fix stream handling in multi-op device tasks

ValeevGroup / tiledarray

A massively-parallel, block-sparse tensor framework written in C++

GNU General Public License v3.0

254 stars 52 forks source link

Closed evaleev closed 1 year ago

evaleev commented 1 year ago

resolves #420 and #422

evaleev commented 1 year ago

currently um_expressions_suite/*cont* and librett_suite/librett_gpu_mem, e.g.

and