StridedBatch for BLOCKED_ELL format

Accelerated Computing GPU-Accelerated Libraries
1

I am trying to use BlockedEll format for my matrix A.
And this will be multiplicated with dense matrix B using cusparseSpMM.
Is there any way I can set the stridedbatch for BlockedEll format?
I tried ‘cusparseCsrSetStridedBatch’ and ‘cusparseCooSetStridedBatch’, but doesn’t work.

2

Hi. Unfortunately, batch computation for BlockedELL hasn’t been supported yet. If you want to do multiple concurrent cusparseSpMM, you can use multiple streams.