Memory leaks?

[mtazzari]

Check all C/CUDA code against memory leakages.

Check that all the mallocs are effectively freed.

[fredRos]

To verify this,

• [x] CPU: run with Intel Inspector: no errors found for d913ec7278afe
• [x] GPU: nvidia tool?

[mtazzari]

I have checked for memory leaks in the GPU code using the memcheck tool provided by NVIDIA with:

cuda-memcheck py.test python/test_galario.py --gpu=1
cuda-memcheck --tool synccheck py.test python/test_galario.py --gpu=1

getting this result:

========= CUDA-MEMCHECK
=================================== test session starts ===================================
platform linux -- Python 3.6.1, pytest-3.1.1, py-1.4.34, pluggy-0.4.0
rootdir: /home/mtazzari/repos/galario/build3, inifile:
collected 39 items

python/test_galario.py .......................................

=============================== 39 passed in 22.95 seconds ================================
========= ERROR SUMMARY: 0 errors

From NVIDIA docs:

3.1. What is Memcheck ? The memcheck tool is a run time error detection tool for CUDA applications. The tool can precisely detect and report out of bounds and misaligned memory accesses to global, local, shared and global atomic instructions in CUDA applications. It can also detect and report hardware reported error information. In addition, the memcheck tool can detect and report memory leaks in the user application.

I don't close the issue since we made changes to the CPU code as well. Before closing, ensure there are no leaks in the CPU version.

[fredRos]

perhaps we can run memcheck or valgrind in travis. It will be much slower but ok