Mojo function

grouped_matmul_amd_kernel_launcher

grouped_matmul_amd_kernel_launcher[c_type: DType, a_type: DType, b_type: DType, layout_c: Layout, layout_a: Layout, layout_b: Layout, transpose_b: Bool, config: MatmulConfig[a_type, b_type, c_type, transpose_b], elementwise_lambda_fn: OptionalReg[fn[dtype: DType, width: Int, *, alignment: Int = 1](IndexList[2], SIMD[dtype, width]) capturing -> None] = OptionalReg[fn[dtype: DType, width: Int, *, alignment: Int = 1](IndexList[2], SIMD[dtype, width]) capturing -> None](None)](c_tensor: LayoutTensor[c_type, layout_c, MutableAnyOrigin], a_tensor: LayoutTensor[a_type, layout_a, MutableAnyOrigin], b_tensor: LayoutTensor[b_type, layout_b, MutableAnyOrigin], a_offsets: NDBuffer[DType.uint32, 1, MutableAnyOrigin], expert_ids: NDBuffer[DType.int32, 1, MutableAnyOrigin], num_active_experts: Int)