This fixes a bug in the upstream code. There is an extra argument in the call to the x_buffer.CopyToAsync function that must be removed. Otherwise the build fails when CUDA is enabled. --- CLBlast/src/routines/level2/xtrsv.cpp 2022-05-21 08:44:13.811816246 -0600 +++ CLBlast_fixed/src/routines/level2/xtrsv.cpp 2022-05-21 08:28:34.601124921 -0600 @@ -154,7 +154,7 @@ } // Retrieves the results - x_buffer.CopyToAsync(queue_, x_size, b_buffer, event_); + x_buffer.CopyToAsync(queue_, x_size, b_buffer); } // =================================================================================================