From 02/07/2014 to 03/08/2014
- 05:30 pm Bug #2802: HOOMD no longer recognizing "--ncpu" as a valid option
On Fri, Mar 7, 2014 at 4:20 PM, wrote:
- 04:20 pm Bug #2802 (Closed): HOOMD no longer recognizing "--ncpu" as a valid option
- This option is to explicitly disable OpenMP. OpenMP support in HOOMD was removed by Josh as of
- 12:59 pm Bug #2802 (Closed): HOOMD no longer recognizing "--ncpu" as a valid option
- In testing some new job scripts, the following command, which has worked for months, is now barfing in the following ...
- 08:39 am Feature #2104: Enable adding/removing bonds dynamically in multi-GPU simulations
- Did you already complete this with your updated bonded group code?
- 08:38 am Task #1486 (Resolved): Make BondData, AngleData and DihedralData instantiations of a class template
- 08:37 am Task #193 (Rejected): Implement generic type name <-> id mapping class
- With the new collapsing of bonded group data, this really isn't needed any more. In shapshots, a simple list of strin...
- 08:36 am Task #338 (In Progress): Convert pair.cgcmm to templated method
- 08:34 am Bug #2101 (Resolved): Optimize for small number of particles per GPU
- 08:33 am Bug #179 (Resolved): Example patch
- 08:33 am Bug #179: Example patch
- Patches are no longer the recommended way to contribute to hoomd. Fork hoomd-blue on github instead.
- 05:09 pm Task #1787 (Resolved): Update license text
- Applied in changeset commit:f2cf8fc95e750117163fb6fd3c0c74724f612948.
- 04:57 pm Bug #2796 (Resolved): pair.table reports incorrect pressure
- Fixed in commit:113cb94802675f7351610078767a24b1dd5ab983. The issue was not with pair.table but with the nvt integrat...
- 11:22 am Task #2645 (Resolved): Use __ldg() for all texture reads on SM 3.5
- Applied in changeset commit:73bfbbe36b5c5750a86d6d92f05767130fe78235.
- 07:55 pm Bug #1608 (Resolved): Bug in init.create_empty
- Applied in changeset commit:7bf89feef4fbbaff4f8b118bc75c388e8747c5ba.
- 11:30 am Feature #2544 (Resolved): update.box_resize() should start from current box parameters
- Applied in changeset commit:3666e6d35933285617b5bbbdbb3b8d3daacfa377.
- 11:14 pm Task #2633: Strong scaling and cache locality
- I think you've figured all these issues out now. Correct?
- 11:13 pm Feature #2544 (In Progress): update.box_resize() should start from current box parameters
- This will be fine as an option. I like the convenience of only needing to specify one parameter in cubic boxes.
- 11:10 pm Task #2645 (In Progress): Use __ldg() for all texture reads on SM 3.5
- I will finish this one up.
- 11:02 pm Task #2644: Shared memory version of all pair force kernels
- I wouldn't bother with updating Tersoff. It's not used by very many people. CGCMM will automatically get this when we...
- 10:56 pm Task #1507 (Resolved): Remove OpenMP
- Applied in changeset commit:1409024222db694f7462e313af74c26d22d58260.
- 09:58 pm Bug #2796 (Resolved): pair.table reports incorrect pressure
- User reports that pair.table is resulting in incorrect pressures (with rigid bodies).
- 06:21 pm Feature #2790: New replicate() command
- As this is 90% in the snapshot and the other 10% is enabling a feature to restore snapshots with different N, this is...
- 06:16 pm Task #291 (Resolved): Option dependant block size tuning
- Applied in changeset commit:94f273e9b174b5db55ec67324b7d1cb6a1d8fba7.
- 01:27 pm Task #2792 (New): Check that MPI_SUPPORTED / MPI_NOT_SUPPORTED docs are current
- In all the updates on MPI functionality, I'm not sure that all the docuemntation got updated along with it. Jens, cou...
- 01:10 pm Task #2791 (New): Remove 1.x kernels
- Support for compute 1.x is deprecated in HOOMD v1.0. Let's remove all the old legacy codepaths specific to that archi...
- 12:23 pm Bug #1994: Data Proxies throwing Seg Faults
- Nothing can be done for 1.0. If OS X 10.9 truly solves the issue then we need to update to build on that OS. Moving t...
- 10:09 am Task #1688 (Resolved): Enable easier usage of hoomd imported into other scripts
- 08:25 am Feature #772: Tabulated Angle Bonds
Can you look into the correctness of the tabulated angles and dihedrals? The documentation might also need s...
- 04:42 pm Feature #2790: New replicate() command
- Note that there is one current limitation. The domain decomposition grid is not updated upon replicating along the ax...
- 03:36 pm Feature #2790 (New): New replicate() command
The branch replicate contains the new replicate feature() on the python level
(commit:9fd4e351) There is al...
- 03:58 pm Bug #2702 (Resolved): HOOMDDumpwriter writes 0's for dihedrals and impropers
- Applied in changeset commit:b1d21111dca6102c07f0a15eacc9bae19b25a190.
- 10:01 am Feature #2781 (Resolved): Autotuning of overlapping communication/computation
- Applied in changeset commit:66d05a81afac7f4e0ceaaa457fa642c971a0ca60.
- 09:55 am Feature #2781: Autotuning of overlapping communication/computation
- The changes are good and average is possibly better than median for tuning binary options. Median is needed for faste...
- 07:40 pm Task #2644 (In Progress): Shared memory version of all pair force kernels
- Standard & DPD pair potentials switched over to warp-synchronous version/autotuning.
Do we have an example script ...
- 07:36 pm Bug #2705 (Resolved): Fix initialization with MPI_CUDA
- Now the behavior is documented on the MPI concepts page.
Plus, Titan GPUDirect/RDMA is broken, so no fix specificall...
- 07:35 pm Bug #2676 (Closed): Implement better distance check for MPI simulations
- The second statement is false, since nlist->compute() will not compute the neighbor list more than once per time step...
- 07:33 pm Task #500 (Closed): OMPI bug
- Haven't seen this lately
- 07:20 pm Feature #2781: Autotuning of overlapping communication/computation
- I actually meant commit:6d90cb7f .
- 07:17 pm Feature #2781 (Resolved): Autotuning of overlapping communication/computation
please have a look at commit:d5f732d5 in branch tune_precompute.
It introduces a slight modification to the...
- 11:46 am Bug #2766 (Resolved): Multi-GPU DPD equilibrates to wrong temperature
- Fixed in commit:59424cbb . I had forgotten to communicate tags for ghosts.
- 03:03 pm Bug #2766 (Resolved): Multi-GPU DPD equilibrates to wrong temperature
- The equilibration temperature using the DPD thermostat is off by a few percent when using MPI simulations.
- 10:00 am Bug #2697 (Rejected): Completely positive lj potential logs as negative
- 08:49 am Task #2466 (Rejected): fix gentoo ebuild
- not worth the effort needed.....
- 08:44 am Bug #2559 (Resolved): init.create_empty + setting particle positions via data access broken in MPI
- 08:42 am Task #1565 (Resolved): Add documentation page for concepts
- 01:30 pm Bug #2703 (Resolved): failing unit test test_nvt_integrator_mpi
- confirmed fixed.
Also available in: Atom