Skip to content
KernelIndex
Search⌘K
Sign in

FP16 matmul

120 eligible runs
gemm

Plain fp16 matrix multiply: a (m×k) @ b (k×n) into a pre-allocated fp16 c (m×n). All outer and inner dimensions are multiples of 16.

Fastest reported
106.9µs
submission 773912olezhka_007 · License unknown · python

Reported evidence · last observed 2026-04-18. Reported by source; not independently reproduced. License unknown.

No package · vendor the source →View source →Run detail →

Current records

HardwareBest knownImplementationRuns
NVIDIA B200106.9 µssubmission 77391253NVIDIA H100219.1 µssubmission 51247228NVIDIA A100627.7 µssubmission 78071827NVIDIA L42.08 mssubmission 78061112
Source-native comparison · GPU NVIDIA B200 · Workload suite of 8 cases · score · Protocol reference-kernels eval.py · 53 results · last observed 2026-05-12Record history →
#
Implementation
Latency
vs #1
Trust
Observed
1
submission 773912olezhka_007
106.9µs
1.00×
Reported · license unknown · source
2026-04-18

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

source mirroredlicense unknownno install recipeView source →Run detail →
2
submission 773910olezhka_007
107.2µs
1.00×
Reported · license unknown · source
2026-04-18

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

source mirroredlicense unknownno install recipeView source →Run detail →
3
submission 611052dannywillowliu-uchi
107.5µs
1.01×
Reported · license unknown · source
2026-03-22stale

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

source mirroredlicense unknownno install recipeView source →Run detail →
4
submission 773908olezhka_007
107.6µs
1.01×
Reported · license unknown · source
2026-04-18

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

source mirroredlicense unknownno install recipeView source →Run detail →
5
submission 773814olezhka_007
109.1µs
1.02×
Reported · license unknown · source
2026-04-18

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

source mirroredlicense unknownno install recipeView source →Run detail →
6
submission 773813olezhka_007
110.1µs
1.03×
Reported · license unknown · source
2026-04-18

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

source mirroredlicense unknownno install recipeView source →Run detail →
7
submission 773808olezhka_007
111.5µs
1.04×
Reported · license unknown · source
2026-04-18

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

source mirroredlicense unknownno install recipeView source →Run detail →
8
submission 780608ajay_a
113.2µs
1.06×
Reported · license unknown · source
2026-05-05

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

source mirroredlicense unknownno install recipeView source →Run detail →
9
submission 512434burtenshaw
113.5µs
1.06×
Reported · license unknown · source
2026-03-02stale

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

source mirroredlicense unknownno install recipeView source →Run detail →
10
submission 779796Shyamsaibethina
113.5µs
1.06×
Reported · license unknown · source
2026-04-23

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

source mirroredlicense unknownno install recipeView source →Run detail →
11
submission 779792Shyamsaibethina
113.6µs
1.06×
Reported · license unknown · source
2026-04-23

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

source mirroredlicense unknownno install recipeView source →Run detail →
12
submission 779725Shyamsaibethina
113.7µs
1.06×
Reported · license unknown · source
2026-04-23

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

source mirroredlicense unknownno install recipeView source →Run detail →
13
submission 773807olezhka_007
113.8µs
1.06×
Reported · license unknown · source
2026-04-18

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

source mirroredlicense unknownno install recipeView source →Run detail →
14
submission 780604ajay_a
114.2µs
1.07×
Reported · license unknown · source
2026-05-05

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

source mirroredlicense unknownno install recipeView source →Run detail →
15
submission 673457Elán Zainos Corona
114.4µs
1.07×
Reported · license unknown · source
2026-03-30stale

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

source mirroredlicense unknownno install recipeView source →Run detail →
16
submission 780585Samuel Nordmann
114.7µs
1.07×
Reported · license unknown · source
2026-05-04

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

source mirroredlicense unknownno install recipeView source →Run detail →
17
submission 779881ajay_a
114.8µs
1.07×
Reported · license unknown · source
2026-04-23

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

source mirroredlicense unknownno install recipeView source →Run detail →
18
submission 779877ajay_a
115.3µs
1.08×
Reported · license unknown · source
2026-04-23

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

source mirroredlicense unknownno install recipeView source →Run detail →
19
submission 779854ajay_a
115.3µs
1.08×
Reported · license unknown · source
2026-04-23

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

source mirroredlicense unknownno install recipeView source →Run detail →
20
submission 780584Samuel Nordmann
129.3µs
1.21×
Reported · license unknown · source
2026-05-04

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

source mirroredlicense unknownno install recipeView source →Run detail →
21
submission 773800olezhka_007
132.7µs
1.24×
Reported · license unknown · source
2026-04-18

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

source mirroredlicense unknownno install recipeView source →Run detail →
22
submission 773796olezhka_007
133.9µs
1.25×
Reported · license unknown · source
2026-04-18

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

source mirroredlicense unknownno install recipeView source →Run detail →
23
submission 673439Elán Zainos Corona
135.2µs
1.26×
Reported · license unknown · source
2026-03-30stale

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

source mirroredlicense unknownno install recipeView source →Run detail →
24
submission 513379JordanNanos
135.9µs
1.27×
Reported · license unknown · source
2026-03-06stale

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

source mirroredlicense unknownno install recipeView source →Run detail →
25
submission 513118JordanNanos
136.4µs
1.28×
Reported · license unknown · source
2026-03-06stale

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

source mirroredlicense unknownno install recipeView source →Run detail →
26
submission 542818Jaber Jaber
136.6µs
1.28×
Reported · license unknown · source
2026-03-13stale

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

source mirroredlicense unknownno install recipeView source →Run detail →
27
submission 539666Jaber Jaber
136.8µs
1.28×
Reported · license unknown · source
2026-03-12stale

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

source mirroredlicense unknownno install recipeView source →Run detail →
28
submission 545257rajesh0042
136.9µs
1.28×
Reported · license unknown · source
2026-03-13stale

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

source mirroredlicense unknownno install recipeView source →Run detail →
29
submission 512855JordanNanos
137.7µs
1.29×
Reported · license unknown · source
2026-03-05stale

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

source mirroredlicense unknownno install recipeView source →Run detail →
30
submission 773756olezhka_007
137.9µs
1.29×
Reported · license unknown · source
2026-04-18

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

source mirroredlicense unknownno install recipeView source →Run detail →
31
submission 673284Elán Zainos Corona
142.3µs
1.33×
Reported · license unknown · source
2026-03-30stale

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

source mirroredlicense unknownno install recipeView source →Run detail →
32
submission 512832mreso
142.9µs
1.34×
Reported · license unknown · source
2026-03-04stale

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

source mirroredlicense unknownno install recipeView source →Run detail →
33
submission 673272Elán Zainos Corona
143.3µs
1.34×
Reported · license unknown · source
2026-03-30stale

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

source mirroredlicense unknownno install recipeView source →Run detail →
34
submission 611026dannywillowliu-uchi
145.1µs
1.36×
Reported · license unknown · source
2026-03-22stale

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

source mirroredlicense unknownno install recipeView source →Run detail →
35
submission 773755olezhka_007
146.5µs
1.37×
Reported · license unknown · source
2026-04-18

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

source mirroredlicense unknownno install recipeView source →Run detail →
36
submission 545198rajesh0042
147.6µs
1.38×
Reported · license unknown · source
2026-03-13stale

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

source mirroredlicense unknownno install recipeView source →Run detail →
37
submission 512471iharryli
148.4µs
1.39×
Reported · license unknown · source
2026-03-03stale

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

source mirroredlicense unknownno install recipeView source →Run detail →
38
submission 670971Elán Zainos Corona
148.6µs
1.39×
Reported · license unknown · source
2026-03-30stale

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

source mirroredlicense unknownno install recipeView source →Run detail →
39
submission 670955Elán Zainos Corona
149.1µs
1.39×
Reported · license unknown · source
2026-03-30stale

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

source mirroredlicense unknownno install recipeView source →Run detail →
40
submission 512818mreso
149.4µs
1.40×
Reported · license unknown · no source
2026-03-04stale

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

no sourcelicense unknownno install recipeRun detail →
41
submission 512328burtenshaw
155.3µs
1.45×
Reported · license unknown · source
2026-03-02stale

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

source mirroredlicense unknownno install recipeView source →Run detail →
42
submission 93692Petr_Rocoss
155.4µs
1.45×
Reported · license unknown · source
2025-11-20stale

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

source mirroredlicense unknownno install recipeView source →Run detail →
43
submission 458326Nick Nuon
157.9µs
1.48×
Reported · license unknown · source
2026-02-04stale

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

source mirroredlicense unknownno install recipeView source →Run detail →
44
submission 773747olezhka_007
169.0µs
1.58×
Reported · license unknown · source
2026-04-18

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

source mirroredlicense unknownno install recipeView source →Run detail →
45
submission 779904D. Guo
172.7µs
1.62×
Reported · license unknown · source
2026-04-24

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

source mirroredlicense unknownno install recipeView source →Run detail →
46
submission 779851ajay_a
188.5µs
1.76×
Reported · license unknown · source
2026-04-23

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

source mirroredlicense unknownno install recipeView source →Run detail →
47
submission 762504rawat_arpit_04702
190.8µs
1.78×
Reported · license unknown · source
2026-04-11

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

source mirroredlicense unknownno install recipeView source →Run detail →
48
submission 93648Petr_Rocoss
203.6µs
1.90×
Reported · license unknown · no source
2025-11-20stale

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

no sourcelicense unknownno install recipeRun detail →
49
submission 93644Petr_Rocoss
211.6µs
1.98×
Reported · license unknown · no source
2025-11-20stale

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

no sourcelicense unknownno install recipeRun detail →
50
submission 779769D. Guo
241.3µs
2.26×
Reported · license unknown · no source
2026-04-23

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

no sourcelicense unknownno install recipeRun detail →
51
submission 512769mreso
5.02ms
46.92×
Reported · license unknown · no source
2026-03-04stale

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

no sourcelicense unknownno install recipeRun detail →
52
submission 782505marciok
5.31ms
49.63×
Reported · license unknown · source
2026-05-12

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

source mirroredlicense unknownno install recipeView source →Run detail →
53
submission 782502marciok
5.31ms
49.65×
Reported · license unknown · no source
2026-05-12

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

no sourcelicense unknownno install recipeRun detail →

Implementations

Implementation
Runtime
Best latency
Evidence
Availability
submission 780718brianyu
python
627.7µs
5.87×
Reported
license unknown · source
submission 545267rajesh0042
python
629.8µs
5.89×
Reported
license unknown · source
submission 780441Kernel-Zhang
python
635.2µs
5.94×
Reported
license unknown · source
submission 674172Elán Zainos Corona
python
641.5µs
6.00×
Reported
license unknown · source
submission 614108dannywillowliu-uchi
python
641.7µs
6.00×
Reported
license unknown · source
submission 554068idanbeck
python
676.5µs
6.33×
Reported
license unknown · source
submission 512299burtenshaw
python
683.0µs
6.39×
Reported
license unknown · source
submission 763575rawat_arpit_04702
python
697.7µs
6.52×
Reported
license unknown · source
Show all 120 implementations ›
submission 93689Petr_Rocoss
python
733.2µs
6.86×
Reported
license unknown · source
submission 779733D. Guo
python
734.2µs
6.87×
Reported
license unknown · source
submission 512470iharryli
python
744.4µs
6.96×
Reported
license unknown · source
submission 512816mreso
python
748.7µs
7.00×
Reported
license unknown · source
submission 782443shivbhatia
python
749.9µs
7.01×
Reported
license unknown · source
submission 777276ddwinterdd
python
825.7µs
7.72×
Reported
license unknown · source
submission 780614brianyu
python
634.2µs
5.93×
Reported
license unknown · source
submission 544874rajesh0042
python
746.5µs
6.98×
Reported
license unknown · source
submission 544928rajesh0042
python
655.3µs
6.13×
Reported
license unknown · source
submission 544931rajesh0042
python
652.3µs
6.10×
Reported
license unknown · source
submission 544949rajesh0042
python
650.0µs
6.08×
Reported
license unknown · source
submission 554041idanbeck
python
691.1µs
6.46×
Reported
license unknown · source
submission 760226rawat_arpit_04702
python
2.30s
21541.87×
Reported
license unknown · source
submission 762432rawat_arpit_04702
python
706.7µs
6.61×
Reported
license unknown · source
submission 762840rawat_arpit_04702
python
701.8µs
6.56×
Reported
license unknown · source
submission 763263rawat_arpit_04702
python
700.8µs
6.55×
Reported
license unknown · source
submission 93643Petr_Rocoss
python
929.6µs
8.69×
Reported
license unknown · source
submission 93647Petr_Rocoss
python
777.4µs
7.27×
Reported
license unknown · source
submission 772639D. Guo
python
741.7µs
6.94×
Reported
license unknown · source
submission 773912olezhka_007
python
106.9µs
1.00×
Reported
license unknown · source
submission 611052dannywillowliu-uchi
python
107.5µs
1.01×
Reported
license unknown · source
submission 780608ajay_a
python
113.2µs
1.06×
Reported
license unknown · source
submission 512434burtenshaw
python
113.5µs
1.06×
Reported
license unknown · source
submission 779796Shyamsaibethina
python
113.5µs
1.06×
Reported
license unknown · source
submission 673457Elán Zainos Corona
python
114.4µs
1.07×
Reported
license unknown · source
submission 780585Samuel Nordmann
python
114.7µs
1.07×
Reported
license unknown · source
submission 513379JordanNanos
python
135.9µs
1.27×
Reported
license unknown · source
submission 542818Jaber Jaber
python
136.6µs
1.28×
Reported
license unknown · source
submission 545257rajesh0042
python
136.9µs
1.28×
Reported
license unknown · source
submission 512832mreso
python
142.9µs
1.34×
Reported
license unknown · source
submission 512471iharryli
python
148.4µs
1.39×
Reported
license unknown · source
submission 93692Petr_Rocoss
python
155.4µs
1.45×
Reported
license unknown · source
submission 458326Nick Nuon
python
157.9µs
1.48×
Reported
license unknown · source
submission 779904D. Guo
python
172.7µs
1.62×
Reported
license unknown · source
submission 762504rawat_arpit_04702
python
190.8µs
1.78×
Reported
license unknown · source
submission 782505marciok
python
5.31ms
49.63×
Reported
license unknown · source
submission 773747olezhka_007
python
169.0µs
1.58×
Reported
license unknown · source
submission 773755olezhka_007
python
146.5µs
1.37×
Reported
license unknown · source
submission 773756olezhka_007
python
137.9µs
1.29×
Reported
license unknown · source
submission 773796olezhka_007
python
133.9µs
1.25×
Reported
license unknown · source
submission 773800olezhka_007
python
132.7µs
1.24×
Reported
license unknown · source
submission 773807olezhka_007
python
113.8µs
1.06×
Reported
license unknown · source
submission 773808olezhka_007
python
111.5µs
1.04×
Reported
license unknown · source
submission 773813olezhka_007
python
110.1µs
1.03×
Reported
license unknown · source
submission 773814olezhka_007
python
109.1µs
1.02×
Reported
license unknown · source
submission 773908olezhka_007
python
107.6µs
1.01×
Reported
license unknown · source
submission 773910olezhka_007
python
107.2µs
1.00×
Reported
license unknown · source
submission 611026dannywillowliu-uchi
python
145.1µs
1.36×
Reported
license unknown · source
submission 779851ajay_a
python
188.5µs
1.76×
Reported
license unknown · source
submission 779854ajay_a
python
115.3µs
1.08×
Reported
license unknown · source
submission 779877ajay_a
python
115.3µs
1.08×
Reported
license unknown · source
submission 779881ajay_a
python
114.8µs
1.07×
Reported
license unknown · source
submission 780604ajay_a
python
114.2µs
1.07×
Reported
license unknown · source
submission 512328burtenshaw
python
155.3µs
1.45×
Reported
license unknown · source
submission 779725Shyamsaibethina
python
113.7µs
1.06×
Reported
license unknown · source
submission 779792Shyamsaibethina
python
113.6µs
1.06×
Reported
license unknown · source
submission 670955Elán Zainos Corona
python
149.1µs
1.39×
Reported
license unknown · source
submission 670971Elán Zainos Corona
python
148.6µs
1.39×
Reported
license unknown · source
submission 673272Elán Zainos Corona
python
143.3µs
1.34×
Reported
license unknown · source
submission 673284Elán Zainos Corona
python
142.3µs
1.33×
Reported
license unknown · source
submission 673439Elán Zainos Corona
python
135.2µs
1.26×
Reported
license unknown · source
submission 780584Samuel Nordmann
python
129.3µs
1.21×
Reported
license unknown · source
submission 512855JordanNanos
python
137.7µs
1.29×
Reported
license unknown · source
submission 513118JordanNanos
python
136.4µs
1.28×
Reported
license unknown · source
submission 539666Jaber Jaber
python
136.8µs
1.28×
Reported
license unknown · source
submission 545198rajesh0042
python
147.6µs
1.38×
Reported
license unknown · source
submission 512472iharryli
python
219.1µs
2.05×
Reported
license unknown · source
submission 614113dannywillowliu-uchi
python
219.1µs
2.05×
Reported
license unknown · source
submission 544953rajesh0042
python
219.4µs
2.05×
Reported
license unknown · source
submission 512817mreso
python
220.4µs
2.06×
Reported
license unknown · source
submission 512443burtenshaw
python
220.5µs
2.06×
Reported
license unknown · source
submission 757537Zeyu Li
python
220.6µs
2.06×
Reported
license unknown · source
submission 757204ethan
python
223.0µs
2.09×
Reported
license unknown · source
submission 779785ajay_a
python
233.4µs
2.18×
Reported
license unknown · source
submission 577884KernelAgent
python
234.8µs
2.20×
Reported
license unknown · source
submission 776985ddwinterdd
python
246.5µs
2.31×
Reported
license unknown · source
submission 93693Petr_Rocoss
python
249.5µs
2.33×
Reported
license unknown · source
submission 780703D. Guo
python
252.2µs
2.36×
Reported
license unknown · source
submission 583248Cookie 🍪
python
287.8µs
2.69×
Reported
license unknown · source
submission 762503rawat_arpit_04702
python
301.0µs
2.82×
Reported
license unknown · source
submission 756979ethan0027
python
317.5µs
2.97×
Reported
license unknown · source
submission 512775mreso
python
6.11ms
57.15×
Reported
license unknown · source
submission 512327burtenshaw
python
248.6µs
2.32×
Reported
license unknown · source
submission 512435burtenshaw
python
226.3µs
2.12×
Reported
license unknown · source
submission 490610KernelAgent
python
1.43ms
13.36×
Reported
license unknown · source
submission 490611KernelAgent
python
351.8µs
3.29×
Reported
license unknown · source
submission 511441KernelAgent
python
340.8µs
3.19×
Reported
license unknown · source
submission 512161KernelAgent
python
302.5µs
2.83×
Reported
license unknown · source
submission 569477KernelAgent
python
272.8µs
2.55×
Reported
license unknown · source
submission 780611brianyu
python
2.08ms
19.42×
Reported
license unknown · source
submission 512469iharryli
python
2.20ms
20.58×
Reported
license unknown · source
submission 512815mreso
python
2.23ms
20.82×
Reported
license unknown · source
submission 544956rajesh0042
python
2.26ms
21.09×
Reported
license unknown · source
submission 512444burtenshaw
python
2.30ms
21.53×
Reported
license unknown · source
submission 93704Petr_Rocoss
python
2.34ms
21.88×
Reported
license unknown · source
submission 512329burtenshaw
python
2.58ms
24.10×
Reported
license unknown · source
submission 512436burtenshaw
python
2.35ms
21.93×
Reported
license unknown · source
submission 93645Petr_Rocoss
python
2.67ms
24.94×
Reported
license unknown · source
submission 93656Petr_Rocoss
python
2.65ms
24.76×
Reported
license unknown · source
submission 93671Petr_Rocoss
python
2.47ms
23.14×
Reported
license unknown · source
submission 93694Petr_Rocoss
python
2.35ms
21.95×
Reported
license unknown · source
submission 580634Cookie 🍪
python
291.1µs
2.72×
Reported
license unknown · no source
submission 487903Cookie 🍪
python
324.5µs
3.03×
Reported
license unknown · no source
submission 93642Petr_Rocoss
python
335.6µs
3.14×
Reported
license unknown · no source
submission 93638Petr_Rocoss
python
345.3µs
3.23×
Reported
license unknown · no source
submission 487818Cookie 🍪
python
9.67ms
90.42×
Reported
license unknown · no source
submission 512818mreso
python
149.4µs
1.40×
Reported
license unknown · no source
submission 93648Petr_Rocoss
python
203.6µs
1.90×
Reported
license unknown · no source
submission 93644Petr_Rocoss
python
211.6µs
1.98×
Reported
license unknown · no source
submission 779769D. Guo
python
241.3µs
2.26×
Reported
license unknown · no source
submission 512769mreso
python
5.02ms
46.92×
Reported
license unknown · no source
submission 782502marciok
python
5.31ms
49.65×
Reported
license unknown · no source

Semantics

Inputs and outputs
afp16 [m, k]
bfp16 [k, n]
cfp16 [m, n]
Axes and behavior
kvariable
mvariable
nvariable
seedvariable
determinismunspecified
constraintsNo mutation or aliasing
Identity
aliasfp16 matmulaliasmatmul_v2sha256ea05bab53252…
Sources: GPU MODE KernelBot (2026-05-12) · June 9 Researcher Reciprocity License v1.0last observed 2026-05-12How records are decidedJSON