Will cusparse begin to support 64 bit int?

[sBc-Random]

I’m currently using cusparse quite heavily in my calculations, and indeed run into situation where there are >2bil nnz. At the moment this is being split up over 8 gpu, so it’s fine (and what’s more, it isn’t possible to store the whole matrix on a single card - 2 bil nnz>=16GB float or 32GB double. However, with the memory of gpus doubling every 12 months or so (and with a 12GB card on the way), it’s not going to take more than a couple of years for these problems to suit GPUs. My question is - does cusparse plan to start supporting passing nnz as an int64?