FusedMHARunnerV2 function mSM check

System Info

FusedMHARunnerV2::FusedMHARunnerV2(MHARunnerFixedParams fixedParams)
: mFixedParams(fixedParams)
{
TLLM_CHECK_WITH_INFO((mSM == kSM_80 || mSM == kSM_86 || mSM == kSM_89 || mSM == kSM_90 || mSM == kSM_120),
"Unsupported architecture");
TLLM_CHECK_WITH_INFO((mFixedParams.dataType == DATA_TYPE_FP16 || mFixedParams.dataType == DATA_TYPE_BF16
|| mFixedParams.dataType == DATA_TYPE_E4M3),
"Unsupported data type");
xmmaKernel = getXMMAKernelsV2(mFixedParams.dataType, mSM);

if (mFixedParams.headSizeV == 0)
{
    mFixedParams.headSizeV = mFixedParams.headSize;
}
// Get device attributes.
int device_id;
cudaGetDevice(&device_id);
cudaDeviceGetAttribute(&mMultiProcessorCount, cudaDevAttrMultiProcessorCount, device_id);
cudaDeviceGetAttribute(&mDeviceL2CacheSize, cudaDevAttrL2CacheSize, device_id);
auto const [free_memory, total_memory] = tensorrt_llm::common::getDeviceMemoryInfo(false);
mTotalDeviceMemory = total_memory;

}

can tensorr-llm only support kSM_80 kSM_86 kSM_89 kSM_90 kSM_120?Can tensorr-llm support kSM_87？The support plan?

Who can help?

No response

Information

• The official example scripts
• My own modified scripts

Tasks

• An officially supported task in the examples folder (such as GLUE/SQuAD, ...)
• My own task or dataset (give details below)

Reproduction

NO

Expected behavior

NO

actual behavior

nothing

additional notes

nothing

Start with FusedMHARunnerV2::FusedMHARunnerV2 and its mSM validation, then trace getXMMAKernelsV2 for the listed architectures. Determine whether kSM_87 has supported kernels and what the project’s support policy requires; done means a clear support decision and an agreed implementation or documentation path.