Issues

• DevicePacker doesn't pack current data after swap when using cudaGraph API

  bug
  cwpearson/stencil#30 · Filter by author cwpearson opened on Mar 24, 2021
  

• Support non-pitched allocations

  enhancement
  cwpearson/stencil#29 · Filter by author cwpearson opened on Mar 18, 2021
  1

  

• STENCIL_SETUP_STATS should track different MPI and CUDA copy sizes

  enhancement
  cwpearson/stencil#25 · Filter by author cwpearson opened on Aug 20, 2020
  

• Handling CUDA_VISIBLE_DEVICES

  enhancementneedsinfoperformanceresource:summit
  cwpearson/stencil#24 · Filter by author cwpearson opened on Aug 18, 2020
  

• STENCIL_SETUP_STATS should record same-GPU, same-node, and off-node comm totals

  enhancement
  cwpearson/stencil#23 · Filter by author cwpearson opened on Aug 17, 2020
  

• CMake 3.18.1 policy warning with unset CMAKE_CUDA_ARCHITECTURES

  bugresource:summit
  cwpearson/stencil#22 · Filter by author cwpearson opened on Aug 10, 2020
  

• Self-exchange with CudaMpi | CudaMpiColocated happens with CudaMpi instead of CudaMpiColocated

  needsinfoperformance
  cwpearson/stencil#20 · Filter by author cwpearson opened on Jun 11, 2020
  

• Test fails if not started with cudaDeviceReset()

  bugneedsinfo
  cwpearson/stencil#19 · Filter by author cwpearson opened on Jun 8, 2020
  

• measure-buf-exchange can't run with --smpiargs="-gpu"

  resource:summit
  cwpearson/stencil#16 · Filter by author cwpearson opened on Mar 5, 2020
  

• Call to a CUDA runtime function without setting device first

  needsinfo
  cwpearson/stencil#15 · Filter by author cwpearson opened on Feb 28, 2020
  

• invalid IPC handle when using CUDA-Aware MPI

  bug
  cwpearson/stencil#12 · Filter by author cwpearson opened on Feb 17, 2020
  2

  

• CUDA Runtime Error(46): all CUDA-capable devices are busy or unavailable on Summit

  bugresource:summit
  cwpearson/stencil#11 · Filter by author cwpearson opened on Feb 10, 2020
  4

  

• Use detected mpi, not the mpirun in $PATH

  bug
  cwpearson/stencil#10 · Filter by author cwpearson opened on Jan 30, 2020
  

• Wrongly assumes all nodes will have the same data placement

  bug
  cwpearson/stencil#9 · Filter by author cwpearson opened on Jan 29, 2020
  

• MPI_Comm_free called after MPI_Finalize

  bug
  cwpearson/stencil#8 · Filter by author cwpearson opened on Jan 29, 2020
  

• Wrongly assume that all ranks share the same GPU topology

  bug
  cwpearson/stencil#7 · Filter by author cwpearson opened on Jan 14, 2020
  

• Compute GPU distance matrix by actual bandwidth

  enhancement
  cwpearson/stencil#6 · Filter by author cwpearson opened on Jan 13, 2020
  

• interface to retrieve lower bound and upper bound of DistributedDomain

  enhancement
  cwpearson/stencil#4 · Filter by author cwpearson opened on Jan 12, 2020
  

• Allow custom allocator for DistributedDomain and LocalDomain

  enhancement
  cwpearson/stencil#2 · Filter by author cwpearson opened on Jan 12, 2020
  1