Skip to content
KernelIndex
Search⌘K

Flux multi axis RoPE frequency computation

28 eligible runs
rope

Multi-axis rotary position embedding frequency computation that generates cos/sin frequency tensors for 3D position encoding (time, height, width). Computes 1D RoPE for each axis with dimensions [16, 56, 56] and concatenates them. Involves frequency band calculation, outer products, sin/cos evaluation, repeat_interleave operations, and concatenation across axes.

Fastest reported
2.19µsmean · 73% of SOL
DatabricksLicense unknown · unknown

Reported evidence · last observed 2026-06-26. Reported by source; not independently reproduced. No public source. License unknown.

No public sourceRun detail →

Current records

Workloadsuite of 16 cases · mean latencyn_axes = 3 · axis0_dim = 16 · axis1_dim = 56 · axis2_dim = 56 · total_dim = 12816 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 · 16 results · last observed 2026-08-12Record history →
#
Implementation
Latency
vs #1
Trust
Observed
1
2.19µs73% SOL
1.00×
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 →
2
2.28µs72% SOL
1.04×
Reported · license unknown · no source
2026-06-29

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

no sourcelicense unknownno install recipeRun detail →
3
2.33µs72% SOL
1.07×
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 →
4
2.47µs70% SOL
1.13×
Reported · license unknown · no source
2026-06-29

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

no sourcelicense unknownno install recipeRun detail →
5
2.50µs70% SOL
1.14×
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 →
6
2.51µs70% SOL
1.15×
Reported · license unknown · no source
2026-05-20

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

no sourcelicense unknownno install recipeRun detail →
7
2.53µs69% SOL
1.16×
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 →
8
2.64µs68% SOL
1.21×
Reported · license unknown · no source
2026-04-30

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

no sourcelicense unknownno install recipeRun detail →
9
2.89µs66% SOL
1.32×
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 →
10
3.06µs65% SOL
1.39×
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 →
11
3.30µs63% SOL
1.51×
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 →
12
3.39µs62% SOL
1.55×
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 →
13
3.47µs61% SOL
1.58×
Reported · license unknown · no source
2026-06-01

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

no sourcelicense unknownno install recipeRun detail →
14
4.02µs57% SOL
1.83×
Reported · license unknown · no source
2026-07-03

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

no sourcelicense unknownno install recipeRun detail →
15
4.37µs55% SOL
1.99×
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 →
16
242.4µs2% SOL
110.63×
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
2.28µs
1.04×
Reported
license unknown · no source
unknown
2.32µs
1.06×
Reported
license unknown · no source
unknown
2.33µs
1.07×
Reported
license unknown · no source
unknown
2.39µs
1.09×
Reported
license unknown · no source
unknown
2.47µs
1.13×
Reported
license unknown · no source
unknown
2.50µs
1.14×
Reported
license unknown · no source
unknown
2.53µs
1.16×
Reported
license unknown · no source
unknown
2.70µs
1.23×
Reported
license unknown · no source
Show all 28 implementations ›
unknown
2.89µs
1.32×
Reported
license unknown · no source
unknown
3.06µs
1.39×
Reported
license unknown · no source
unknown
3.30µs
1.51×
Reported
license unknown · no source
unknown
3.47µs
1.58×
Reported
license unknown · no source
unknown
242.4µs
110.63×
Reported
license unknown · no source
unknown
2.30µs
1.05×
Reported
license unknown · no source
unknown
2.34µs
1.07×
Reported
license unknown · no source
unknown
2.38µs
1.09×
Reported
license unknown · no source
unknown
2.48µs
1.13×
Reported
license unknown · no source
unknown
2.51µs
1.15×
Reported
license unknown · no source
unknown
2.52µs
1.15×
Reported
license unknown · no source
unknown
2.63µs
1.20×
Reported
license unknown · no source
unknown
2.64µs
1.21×
Reported
license unknown · no source
unknown
2.92µs
1.33×
Reported
license unknown · no source
unknown
3.39µs
1.55×
Reported
license unknown · no source
unknown
4.02µs
1.83×
Reported
license unknown · no source
unknown
4.37µs
1.99×
Reported
license unknown · no source
unknown
2.19µs
1.00×
Reported
license unknown · no source
unknown
2.20µs
1.00×
Reported
license unknown · no source
unknown
2.22µs
1.01×
Reported
license unknown · no source

Semantics

Inputs and outputs
idsfp32 [seq_len, n_axes]
thetafp32 scalar
freqs_cosfp32 [seq_len, total_dim]
freqs_sinfp32 [seq_len, total_dim]
Axes and behavior
n_axesconstant = 3
seq_lenvariable
axis0_dimconstant = 16
axis1_dimconstant = 56
axis2_dimconstant = 56
total_dimtotal_dim = axis0_dim + axis1_dim + axis2_dim
determinismunspecified
constraintsNo mutation or aliasing
Identity
alias034_flux_multi_axis_rope_frequency_computationmodelflux-1-fill-devsha256d50aa5a7e6ab…
Sources: NVIDIA SOL-ExecBench (2026-08-12)last observed 2026-08-12How records are decidedJSON