Regarding the use of cudaMemcpyAsync and cudaStreamSynchronize, the data is not copied from the device back to the host.

[weijizhen]

Problem: When I use cudaMemcpyAsync, the data is not copied from the device back to the host.

Should I, or roughly where, modify the GPGPU-Sim source code?

[JRPan]

It should be supported. Something else is probably wrong.