Error of example 23 gemm_operand_reduction_fusion with splitK

[18321961708]

I have some questions about example23:

1. does this kernel GemmWithKReduction only support with reductionForA but not for product result?
2. When I run example with command "./23_ampere_gemm_operand_reduction_fusion --save-workspace --split-k-slices=2 --parallel-split-k" I will get result "ERROR - results miscompared. Results written to '23_ampere_gemm_operand_reduction_fusion1024x1024x1024.dat'. ID,M,N,K,SplitK-Slices,Parallel-SplitK,Runtime gemm_1,1024,1024,1024,2,1,0". And I wonder if kernel ReduceVectorSplitKKernel can help me with reduction of A*B result.