QMCPACK / QMCPACK/qmcpack

CompositeSPOSet does support OMPTarget and just crashes

Open
#3,556 0 comments 0 reactions 0 assignees View on GitHub
bug
Dominant language
C++
Stars
403
Forks
154
Avg merge
1d 12h
Merged PRs (30d)
82

Description

CompositeSPOSet causes a crash if code actually tries to evaluate with it when ENABLE_OFFLOAD=1 ENABLE_CUDA=1.

**To Reproduce**
```shell
epd@leconte:~/qmcpack/build_offload_leconte/src/QMCWaveFunctions/tests$ ./test_wavefunction_sposet 'CompositeSPO::diamond_1x1x1'
Rank = 0 Free Memory = 498393 MB
WARNING More CUDA devices than the number of MPI ranks. Some devices will be left idle.
There is potential performance issue with the GPU affinity. Use CUDA_VISIBLE_DEVICE or MPI launcher to expose desired devices.
WARNING More OpenMP devices than the number of MPI ranks. Some devices will be left idle.
There is potential performance issue with the GPU affinity.
WARNING The number of MPI ranks on the node is not divisible by the number of accelerators. Imbalanced load may cause performance l\
oss.
Filters: CompositeSPO::diamond_1x1x1 CompositeSPO::diamond_1x1x1
CUDA error: an illegal memory access was encountered
Libomptarget error: Call to targetDataEnd failed, abort target.
Libomptarget error: Failed to process data after launching the kernel.
Libomptarget error: Run with LIBOMPTARGET_INFO=4 to dump host-target pointer mappings.
SplineC2COMPTarget.cpp:596:5: Libomptarget fatal error 1: failure of target construct while offloading is mandatory

~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
test_wavefunction_sposet is a Catch v2.13.6 host application.
Run with -? for options

-------------------------------------------------------------------------------
CompositeSPO::diamond_1x1x1
-------------------------------------------------------------------------------
../src/QMCWaveFunctions/tests/test_CompositeSPOSet.cpp:22
...............................................................................

../src/QMCWaveFunctions/tests/test_CompositeSPOSet.cpp:22: FAILED:
{Unknown expression after the reported line}
due to a fatal error condition:
SIGABRT - Abort (abnormal termination) signal

===============================================================================
test cases: 1 | 1 failed
assertions: 3 | 2 passed | 1 failed
```

**Expected behavior**
Not crash, evaluate splines in the locality expected.

**System:**
Leconte

Contributor guide

Open the contributing guide

Assessment

This issue has not been assessed yet.

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.