CUDA Application runs fine, but returns error when using "nvprof"

Using CUDA 9.2 on a Tesla V100. Application runs totally fine by itself, returning without any error. However, if I try to run nvprof on it, I get a CUDA error (Error Code 6) on a “cufftExecR2C” call. Also get warnings about insufficient device buffer space and insufficient semaphore space. I can’t seem to fix these, but don’t even know if it related to the problem. Any ideas? Error output snippet is below. Also, if it matter, I am using cuda dynamic parallelism.

CUDA error at StreamTest.cu:2143 code=6(CUFF_EXEC_FAILED) “cufftExecR2C(…”
Warning: 3 records have invalid timestamps due to insufficient device buffer space. You can configure the buffer space using the option --device-buffer-size.
Warning: 3 records have invalid timestamps due to insufficient semaphore pool size. You can configure the pool size using the option --profiling-semaphore-pool-size.

Hi, brainkinmd

cdp profile is not supported on Volta.
Please refer the https://docs.nvidia.com/cuda/profiler-users-guide/index.html#profiler-known-issues

When profiling an application that uses CUDA Dynamic Parallelism (CDP) there are several limitations to the profiling tools.
----Tracing is not supported on devices with compute capability 7.0.