Possible memory leak

[DavidDiazGuerra]

There might be a memory leak in gpuRIR_cuda::cuda_simulateRIR since the memory of orV_src is not free at the end of the function:

gpuRIR/src/gpuRIR_cuda.cu

Lines 917 to 927 in 283235d

	// Free memory
	gpuErrchk( cudaFree(pos_src) );
	gpuErrchk( cudaFree(pos_rcv) );
	gpuErrchk( cudaFree(orV_rcv) );
	gpuErrchk( cudaFree(amp) );
	gpuErrchk( cudaFree(tau) );
	gpuErrchk( cudaFree(tau_dp) );
	gpuErrchk( cudaFree(rirISM) );
	gpuErrchk( cudaFree(A) );
	gpuErrchk( cudaFree(alpha) );
	gpuErrchk( cudaFree(rirDiff) );

[shenbuguanni]

yes, I get some out of memory error like :
GPUassert: out of memory xxxx/gpuRIR/src/gpuRIR_cuda.cu 627
What effect does this have?

[DavidDiazGuerra]

I don't think you're running out of memory because of this except if you're calling gpuRIR_cuda::cuda_simulateRIRmany times with a really high number of directive sources. The problem is probably that you have already allocated a big amount of memory out of the library (for example with a deep-learning framework) or that you're trying to simulate a too-long RIR with the ISM.

[shenbuguanni]

Yes, Because I want to generate a large number of RIRs in multiple processes at the same time, I will run many python scripts at the same time, will this error affect the generation of RIR data? If not, this is the way it is

[DavidDiazGuerra]

The last version of the master branch should have the memory leak fixed now, but you'll probably keep having that error if you create so many processes that you run out of GPU memory. That's not a bug in the library, it's just your computer running out of GPU memory, the same as if it was running out of RAM.