TensorFlow now available
Compliments of the season. Many of you may be asking where I have been and what I have been up to. For the past year and a bit, I’ve been working with the Astronomy crowd on the deployment of their…
Update on BeegFS storage
The disk cluster is configured and running. Four Dell R740DX servers each with a combination of twelve 10TB NLSAS and two 240GB SSD drives are providing a 345TB scratch volume, an 8TB /home volume and a 4TB software volume. Each …
Infiniband cabling
Thermal regulation and air flow are two important parts of any HPC installation. Adding Infiniband increases the amount of physical material at the back of the servers which can obstruct fans, reduce airflow and lead to overheating and damaged equipment. …
Infiniband Fabric
We’ve started installing and configuring the Infiniband connections before setting up BGFS storage. The fabric speed is 100Gb/s! …
HPC update
Following the move of Hex to Bremner we have continued installing the new HPC kit. A total of 4 Dell R740XD storage servers, 26 Dell 40 core C6420 compute nodes and 4 Dell C4140 GPU nodes with four P100 cards…
Hex cluster move
On Sunday the 16th of September Hex was moved from the Upper Campus data centre to the Bremner data centre. The move went smoothly and Hex was back up and running by Monday morning. There were some minor issues; the …
HPC downtime
There will be a full shut down of the Hex HPC cluster this coming weekend. We will be placing Hex into draining mode at 16:00 Friday 14 September and no new jobs will be accepted. The cluster will be shut…
HPC upgrade update
The data center where hex will reside is still being worked on. The spiders nest of network cables under the raised floor has been cleaned out to allow better air flow. All networking will now run in overhead…
HPC series comparison
A comparison of the speed of the various cores in our clusters. Series100 (new): Intel Xeon Gold 6138, 2GHz Series600 (hex): AMD Opetron 6376, 2.3GHz Series400 (hal): Intel Xeon 5506, 2.1GHz SeriesGPU (new): Intel Xeon Gold …
New cluster nodes
It has been a long time in coming, several years in fact, but the first of our new worker nodes went live today on our POC cluster. There is still much to be done, Infiniband is not yet ready, the…