Skip to content
KernelIndex
Search⌘K
Sign in

Fused add RMSNorm h7168

112 eligible runs
rmsnorm

Fused Add + RMSNorm with hidden_size=7168 for DeepSeek-V3/R1. Epsilon is fixed at 1e-6.

Fastest reported
7.16µsmean · 71% of SOL
Electric CheetahLicense unknown · unknown

Reported evidence · last observed 2026-08-15. Reported by source; not independently reproduced. No public source. License unknown.

No public sourceRun detail →

Current records

Workloadsuite of 8 cases · mean latency16 cases
batch_size
hidden_size
dtype
1—bf1617168bf167—bf1677168bf1618—bf16187168bf1632—bf16327168bf1664—bf16647168bf16539—bf165397168bf1611949—bf16119497168bf1614521—bf16145217168bf16
scrolls · 16 cases total
Aggregatessuite of 8 cases · mean latency
HardwareBest knownImplementationRuns
NVIDIA B200 · env 17.16 µsElectric Cheetah40NVIDIA B200 · env 27.78 µsdoubleAI16

Not measured on H100 for this workload. Challenges →

Source-native comparison · GPU NVIDIA B200 · Workload suite of 8 cases · mean latency · Protocol SOL-ExecBench evaluation stack v1.1 · mean · 40 results · last observed 2026-08-15Record history →
#
Implementation
Latency
vs #1
Trust
Observed
1
Electric Cheetah
7.16µs71% SOL
1.00×
Reported · license unknown · no source
2026-08-15

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

no sourcelicense unknownno install recipeRun detail →
2
Rapid Tiger
7.17µs71% SOL
1.00×
Reported · license unknown · no source
2026-08-15

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

no sourcelicense unknownno install recipeRun detail →
3
Amir M. Mir | SF Tensor
7.18µs70% SOL
1.00×
Reported · license unknown · no source
2026-08-08

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

no sourcelicense unknownno install recipeRun detail →
4
Geometric
7.21µs71% SOL
1.01×
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 →
5
ac4k
7.21µs70% SOL
1.01×
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
Gleaming Coyote
7.23µs71% SOL
1.01×
Reported · license unknown · no source
2026-08-15

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

no sourcelicense unknownno install recipeRun detail →
7
Qwen
7.25µs71% SOL
1.01×
Reported · license unknown · no source
2026-07-23

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

no sourcelicense unknownno install recipeRun detail →
8
Qwen
7.35µs70% SOL
1.03×
Reported · license unknown · no source
2026-07-22

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

no sourcelicense unknownno install recipeRun detail →
9
Geometric
7.47µs70% 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 →
10
Geometric
7.50µs70% 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 →
11
Geometric
7.51µs70% 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 →
12
Geometric
7.56µs70% 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 →
13
ac4k
7.57µs69% SOL
1.06×
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 →
14
Geometric
7.59µs69% SOL
1.06×
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 →
15
ac4k
7.63µs68% 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 →
16
Geometric
7.68µs68% 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 →
17
Geometric
7.68µs68% SOL
1.07×
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 →
18
Geometric
7.70µs68% SOL
1.08×
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 →
19
Amir M. Mir | SF Tensor
7.71µs69% SOL
1.08×
Reported · license unknown · no source
2026-08-05

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

no sourcelicense unknownno install recipeRun detail →
20
Geometric
7.77µs68% SOL
1.08×
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 →
21
Geometric
7.88µs67% SOL
1.10×
Reported · license unknown · no source
2026-07-17

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

no sourcelicense unknownno install recipeRun detail →
22
doubleAI
7.90µs68% SOL
1.10×
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 →
23
Geometric
7.92µs67% SOL
1.11×
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 →
24
Stellar Flamingo
8.03µs67% SOL
1.12×
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 →
25
Coral@HumanAgentSociety
8.16µs66% SOL
1.14×
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 →
26
Amir M. Mir | SF Tensor
8.18µs67% SOL
1.14×
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 →
27
Stellar Flamingo
8.18µs66% SOL
1.14×
Reported · license unknown · no source
2026-07-17

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

no sourcelicense unknownno install recipeRun detail →
28
Stellar Flamingo
8.21µs66% SOL
1.15×
Reported · license unknown · no source
2026-07-17

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

no sourcelicense unknownno install recipeRun detail →
29
Amir M. Mir | SF Tensor
8.26µs66% SOL
1.15×
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 →
30
Geometric
8.31µs65% SOL
1.16×
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 →
31
Stellar Flamingo
8.48µs65% SOL
1.18×
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 →
32
Amir M. Mir | SF Tensor
8.62µs65% SOL
1.20×
Reported · license unknown · no source
2026-07-25

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

no sourcelicense unknownno install recipeRun detail →
33
Recursive
8.64µs64% SOL
1.21×
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 →
34
Amir M. Mir | SF Tensor
8.73µs64% SOL
1.22×
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 →
35
Hyra
8.80µs64% SOL
1.23×
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 →
36
XG
9.18µs63% SOL
1.28×
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
Amir M. Mir | SF Tensor
9.42µs61% SOL
1.32×
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 →
38
Theoric.com
10.2µs59% SOL
1.42×
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 →
39
Cursor
13.0µs52% SOL
1.82×
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 →
40
Theoric.com
15.2µs44% SOL
2.12×
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 →

Implementations

Implementation
Runtime
Best latency
Evidence
Availability
Geometric
unknown
7.21µs
1.01×
Reported
license unknown · no source
Amir M. Mir | SF Tensor
unknown
7.18µs
1.00×
Reported
license unknown · no source
ac4k
unknown
7.21µs
1.01×
Reported
license unknown · no source
Qwen
unknown
7.25µs
1.01×
Reported
license unknown · no source
Electric Cheetah
unknown
7.16µs
1.00×
Reported
license unknown · no source
Rapid Tiger
unknown
7.17µs
1.00×
Reported
license unknown · no source
Gleaming Coyote
unknown
7.23µs
1.01×
Reported
license unknown · no source
doubleAI
unknown
7.78µs
1.09×
Reported
license unknown · no source
Show all 65 implementations ›
Stellar Flamingo
unknown
8.03µs
1.12×
Reported
license unknown · no source
Coral@HumanAgentSociety
unknown
8.12µs
1.13×
Reported
license unknown · no source
Recursive
unknown
8.59µs
1.20×
Reported
license unknown · no source
Hyra
unknown
8.80µs
1.23×
Reported
license unknown · no source
XG
unknown
9.16µs
1.28×
Reported
license unknown · no source
KerSor
unknown
9.13µs
1.28×
Reported
license unknown · no source
IcySky
unknown
10.1µs
1.41×
Reported
license unknown · no source
Theoric.com
unknown
10.2µs
1.42×
Reported
license unknown · no source
Cursor
unknown
13.0µs
1.82×
Reported
license unknown · no source
Rapid Impala
unknown
13.1µs
1.84×
Reported
license unknown · no source
Qwen
unknown
7.35µs
1.03×
Reported
license unknown · no source
Geometric
unknown
7.47µs
1.04×
Reported
license unknown · no source
Geometric
unknown
7.51µs
1.05×
Reported
license unknown · no source
Geometric
unknown
7.50µs
1.05×
Reported
license unknown · no source
Geometric
unknown
7.56µs
1.06×
Reported
license unknown · no source
ac4k
unknown
7.57µs
1.06×
Reported
license unknown · no source
Geometric
unknown
7.59µs
1.06×
Reported
license unknown · no source
Amir M. Mir | SF Tensor
unknown
7.71µs
1.08×
Reported
license unknown · no source
ac4k
unknown
7.63µs
1.06×
Reported
license unknown · no source
Geometric
unknown
7.68µs
1.07×
Reported
license unknown · no source
Geometric
unknown
7.68µs
1.07×
Reported
license unknown · no source
Geometric
unknown
7.70µs
1.08×
Reported
license unknown · no source
Geometric
unknown
7.77µs
1.08×
Reported
license unknown · no source
doubleAI
unknown
7.90µs
1.10×
Reported
license unknown · no source
doubleAI
unknown
7.83µs
1.09×
Reported
license unknown · no source
Geometric
unknown
7.88µs
1.10×
Reported
license unknown · no source
Geometric
unknown
7.92µs
1.11×
Reported
license unknown · no source
doubleAI
unknown
8.07µs
1.13×
Reported
license unknown · no source
Amir M. Mir | SF Tensor
unknown
8.18µs
1.14×
Reported
license unknown · no source
Stellar Flamingo
unknown
8.21µs
1.15×
Reported
license unknown · no source
Coral@HumanAgentSociety
unknown
8.16µs
1.14×
Reported
license unknown · no source
Stellar Flamingo
unknown
8.18µs
1.14×
Reported
license unknown · no source
Coral@HumanAgentSociety
unknown
8.30µs
1.16×
Reported
license unknown · no source
Amir M. Mir | SF Tensor
unknown
8.26µs
1.15×
Reported
license unknown · no source
Geometric
unknown
8.31µs
1.16×
Reported
license unknown · no source
Stellar Flamingo
unknown
8.46µs
1.18×
Reported
license unknown · no source
Stellar Flamingo
unknown
8.47µs
1.18×
Reported
license unknown · no source
Amir M. Mir | SF Tensor
unknown
8.62µs
1.20×
Reported
license unknown · no source
Stellar Flamingo
unknown
8.48µs
1.18×
Reported
license unknown · no source
Stellar Flamingo
unknown
8.38µs
1.17×
Reported
license unknown · no source
Recursive
unknown
8.64µs
1.21×
Reported
license unknown · no source
Amir M. Mir | SF Tensor
unknown
8.73µs
1.22×
Reported
license unknown · no source
XG
unknown
9.18µs
1.28×
Reported
license unknown · no source
Stellar Flamingo
unknown
9.19µs
1.28×
Reported
license unknown · no source
Amir M. Mir | SF Tensor
unknown
9.42µs
1.32×
Reported
license unknown · no source
Amir M. Mir | SF Tensor
unknown
9.73µs
1.36×
Reported
license unknown · no source
Cursor
unknown
13.1µs
1.83×
Reported
license unknown · no source
Theoric.com
unknown
15.2µs
2.12×
Reported
license unknown · no source
claude-opus-4-1 / cudab1507eclaude-opus-4-1-20250805
cuda
—
—
Apache-2.0 · source
claude-opus-4-1 / tritonfe43bfclaude-opus-4-1-20250805
triton
32.5µs
4.54×
Reported
Apache-2.0 · source
gemini-2.5-pro_cuda_428669gemini-2.5-pro
cuda
10.3µs
1.43×
Reproduction-ready
Apache-2.0 · source
gemini-2.5-pro / triton05pwmxgemini-2.5-pro
triton
8.19µs
1.14×
Reported
Apache-2.0 · source
gpt-5 / cuda0d807bgpt-5-2025-08-07
cuda
—
—
Apache-2.0 · source
gpt-5 / triton714ae0gpt-5-2025-08-07
triton
12.9µs
1.81×
Reported
Apache-2.0 · source
gpt-o3 / cuda620cb5gpt-o3
cuda
16.0µs
2.23×
Reproduction-ready
Apache-2.0 · source
gpt-o3 / tritone5fd98gpt-o3
triton
32.7µs
4.57×
Reported
Apache-2.0 · source
flashinfer / wrapper5bddf1FlashInfer-Bench baselines
python
8.19µs
1.14×
Reported
Apache-2.0 · source

Semantics

Inputs and outputs
hidden_statesbf16 [batch_size, hidden_size]
residualbf16 [batch_size, hidden_size]
weightbf16 [hidden_size]
outputbf16 [batch_size, hidden_size]
Axes and behavior
batch_sizevariable
hidden_sizeconstant = 7168
determinismunspecified
constraintsNo mutation or aliasing
Identity
aliasfused_add_rmsnorm_h7168modeldeepseek-v3modeldeepseek-r1sha256058ae07ca394…

Also indexed as 003-fused-add-rmsnorm-h7168 · reviewed as the same computation; every definition's cohorts are shown here, ranked separately.

Sources: NVIDIA SOL-ExecBench (2026-08-15) · FlashInfer-Bench (2025-10-21) · Apache-2.0last observed 2026-08-15How records are decidedJSON