Using new cusparse<T>gemvi() routine

I’ve seen that cusparse can now handle dense matrix x sparse vector multiplies via the cusparsegemvi() function. I can’t seem to find any documentation on it though. I’ve looked through the cusparse doc for the 7.5 toolkit, but it doesn’t seem to be in there.

Can you direct me to the documentation?

Placing the link to the GEMVI function reference in the cuSPARSE documentation for anyone coming across this question: