Skip to content
KernelIndex
Search⌘K

Rotary position embedding

53 eligible runs
rope

RoPE (Rotary Position Embedding) computation with llama3 scaling that computes cos/sin embeddings from inverse frequencies and position IDs. Uses frequency-based scaling with high_freq_factor=4.0, low_freq_factor=1.0, factor=8.0 for extended context support up to 131072 positions. Returns concatenated cos and sin embeddings.

Fastest reported
1.86µsmean · 77% of SOL
RecursiveLicense unknown · unknown

Reported evidence · last observed 2026-06-07. Reported by source; not independently reproduced. No public source. License unknown.

No public sourceRun detail →

Current records

Workloadsuite of 16 cases · mean latencyhead_dim = 128 · half_head_dim = 6416 cases

Not measured on H100 for this workload. Challenges →

Source-native comparison · GPU NVIDIA B200 · Workload suite of 16 cases · mean latency · Protocol SOL-ExecBench evaluation stack v1.0 · mean · 29 results · last observed 2026-08-13Record history →
#
Implementation
Latency
vs #1
Trust
Observed
1
1.86µs77% SOL
1.00×
Reported · license unknown · no source
2026-06-07

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

no sourcelicense unknownno install recipeRun detail →
2
1.89µs77% SOL
1.01×
Reported · license unknown · no source
2026-06-26

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

no sourcelicense unknownno install recipeRun detail →
3
1.99µs77% SOL
1.07×
Reported · license unknown · no source
2026-07-01

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

no sourcelicense unknownno install recipeRun detail →
4
2.07µs76% SOL
1.11×
Reported · license unknown · no source
2026-06-28

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

no sourcelicense unknownno install recipeRun detail →
5
2.37µs71% SOL
1.27×
Reported · license unknown · no source
2026-07-08

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

no sourcelicense unknownno install recipeRun detail →
6
2.51µs71% SOL
1.35×
Reported · license unknown · no source
2026-07-02

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

no sourcelicense unknownno install recipeRun detail →
7
2.85µs68% SOL
1.53×
Reported · license unknown · no source
2026-07-05

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

no sourcelicense unknownno install recipeRun detail →
7=
2.85µs68% SOL
1.53×
Reported · license unknown · no source
2026-06-30

Measured exactly what you asked; tied ranks share a number. Reported by source; not independently reproduced.

no sourcelicense unknownno install recipeRun detail →
8
2.88µs67% SOL
1.54×
Reported · license unknown · no source
2026-05-31

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

no sourcelicense unknownno install recipeRun detail →
9
2.88µs67% SOL
1.55×
Reported · license unknown · no source
2026-05-31

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

no sourcelicense unknownno install recipeRun detail →
10
2.89µs67% SOL
1.55×
Reported · license unknown · no source
2026-04-21

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

no sourcelicense unknownno install recipeRun detail →
11
2.89µs67% SOL
1.55×
Reported · license unknown · no source
2026-05-31

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

no sourcelicense unknownno install recipeRun detail →
12
2.89µs67% SOL
1.55×
Reported · license unknown · no source
2026-05-31

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

no sourcelicense unknownno install recipeRun detail →
13
2.92µs67% SOL
1.57×
Reported · license unknown · no source
2026-05-19

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

no sourcelicense unknownno install recipeRun detail →
14
2.97µs67% SOL
1.59×
Reported · license unknown · no source
2026-05-19

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

no sourcelicense unknownno install recipeRun detail →
15
3.40µs63% SOL
1.82×
Reported · license unknown · no source
2026-05-28

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

no sourcelicense unknownno install recipeRun detail →
16
3.48µs63% SOL
1.86×
Reported · license unknown · no source
2026-04-27

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

no sourcelicense unknownno install recipeRun detail →
17
3.67µs62% SOL
1.97×
Reported · license unknown · no source
2026-04-15

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

no sourcelicense unknownno install recipeRun detail →
18
3.73µs61% SOL
2.00×
Reported · license unknown · no source
2026-04-14

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

no sourcelicense unknownno install recipeRun detail →
19
4.19µs58% SOL
2.25×
Reported · license unknown · no source
2026-07-07

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

no sourcelicense unknownno install recipeRun detail →
20
4.70µs55% SOL
2.52×
Reported · license unknown · no source
2026-05-06

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

no sourcelicense unknownno install recipeRun detail →
21
4.83µs54% SOL
2.59×
Reported · license unknown · no source
2026-04-15

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

no sourcelicense unknownno install recipeRun detail →
22
4.84µs54% SOL
2.60×
Reported · license unknown · no source
2026-06-25

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

no sourcelicense unknownno install recipeRun detail →
23
4.87µs54% SOL
2.61×
Reported · license unknown · no source
2026-06-25

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

no sourcelicense unknownno install recipeRun detail →
24
5.01µs53% SOL
2.69×
Reported · license unknown · no source
2026-06-14

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

no sourcelicense unknownno install recipeRun detail →
25
5.01µs53% SOL
2.69×
Reported · license unknown · no source
2026-06-09

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

no sourcelicense unknownno install recipeRun detail →
26
5.67µs49% SOL
3.04×
Reported · license unknown · no source
2026-04-20

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

no sourcelicense unknownno install recipeRun detail →
27
24.4µs17% SOL
13.07×
Reported · license unknown · no source
2026-04-15

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

no sourcelicense unknownno install recipeRun detail →
28
41.6µs11% SOL
22.29×
Reported · license unknown · no source
2026-05-28

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

no sourcelicense unknownno install recipeRun detail →

Implementations

Implementation
Runtime
Best latency
Evidence
Availability
unknown
1.95µs
1.16×
Reported
license unknown · no source
unknown
1.88µs
1.11×
Reported
license unknown · no source
unknown
1.86µs
1.10×
Reported
license unknown · no source
unknown
2.07µs
1.22×
Reported
license unknown · no source
unknown
2.35µs
1.39×
Reported
license unknown · no source
unknown
2.51µs
1.49×
Reported
license unknown · no source
unknown
2.58µs
1.53×
Reported
license unknown · no source
unknown
2.79µs
1.66×
Reported
license unknown · no source
Show all 53 implementations ›
unknown
2.85µs
1.69×
Reported
license unknown · no source
unknown
2.85µs
1.69×
Reported
license unknown · no source
unknown
2.88µs
1.71×
Reported
license unknown · no source
unknown
2.89µs
1.71×
Reported
license unknown · no source
unknown
3.04µs
1.80×
Reported
license unknown · no source
unknown
3.40µs
2.01×
Reported
license unknown · no source
unknown
3.73µs
2.21×
Reported
license unknown · no source
unknown
4.19µs
2.48×
Reported
license unknown · no source
unknown
4.70µs
2.79×
Reported
license unknown · no source
unknown
4.83µs
2.86×
Reported
license unknown · no source
unknown
4.84µs
2.87×
Reported
license unknown · no source
unknown
5.01µs
2.97×
Reported
license unknown · no source
unknown
5.01µs
2.97×
Reported
license unknown · no source
unknown
24.4µs
14.44×
Reported
license unknown · no source
unknown
1.99µs
1.18×
Reported
license unknown · no source
unknown
1.89µs
1.12×
Reported
license unknown · no source
unknown
1.87µs
1.11×
Reported
license unknown · no source
unknown
2.09µs
1.24×
Reported
license unknown · no source
unknown
2.37µs
1.40×
Reported
license unknown · no source
unknown
2.40µs
1.42×
Reported
license unknown · no source
unknown
2.40µs
1.42×
Reported
license unknown · no source
unknown
2.40µs
1.42×
Reported
license unknown · no source
unknown
2.56µs
1.52×
Reported
license unknown · no source
unknown
2.79µs
1.65×
Reported
license unknown · no source
unknown
2.80µs
1.66×
Reported
license unknown · no source
unknown
2.89µs
1.71×
Reported
license unknown · no source
unknown
2.88µs
1.71×
Reported
license unknown · no source
unknown
2.92µs
1.73×
Reported
license unknown · no source
unknown
2.89µs
1.71×
Reported
license unknown · no source
unknown
2.89µs
1.71×
Reported
license unknown · no source
unknown
2.97µs
1.76×
Reported
license unknown · no source
unknown
3.03µs
1.80×
Reported
license unknown · no source
unknown
3.21µs
1.90×
Reported
license unknown · no source
unknown
3.45µs
2.04×
Reported
license unknown · no source
unknown
3.48µs
2.06×
Reported
license unknown · no source
unknown
3.67µs
2.18×
Reported
license unknown · no source
unknown
3.73µs
2.21×
Reported
license unknown · no source
unknown
4.51µs
2.67×
Reported
license unknown · no source
unknown
4.87µs
2.88×
Reported
license unknown · no source
unknown
5.17µs
3.06×
Reported
license unknown · no source
unknown
5.67µs
3.36×
Reported
license unknown · no source
unknown
41.6µs
24.63×
Reported
license unknown · no source
unknown
1.69µs
1.00×
Reported
license unknown · no source
unknown
2.37µs
1.40×
Reported
license unknown · no source
unknown
2.30µs
1.36×
Reported
license unknown · no source

Semantics

Inputs and outputs
position_idsint64 [batch_size, seq_len]
inv_freqfp32 [half_head_dim]
attention_scalingfp32 scalar
cos_sinbf16 [batch_size, seq_len, head_dim, 2]
Axes and behavior
seq_lenvariable
head_dimconstant = 128
batch_sizevariable
half_head_dimhalf_head_dim = head_dim // 2
determinismunspecified
constraintsNo mutation or aliasing
Identity
alias011_rotary_position_embeddingmodelhermes-4-70bsha2561fb1795aa22b…
Sources: NVIDIA SOL-ExecBench (2026-08-13)last observed 2026-08-13How records are decidedJSON