Details
-
Sub-task
-
Status: Open
-
Major
-
Resolution: Unresolved
-
None
-
None
-
None
-
None
Description
Either by adding custom kernel or using cuSparse API.
See org.apache.sysml.runtime.matrix.data.LibMatrixCUDA's matmultTSMM() method. Please move this to Phase 2 if custom kernel is required