Important note! A small bug has slipped into the regression test script which causes the "complex/nbnxn_vsites" test to fail in some cases. It happens *only* when running the tests on a machine with supported GPUs using a CUDA-enabled build (but you won't see it if you have exactly 6 GPUs :).
So, when you run "make check" or manually run the regression tests script on a GPU build, please ignore the "complex/nbnxn_vsites" failure. Even better, rerun the test with forcing no GPUs to be used e.g. with the CUDA_VISIBLE_DEVICES="" environment variable. This way all tests will only use the CPU (in fact this is always recommended in order to test both CPU and GPU kernels). Cheers, -- Szilárd On Sun, Jun 29, 2014 at 10:53 PM, Mark Abraham <mark.j.abra...@gmail.com> wrote: > Hi GROMACS users, > > The official release of GROMACS 5.0 is available! > > What new things can you expect? > > > * GROMACS now uses a combination of *C++98* and C99, so you will need a > working C++ compiler that is intended to work with your C compiler > > > * Building GROMACS now requires *CMake version 2.8.8*, which will mean some > of you will need to update your version in order to build GROMACS. > > > * The native GPU port available in GROMACS 4.6 supports a wider range of > simulation types, and now requires CUDA 4.0 > > > * The *Verlet cut-off scheme is now the default* in GROMACS. One of our > design principles is that GROMACS will do correct simulations by default, > and the use of this scheme makes that easy and fast. The group cut-off > scheme will be removed in a future version, so start planning for that! > > * You can now treat *Lennard-Jones interactions with PME*, which is expected > to provide a much better treatment of spatially heterogeneous long-ranged > dispersion interactions at cost comparable to that of a cutoff. LJPME is > supported for both cut-off schemes, and almost all kinds of simullations > > * GROMACS deploys the Random123 stateless random-number generator > > > * GROMACS now uses the stochastic dynamics (sd1) integrator of Goga & > Berendsen > > > * GROMACS tools are now bundled together into *just two binaries* - mdrun > and gmx. The latter contains most of the tools familiar to you, perhaps with > new names, and a couple of new ones. For the time being, the install > procedure will also create symbolic links so that your old scripts will keep > working for a time, but these will go away at some future point! > > * There is support for the new TNG compressed trajectory format > > * GROMACS contains a feature for "computational electrophysiology" > simulations, to simulate a steady-state ion concentration difference between > compartments of a double membrane > > > * GROMACS supports some special interactions for coarse-grained interactions > with Martini force fields > > * GROMACS support for Interactive Molecular Dynamics (IMD) > > > * A handful of the tools have been re-implemented using a *new C++ analysis > framework*, which we hope will provide a better platform for maintaining and > adding to their functionality in the future. > > > * There is an enhanced testing suite, with *unit tests* for most new C++ > code. > > > * There is preliminary support for Intel Xeon Phi (only in native mode, and > performance will only be slightly improved) > > * GROMACS has removed support for particle decomposition - the simulation > types that required it are now limited to a single MPI rank (but perhaps > multiple OpenMP threads) > > > Please see the link to the release notes below for more details. All the > content of GROMACS 4.6.6 (whose release is imminent) is present, apart from > features that have been removed. > > You can find the code, manual, release notes, installation instructions and > test suite at the links below. > > ftp://ftp.gromacs.org/pub/gromacs/gromacs-5.0.tar.gz > > ftp://ftp.gromacs.org/pub/manual/manual-5.0.pdf > > http://www.gromacs.org/About_Gromacs/Release_Notes/Versions_5.0.x#Release_notes_for_5.0 > > http://www.gromacs.org/Documentation/Installation_Instructions_for_5.0 > > http://gerrit.gromacs.org/download/regressiontests-5.0.tar.gz > > > Happy simulating! > > > Mark Abraham > > GROMACS development manager > > _______________________________________________ > Announcements from the Gromacs team > https://maillist.sys.kth.se/mailman/listinfo/gromacs.org_gmx-announce -- Gromacs Users mailing list * Please search the archive at http://www.gromacs.org/Support/Mailing_Lists/GMX-Users_List before posting! * Can't post? Read http://www.gromacs.org/Support/Mailing_Lists * For (un)subscribe requests visit https://maillist.sys.kth.se/mailman/listinfo/gromacs.org_gmx-users or send a mail to gmx-users-requ...@gromacs.org.