For some batched routines, we need a vector of onemklTranspose, onemklDiag, or onemklSide.
I allocate them in the C interface with the function convert but it's never deallocated after the call to the oneMKL routine, like here with convert(uplo, group_count).
For some batched routines, we need a vector of
onemklTranspose
,onemklDiag
, oronemklSide
. I allocate them in the C interface with the functionconvert
but it's never deallocated after the call to the oneMKL routine, like here withconvert(uplo, group_count)
.