Skip to content
KernelIndex
Search⌘K

submission 549188

ritvik412 · python · License unknown

Use it

Vendorable · source mirrored · license unknownView source →

No package. Vendor the mirrored source: 28 lines, June 9 Researcher Reciprocity License v1.0.

submission_v4.py
curl "https://kernelindex.com/api/v1/implementations/kernelbot-amd-moe-mxfp4-549188?include=source"
interfacepython
Compatibility
measured onAMD Instinct MI355X
declared hardwareAMD Instinct MI355X
architecturesgfx950
dtypesbf16, fp32, fp8_e8m0, int32, mxfp4

Benchmark evidence

1 measurement across 1 GPU, fastest first.

Operation / workload
Hardware
Latency
Rank
Observed
AMD MXFP4 MoEsuite of 7 cases
AMD Instinct MI355X
187.1µs
#718 of 782
2026-03-14

Reported · How evidence levels are derived →

Source and license

sourceavailable
revision digestsha256:3672230127873d6ae7131e6da5c87e1b095abd15ee0ed153ba3dc7b5b8976bbd
license declaredunknown
license concludedunknown
authorsritvik412
imported2026-08-26

Kernel source

submission_v4.py28 lines
#!POPCORN leaderboard amd-moe-mxfp4
#!POPCORN gpu MI355X

from task import input_t, output_t
from aiter import ActivationType, QuantType
from aiter.fused_moe import fused_moe

def custom_kernel(data: input_t) -> output_t:
    """
    Maximized Python Dispatch.
    """
    return fused_moe(
        data[0],   # hidden_states
        data[5],   # gate_up_weight_shuffled
        data[6],   # down_weight_shuffled
        data[9],   # topk_weights
        data[10],  # topk_ids
        expert_mask=None,
        activation=ActivationType.Silu,
        quant_type=QuantType.per_1x32,
        doweight_stage1=False,
        w1_scale=data[7],  # gate_up_weight_scale_shuffled
        w2_scale=data[8],  # down_weight_scale_shuffled
        a1_scale=None,
        a2_scale=None,
        hidden_pad=0,
        intermediate_pad=0,
    )

Source code from GPU Mode and the KernelBot dataset · June 9 Researcher Reciprocity License v1.0

Best evidence level for this revision: reported

JSON