Skip to content
KernelIndex
Search⌘K
Sign in

Mamba conv1d with gating

38 eligible runs
moe

Fused 1D causal convolution with gated projection for Mamba layers. Performs gated linear projection (hidden_size -> 2*intermediate_size), splits into hidden states and gate, applies causal 1D depthwise convolution, SiLU activation, and optional attention masking.

Fastest reported
622.1µsmean · 79% of SOL
RIAC as wellLicense unknown · unknown

Reported evidence · last observed 2026-08-19. Reported by source; not independently reproduced. No public source. License unknown.

No public sourceRun detail →

Current records

Workloadsuite of 16 cases · mean latencyhidden_size = 8192 · proj_out_size = 32768 · conv_kernel_size = 4 · intermediate_size = 1638416 cases
seq_len
batch_size
dtype
1281bf161282bf161288bf1612864bf162562bf162568bf1625632bf162932bf163734bf1651216bf165414bf1610241bf1610242bf1610248bf1621311bf1681921bf16
scrolls · 16 cases total
Aggregatessuite of 16 cases · mean latency
HardwareBest knownImplementationRuns
NVIDIA B200 · env 1622.1 µsRIAC as well27NVIDIA B200 · env 2674.0 µsdoubleAI11

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 v1.1 · mean · 27 results · last observed 2026-08-24Record history →
#
Implementation
Latency
vs #1
Trust
Observed
1
RIAC as well
622.1µs79% SOL
1.00×
Reported · license unknown · no source
2026-08-19

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

no sourcelicense unknownno install recipeRun detail →
2
Amir M. Mir | SF Tensor
651.1µs81% SOL
1.05×
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 →
3
ac4k
673.0µs78% SOL
1.08×
Reported · license unknown · no source
2026-07-11

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

no sourcelicense unknownno install recipeRun detail →
4
doubleAI
704.8µs75% SOL
1.13×
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 →
5
Hyra
711.2µs72% SOL
1.14×
Reported · license unknown · no source
2026-07-15

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

no sourcelicense unknownno install recipeRun detail →
6
Geometric
712.0µs69% SOL
1.14×
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 →
7
Geometric
718.2µs68% SOL
1.15×
Reported · license unknown · no source
2026-08-19

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

no sourcelicense unknownno install recipeRun detail →
8
Databricks
735.7µs69% SOL
1.18×
Reported · license unknown · no source
2026-07-12

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

no sourcelicense unknownno install recipeRun detail →
9
Geometric
745.6µs64% SOL
1.20×
Reported · license unknown · no source
2026-08-22

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

no sourcelicense unknownno install recipeRun detail →
10
Amir M. Mir | SF Tensor
745.6µs64% SOL
1.20×
Reported · license unknown · no source
2026-07-21

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

no sourcelicense unknownno install recipeRun detail →
11
Geometric
749.6µs64% SOL
1.20×
Reported · license unknown · no source
2026-08-16

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

no sourcelicense unknownno install recipeRun detail →
12
Geometric
756.4µs63% SOL
1.22×
Reported · license unknown · no source
2026-08-24

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

no sourcelicense unknownno install recipeRun detail →
13
Geometric
758.0µs63% SOL
1.22×
Reported · license unknown · no source
2026-08-13

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

no sourcelicense unknownno install recipeRun detail →
14
ac4k
758.7µs67% SOL
1.22×
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
759.5µs63% SOL
1.22×
Reported · license unknown · no source
2026-08-17

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

no sourcelicense unknownno install recipeRun detail →
16
Amir M. Mir | SF Tensor
763.9µs62% SOL
1.23×
Reported · license unknown · no source
2026-07-11

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

no sourcelicense unknownno install recipeRun detail →
17
Databricks
764.9µs66% SOL
1.23×
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 →
18
Geometric
766.8µs63% SOL
1.23×
Reported · license unknown · no source
2026-08-11

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

no sourcelicense unknownno install recipeRun detail →
19
Geometric
769.7µs62% SOL
1.24×
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 →
20
Geometric
775.4µs61% SOL
1.25×
Reported · license unknown · no source
2026-08-19

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

no sourcelicense unknownno install recipeRun detail →
21
Amir M. Mir | SF Tensor
781.5µs60% SOL
1.26×
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 →
22
Recursive
793.9µs60% 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 →
23
Geometric
806.6µs58% SOL
1.30×
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 →
24
Geometric
811.2µs57% SOL
1.30×
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 →
25
Cursor
925.9µs48% SOL
1.49×
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
TH3CHARLie | TileGym
1.12ms37% SOL
1.80×
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 →
27
ac4k
2.38ms21% SOL
3.82×
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 →

Implementations

Implementation
Runtime
Best latency
Evidence
Availability
Amir M. Mir | SF Tensor
unknown
651.1µs
1.05×
Reported
license unknown · no source
ac4k
unknown
673.0µs
1.08×
Reported
license unknown · no source
RIAC as well
unknown
622.1µs
1.00×
Reported
license unknown · no source
doubleAI
unknown
674.0µs
1.08×
Reported
license unknown · no source
Databricks
unknown
711.1µs
1.14×
Reported
license unknown · no source
Hyra
unknown
711.2µs
1.14×
Reported
license unknown · no source
Geometric
unknown
712.0µs
1.14×
Reported
license unknown · no source
Recursive
unknown
753.5µs
1.21×
Reported
license unknown · no source
Show all 38 implementations ›
Argus
unknown
796.5µs
1.28×
Reported
license unknown · no source
Cursor
unknown
894.4µs
1.44×
Reported
license unknown · no source
TH3CHARLie | TileGym
unknown
1.12ms
1.80×
Reported
license unknown · no source
Igor Bragaia
unknown
1.53ms
2.47×
Reported
license unknown · no source
ac4k
unknown
696.0µs
1.12×
Reported
license unknown · no source
Databricks
unknown
735.7µs
1.18×
Reported
license unknown · no source
Geometric
unknown
718.2µs
1.15×
Reported
license unknown · no source
ac4k
unknown
758.7µs
1.22×
Reported
license unknown · no source
Databricks
unknown
764.9µs
1.23×
Reported
license unknown · no source
Geometric
unknown
745.6µs
1.20×
Reported
license unknown · no source
Amir M. Mir | SF Tensor
unknown
744.1µs
1.20×
Reported
license unknown · no source
Amir M. Mir | SF Tensor
unknown
746.7µs
1.20×
Reported
license unknown · no source
Amir M. Mir | SF Tensor
unknown
745.6µs
1.20×
Reported
license unknown · no source
Geometric
unknown
749.6µs
1.20×
Reported
license unknown · no source
Geometric
unknown
758.0µs
1.22×
Reported
license unknown · no source
Geometric
unknown
756.4µs
1.22×
Reported
license unknown · no source
Geometric
unknown
766.8µs
1.23×
Reported
license unknown · no source
Geometric
unknown
759.5µs
1.22×
Reported
license unknown · no source
Geometric
unknown
769.7µs
1.24×
Reported
license unknown · no source
Amir M. Mir | SF Tensor
unknown
763.9µs
1.23×
Reported
license unknown · no source
Geometric
unknown
775.4µs
1.25×
Reported
license unknown · no source
Amir M. Mir | SF Tensor
unknown
781.5µs
1.26×
Reported
license unknown · no source
Recursive
unknown
793.9µs
1.28×
Reported
license unknown · no source
Amir M. Mir | SF Tensor
unknown
790.4µs
1.27×
Reported
license unknown · no source
Geometric
unknown
806.6µs
1.30×
Reported
license unknown · no source
Geometric
unknown
811.2µs
1.30×
Reported
license unknown · no source
doubleAI
unknown
816.1µs
1.31×
Reported
license unknown · no source
Cursor
unknown
925.9µs
1.49×
Reported
license unknown · no source
ac4k
unknown
2.38ms
3.82×
Reported
license unknown · no source
doubleAI
unknown
704.8µs
1.13×
Reported
license unknown · no source

Semantics

Inputs and outputs
hidden_statesbf16 [batch_size, seq_len, hidden_size]
attention_maskbf16 [batch_size, seq_len]
in_proj_weightbf16 [proj_out_size, hidden_size]
in_proj_biasbf16 [proj_out_size]
conv1d_weightbf16 [intermediate_size, 1, conv_kernel_size]
conv1d_biasbf16 [intermediate_size]
output_hidden_statesbf16 [batch_size, intermediate_size, seq_len]
gatebf16 [batch_size, intermediate_size, seq_len]
Axes and behavior
seq_lenvariable
batch_sizevariable
hidden_sizeconstant = 8192
proj_out_sizeproj_out_size = intermediate_size * 2
conv_kernel_sizeconstant = 4
intermediate_sizeconstant = 16384
determinismunspecified
constraintsNo mutation or aliasing
Identity
alias029_mamba_conv1d_with_gatingmodelai21-jamba-large-1-7sha2564be157544e84…
Sources: NVIDIA SOL-ExecBench (2026-08-24)last observed 2026-08-24How records are decidedJSON