Mark Abraham





04:52 PM GROMACS Bug #3100: crash with GPU comm DD
Alan Gray wrote:
> @Mark in is1DAnd1PulseDD() you are only returning true if (dd.comm->maxpulse == 1), but in these ...
01:06 PM GROMACS Task #2971 (In Progress): Rework TPR reading to allow reading of raw bytes from disk and communication of complete information at setup time
Szilárd Páll wrote:
> Paul Bauer wrote:
> > It was decided to not focus on this for GROMACS 2020.
> Has this d...
01:04 PM GROMACS Task #3077: PME/PP GPU Comms unique pointer deletion causes seg fault when CUDA calls exist in destructor
I've not seen any issues with such patches
01:03 PM GROMACS Bug #3099 (Fix uploaded): compiler flags missing from the log report
01:03 PM GROMACS Bug #3099: compiler flags missing from the log report
This fix has been in gerrit waiting for review for a month, which presum...
10:07 AM GROMACS Revision b396a73c: Refactor to create a GpuTaskAssignments object
This change is purely reorganization, nothing is reordered or
functionality changed.
This expands on the previous ve...


07:13 PM GROMACS Revision 6fb2a76e: Use GPU halo exchange only when compatible DD is available
Change-Id: I624861e61e6f1b6854172102b12a4f9b51b94101
06:20 PM GROMACS Revision e5a48407: Further improve getDDGridSetup
Merged and simplified two similar setup functions that were sharing
the setting up of the number of PME-only ranks an...


03:21 PM GROMACS Bug #3097 (New): nbnxm grid issue with regressiontest complex/nbnxn_rzero with gpubufferops path
The nightly tests of master HEAD are running the gpubufferops and gpucomm matrices, and in them we see e.g. at http:/...
01:41 PM GROMACS Bug #3091 (Closed): simulator comparison test failing in gpucomm matrix

Also available in: Atom