Hosted at the University of Namur, this system currently consists of 1536 cores spread across 30 AMD Epyc Naples and 32 Intel Sandy Bridge compute nodes.
The AMD nodes include:
- 24 nodes with a single 32-core AMD Epyc 7551P @ 2.0 GHz and 256 GB RAM
- 4 nodes with the same CPUs but 512 GB RAM
- 2 nodes with dual 32-core AMD Epyc 7501 @ 2.0 GHz and 2 TB RAM
The Intel nodes feature dual 8-core Xeon E5-2660 @ 2.2 GHz and 64 or 128 GB RAM (8 nodes).
GPU resources include:
- 1 node with 3× NVIDIA RTX A6000 (48 GB GDDR6 each, Ampere), a 32-core AMD Epyc 7502P and 256 GB RAM
- 1 node with 8× NVIDIA A40 (48 GB GDDR6 each, Ampere), dual 32-core AMD Epyc 7543 and 512 GB RAM
All nodes are interconnected via a 10 Gigabit Ethernet network and have access to three NFS filesystems totaling 100 TB.
Suitable for:
Shared-memory parallel jobs (OpenMP, Pthreads) or resource-intensive sequential workloads, especially large-memory jobs, and GPU-accelerated workloads (CUDA) on the GPU nodes.
Resources
- Home directory (200 GB quota per user)
- Working directory
/workdir(400 GB,$WORKDIR) - Local working directory
/scratch($LOCALSCRATCH), dynamically defined in jobs - Nodes have internet access
- default queue* — Max 15 days
- hmem queue* — ≥ 64 GB/core, max 15 days
- gpu queue* — Max 5 days, 8 GPUs/user
- Max 128 CPUs/user across all partitions
- Generic resource*:
gpu— request a card type with--gres=gpu:QuadroRTXA6000:1or--gres=gpu:TeslaA40:1
Access / Support
SSH to hercules2.ptci.unamur.be (port 22) via your university gateway using your login and an id_rsa.ceci key.
Support : ptci.support@unamur.be
Server SSH key fingerprints:
What’s this ?
RSA:
SHA256:LzByp8XBhpgy+2lB1DZcpieYUCSq8FEfLBLPm+WB8xg
ED25519:
SHA256:fHuc0Y+QuAZW2FrI9NXrfDt2CeDmVWD6wHeDW4I3ztw
ECDSA:
SHA256:SyLaaBe7CuO7Dpa6vJa0vbAUxnYSpl30xaJo5yBF//c
