Mojo function

naive_grouped_matmul

naive_grouped_matmul[c_type: DType, c_shape: DimList, a_type: DType, a_shape: DimList, b_type: DType, b_shape: DimList, //, *, transpose_b: Bool = True, elementwise_lambda_fn: OptionalReg[fn[DType, Int, Int](IndexList[2], SIMD[$0, $1]) capturing -> None] = OptionalReg[fn[DType, Int, Int](IndexList[2], SIMD[$0, $1]) capturing -> None]({:i1 0, 1})](c: NDBuffer[c_type, 2, MutableAnyOrigin, c_shape], a: NDBuffer[a_type, 2, MutableAnyOrigin, a_shape], b: NDBuffer[b_type, 3, MutableAnyOrigin, b_shape], a_offsets: NDBuffer[uint32, 1, MutableAnyOrigin], expert_ids: NDBuffer[uint32, 1, MutableAnyOrigin], max_num_tokens_per_expert: Int, num_active_experts: Int, ctx: DeviceContext)