We run demo( "concurrentkernels") in samples.git, it cost nearly 13.2GB virtual memory（orin platform) , anything goes wrong?

Accelerated Computing CUDA CUDA Programming and Performance
1

we use cuda sample that the library contained, on origin platform, the sample code cost nearly 13.2GB virtual memory. anything goes wrong?
./concurrentKernels
top
PID PPID USER STAT VSZ %VSZ CPU %CPU
10202 2457 root R 13.2g 102.8 10 7.5

2

Is this related to the cuda-memcheck tool or is the memory issue without using any tools? If no tools are used, this should be moved to the cuda programming and performance forum.

3

i haven’t use any cuda-memcheck tool , this memory-check is test with linux top command.

4

virtual memory is just carving out an address space. It’s not a problem.