56% SOL
1.06×
Reported · license unknown · no source
2026-08-03

Measured exactly what you asked. Reported by source; not independently reproduced.

no sourcelicense unknownno install recipeRun detail →
28
Geometric
41.1µs56% SOL
1.06×
Reported · license unknown · no source
2026-07-30

Measured exactly what you asked. Reported by source; not independently reproduced.

no sourcelicense unknownno install recipeRun detail →
29
Geometric
41.1µs56% SOL
1.06×
Reported · license unknown · no source
2026-08-01

Measured exactly what you asked. Reported by source; not independently reproduced.

no sourcelicense unknownno install recipeRun detail →
30
Geometric
41.2µs56% SOL
1.07×
Reported · license unknown · no source
2026-08-03

Measured exactly what you asked. Reported by source; not independently reproduced.

no sourcelicense unknownno install recipeRun detail →
31
Geometric
41.2µs56% SOL
1.07×
Reported · license unknown · no source
2026-07-30

Measured exactly what you asked. Reported by source; not independently reproduced.

no sourcelicense unknownno install recipeRun detail →
32
Geometric
41.3µs56% SOL
1.07×
Reported · license unknown · no source
2026-08-03

Measured exactly what you asked. Reported by source; not independently reproduced.

no sourcelicense unknownno install recipeRun detail →
33
Geometric
41.3µs56% SOL
1.07×
Reported · license unknown · no source
2026-08-03

Measured exactly what you asked. Reported by source; not independently reproduced.

no sourcelicense unknownno install recipeRun detail →
34
Geometric
41.5µs56% SOL
1.07×
Reported · license unknown · no source
2026-08-02

Measured exactly what you asked. Reported by source; not independently reproduced.

no sourcelicense unknownno install recipeRun detail →
35
Geometric
41.6µs55% SOL
1.08×
Reported · license unknown · no source
2026-07-24

Measured exactly what you asked. Reported by source; not independently reproduced.

no sourcelicense unknownno install recipeRun detail →
36
Coral@HumanAgentSociety
41.8µs55% SOL
1.08×
Reported · license unknown · no source
2026-07-10

Measured exactly what you asked. Reported by source; not independently reproduced.

no sourcelicense unknownno install recipeRun detail →
37
Geometric
42.0µs55% SOL
1.09×
Reported · license unknown · no source
2026-07-31

Measured exactly what you asked. Reported by source; not independently reproduced.

no sourcelicense unknownno install recipeRun detail →
38
doubleAI
42.2µs54% SOL
1.09×
Reported · license unknown · no source
2026-07-10

Measured exactly what you asked. Reported by source; not independently reproduced.

no sourcelicense unknownno install recipeRun detail →
39
Hyra
42.6µs54% SOL
1.10×
Reported · license unknown · no source
2026-07-18

Measured exactly what you asked. Reported by source; not independently reproduced.

no sourcelicense unknownno install recipeRun detail →
40
OpenKernels-Ruibo
43.2µs53% SOL
1.12×
Reported · license unknown · no source
2026-07-10

Measured exactly what you asked. Reported by source; not independently reproduced.

no sourcelicense unknownno install recipeRun detail →
41
Argus-sufeng
43.4µs53% SOL
1.12×
Reported · license unknown · no source
2026-07-10

Measured exactly what you asked. Reported by source; not independently reproduced.

no sourcelicense unknownno install recipeRun detail →
42
Amir M. Mir | SF Tensor
43.8µs53% SOL
1.13×
Reported · license unknown · no source
2026-07-28

Measured exactly what you asked. Reported by source; not independently reproduced.

no sourcelicense unknownno install recipeRun detail →
43
Amir M. Mir | SF Tensor
45.1µs52% SOL
1.17×
Reported · license unknown · no source
2026-07-20

Measured exactly what you asked. Reported by source; not independently reproduced.

no sourcelicense unknownno install recipeRun detail →
44
Stellar Flamingo
45.2µs51% SOL
1.17×
Reported · license unknown · no source
2026-07-10

Measured exactly what you asked. Reported by source; not independently reproduced.

no sourcelicense unknownno install recipeRun detail →
45
Theoric.com
47.8µs49% SOL
1.24×
Reported · license unknown · no source
2026-07-28

Measured exactly what you asked. Reported by source; not independently reproduced.

no sourcelicense unknownno install recipeRun detail →
46
Amir M. Mir | SF Tensor
47.8µs49% SOL
1.24×
Reported · license unknown · no source
2026-07-10

Measured exactly what you asked. Reported by source; not independently reproduced.

no sourcelicense unknownno install recipeRun detail →
47
Cursor
47.9µs49% SOL
1.24×
Reported · license unknown · no source
2026-07-10

Measured exactly what you asked. Reported by source; not independently reproduced.

no sourcelicense unknownno install recipeRun detail →

Implementations

Implementation
Runtime
Best latency
Evidence
Availability
RIAC as well
unknown
38.6µs
1.09×
Reported
license unknown · no source
Geometric
unknown
39.0µs
1.10×
Reported
license unknown · no source
Amir M. Mir | SF Tensor
unknown
39.0µs
1.10×
Reported
license unknown · no source
Qwen
unknown
40.5µs
1.14×
Reported
license unknown · no source
Geometric
unknown
38.8µs
1.09×
Reported
license unknown · no source
Recursive
unknown
39.9µs
1.12×
Reported
license unknown · no source
Qwen
unknown
40.4µs
1.13×
Reported
license unknown · no source
doubleAI
unknown
41.2µs
1.16×
Reported
license unknown · no source
Show all 71 implementations ›
ac4k
unknown
41.0µs
1.15×
Reported
license unknown · no source
Coral@HumanAgentSociety
unknown
41.1µs
1.16×
Reported
license unknown · no source
OpenKernels-Ruibo
unknown
42.3µs
1.19×
Reported
license unknown · no source
Argus-sufeng
unknown
42.6µs
1.20×
Reported
license unknown · no source
Hyra
unknown
42.6µs
1.20×
Reported
license unknown · no source
ZCode
unknown
43.4µs
1.22×
Reported
license unknown · no source
Stellar Flamingo
unknown
43.7µs
1.23×
Reported
license unknown · no source
Cursor
unknown
47.0µs
1.32×
Reported
license unknown · no source
Theoric.com
unknown
47.8µs
1.34×
Reported
license unknown · no source
Amir M. Mir | SF Tensor
unknown
39.2µs
1.10×
Reported
license unknown · no source
Geometric
unknown
39.6µs
1.11×
Reported
license unknown · no source
Geometric
unknown
39.5µs
1.11×
Reported
license unknown · no source
Amir M. Mir | SF Tensor
unknown
39.6µs
1.11×
Reported
license unknown · no source
Geometric
unknown
39.6µs
1.11×
Reported
license unknown · no source
Geometric
unknown
39.7µs
1.11×
Reported
license unknown · no source
Geometric
unknown
40.4µs
1.14×
Reported
license unknown · no source
Geometric
unknown
40.6µs
1.14×
Reported
license unknown · no source
Geometric
unknown
40.7µs
1.14×
Reported
license unknown · no source
Recursive
unknown
40.6µs
1.14×
Reported
license unknown · no source
Geometric
unknown
40.8µs
1.15×
Reported
license unknown · no source
Geometric
unknown
40.8µs
1.15×
Reported
license unknown · no source
Geometric
unknown
40.9µs
1.15×
Reported
license unknown · no source
Geometric
unknown
41.0µs
1.15×
Reported
license unknown · no source
Geometric
unknown
41.0µs
1.15×
Reported
license unknown · no source
Geometric
unknown
41.0µs
1.15×
Reported
license unknown · no source
Geometric
unknown
41.0µs
1.15×
Reported
license unknown · no source
Geometric
unknown
41.0µs
1.15×
Reported
license unknown · no source
Geometric
unknown
41.1µs
1.16×
Reported
license unknown · no source
Geometric
unknown
41.0µs
1.15×
Reported
license unknown · no source
Geometric
unknown
41.1µs
1.16×
Reported
license unknown · no source
ac4k
unknown
41.0µs
1.15×
Reported
license unknown · no source
Geometric
unknown
41.2µs
1.16×
Reported
license unknown · no source
Geometric
unknown
41.3µs
1.16×
Reported
license unknown · no source
Geometric
unknown
41.2µs
1.16×
Reported
license unknown · no source
Geometric
unknown
41.3µs
1.16×
Reported
license unknown · no source
Geometric
unknown
41.5µs
1.17×
Reported
license unknown · no source
doubleAI
unknown
41.4µs
1.16×
Reported
license unknown · no source
Geometric
unknown
41.6µs
1.17×
Reported
license unknown · no source
Geometric
unknown
42.0µs
1.18×
Reported
license unknown · no source
Coral@HumanAgentSociety
unknown
41.7µs
1.17×
Reported
license unknown · no source
Coral@HumanAgentSociety
unknown
41.8µs
1.17×
Reported
license unknown · no source
OpenKernels-Ruibo
unknown
42.4µs
1.19×
Reported
license unknown · no source
doubleAI
unknown
42.2µs
1.19×
Reported
license unknown · no source
doubleAI
unknown
42.4µs
1.19×
Reported
license unknown · no source
OpenKernels-Ruibo
unknown
43.2µs
1.22×
Reported
license unknown · no source
Amir M. Mir | SF Tensor
unknown
43.8µs
1.23×
Reported
license unknown · no source
Argus-sufeng
unknown
43.4µs
1.22×
Reported
license unknown · no source
Amir M. Mir | SF Tensor
unknown
45.1µs
1.27×
Reported
license unknown · no source
Stellar Flamingo
unknown
44.9µs
1.26×
Reported
license unknown · no source
Stellar Flamingo
unknown
45.2µs
1.27×
Reported
license unknown · no source
Stellar Flamingo
unknown
46.0µs
1.29×
Reported
license unknown · no source
Cursor
unknown
47.9µs
1.35×
Reported
license unknown · no source
Amir M. Mir | SF Tensor
unknown
47.8µs
1.34×
Reported
license unknown · no source
Amir M. Mir | SF Tensor
unknown
48.0µs
1.35×
Reported
license unknown · no source
claude-opus-4-1_cuda_b43068claude-opus-4-1-20250805
cuda
—
—
Apache-2.0 · source
claude-opus-4-1 / triton8f159aclaude-opus-4-1-20250805
triton
289.0µs
8.12×
Reported
Apache-2.0 · source
gemini-2.5-pro / cudaadc04bgemini-2.5-pro
cuda
—
—
Apache-2.0 · source
gemini-2.5-pro / tritonq84sirgemini-2.5-pro
triton
128.2µs
3.60×
Reported
Apache-2.0 · source
gpt-5 / cudabd7484gpt-5-2025-08-07
cuda
1.11ms
31.25×
Reproduction-ready
Apache-2.0 · source
gpt-5 / triton1bd4a7gpt-5-2025-08-07
triton
1.06ms
29.92×
Reported
Apache-2.0 · source
gpt-o3 / cuda2ad247gpt-o3
cuda
—
—
Apache-2.0 · source
gpt-o3_triton_a47205gpt-o3
triton
304.7µs
8.57×
Reported
Apache-2.0 · source
torch_matmul_254647FlashInfer-Bench baselines
python
35.6µs
1.00×
Reported
Apache-2.0 · source

Semantics

Inputs and outputs
afp16 [m, k]
bfp16 [n, k]
cfp16 [m, n]
Axes and behavior
kconstant = 14336
mvariable
nconstant = 4096
determinismunspecified
constraintsNo mutation or aliasing
Identity
aliasgemm_n4096_k14336modelllama-3-1-8bsha256603ae61c1cf5…

Also indexed as 008-gemm-n4096-k14336 · reviewed as the same computation; every definition's cohorts are shown here, ranked separately.

Sources: FlashInfer-Bench (2025-10-20) · Apache-2.0 · NVIDIA SOL-ExecBench (2026-08-23)last observed 2026-08-23How records are decidedJSON