Berlin and Göttingen, No singularity and apptainer filesystem overlays
Filesystem overlays were temporarily disabled on Emmy/Grete and Lise due to maintenance. Maintenance is now complete for Emmy/Grete and overlays have been re-enabled on Emmy/Grete.
This mainly affects Singularity and Apptainer users. If you use Singularity and Apptainer, the following limitations now apply:
* The targets of bind mounts must already exist in the SIF container (/proc, /dev, ... , and any user specified bind mounts).
* SIF containers completely read-only during execution unless they are first extracted into a sandbox.
Lise in Berlin, GPU-Partition for Power Users
For details, see the description of the NVIDIA-A100 hardware at the link Power user GPU.
Emmy in Göttingen, June maintenance
The core switches of the Emmy system were the root cause of several network related problems (luckily mostly not visible to the user side). This means we want to replace them. The core switch replacement requires a full downtime of HPC operation in Goettingen. We will combine this work with several maintenance and update tasks on the systems. The downtime will start on June 12th 8am. We expect the system to be available again in the course of Thursday June 15th.
Grete in Göttingen available
We’re happy to announce the beginning of regular user operation for our new GPU cluster, “Grete” in Göttingen.
The main part of the cluster is available via the new partition
grete, consisting of 33 nodes equipped with 4 NVIDIA Tesla A100 40 GB GPUs, 2 AMD Epyc CPUs, and an Infiniband HDR interconnect. In addition, “Grete” has a dedicated new login node,
glogin9, which is also available via its DNS alias glogin-gpu.hlrn.de.
For smaller single GPU jobs and interactive usage, we provide additional partitions.
More information about using the new GPU system can be found here, and the accounting information has been extended to include the GPUs and MIG slices. Please do not hesitate to contact us if you have questions or need support migrating suitable applications to the GPU system.
The existing GPU nodes ggpu[01-03] with Nvidia V100 32GB GPUs will be migrated to the same site (“RZGö”) as “Grete” in mid-May.