Hi GROMACS users,
Thanks to all of you who tested the GROMACS 2023 beta release! Thanks to you, we improved a couple of things on the beta release and are proud to announce the first release candidate of GROMACS 2023!
We are making this available to you to get an early taste of how GROMACS 2023 will look and work, and most importantly to get feedback from you about how well things work. While we try our hardest to keep the quality of GROMACS as high as possible, we’re human, we overlook things while doing other things, and we need your many pairs of eyes to help to build a tool that we can all use to do good science! We really need you to test your kinds of simulation on your hardware, both for correctness and performance. This is particularly important if you are using “interesting” hardware or compilers because we can’t test all of them!
Please do not use this version for doing science you plan to publish - it needs more testing before it’s reliable enough for that. Similarly, please don’t use this version as a base for a project that bundles or forks GROMACS.
What new things can you expect? (See the release notes for more details.)
The SYCL GPU implementation, which is the GPU portability layer that supports all major GPU platforms, has received major extensions in support for both platforms and features. To ensure portability in practice, the GROMACS GPU portability layer is actively developed with multiple SYCL implementations (hipSYCL, oneAPI DPC++, IntelLLVM) and regularly tested on multiple GPU backends.
- SYCL supports more GPU offload features: bonded forces and direct GPU-GPU communication with GPU-aware MPI.
- SYCL hardware support includes AMD (including RDNA support added here) and Intel for production as well as NVIDIA GPUs (not for production).
- SYCL optimizations targeting important HPC platforms.
PME decomposition has been optimized and extended to support offloading the entire PME calculation to multiple GPUs, including the FFT computation; when combined with cuFFTmp, this enables much improved strong scaling (experimental feature).
CUDA Graph support has been added to execute GPU-resident single-/multi-GPU simulations using thread-MPI entirely on the GPU to improve performance (experimental feature).
Apple M1/M2 GPUs are now supported via the OpenCL GPU backend.
New ensemble temperature mdp options allow setting the temperature of the ensemble for simulations without temperature coupling or with different reference temperatures.
With gmx dssp, GROMACS now has a native implementation of the DSSP algorithm, which replaces gmx do_dssp.
Check the installation guide, user guide, and reference manual for the details.
There are lots of other new things, and a few old things removed - please see the release notes for the complete list. All the content of GROMACS 2022.4 is present, apart from features that have been removed. Your testing helps us to harden the final release. To follow progress and give further feedback, have a look at the 2023 milestone on gitlab and report issues there .
If all goes to plan, we hope to ship the final 2023 release at the end of January/beginning of February, but that relies on people joining in and helping us to test! We hope you will consider making that contribution, so that we can continue to deliver high-quality free simulation software that will be useful to you.
You can find the code, manual, release notes, installation instructions and test suite at the links below.
(includes install guide, user guide, reference manual)