- Turing is named after the legendary pioneer of compute science Alan Turing.
The Turing cluster went online around 2013, and had received several major expansions (in 2015 and 2016).
- Wahab is named after ODU's great computer science educator Dr. Hussein Abdel-Wahab.
The Wahab cluster went online around 2018, it is currently ODU's newest cluster. We gratefully acknowledge the support of U.S. National Science Foundation (MRI grant #1828593) for the acquisition of Wahab cluster.
Turing & Wahab are both traditional HPC computation-focused cluster.
- Both Wahab and Turing clusters use SLURM job scheduler to manage workloads on their computing resources.
- They are both equipped with Infiniband network for high-speed intercluster communication;
- They are both equipped with a Lustre storage device for high-performance short term storage;
- They share same home directory and RC (long-term research storage) directory;
- They do not share the Lustre storage;
- Their compute node can only talk to other nodes within the same cluster.
Here is a summary table of the specifications of the two clusters:
| Cluster |
Num of CPU cores |
Num of GPUs |
Memory (aggregated) |
Scratch space |
| Wahab |
~6320 |
72 |
60 TB |
350 TB |
| Turing |
~5600 |
36 |
34 TB |
180 TB |
More details can be found in our main website.
¶ Are Turing and Wahab Suitable for My Computation?
You should consider running your application on Turing or Wahab if the following conditions satisfy your needs:
-
It runs on a Linux Platform
We do not support any other platform at this point and it is unlikely that we will in the foreseeable feature.
-
It benefits from parallel computing
Your application could utilize more than one cpu core from multi-processing, multi-threading, MPI or a similar technique.
-
It does not require HDFS
Turing carry some ability to run application from Hadoop eco system, however, it does not support HDFS.
- Hadoop cluster is a small, teaching focused cluster
- It runs hortonworks hadoop distribution and
- It provides standard HDFS, map reduce, spark, hbase etc...
- It is equipped with Infiniband networking device for communication
- Infiniband
- The Infiniband network provides high speed message passing between compute nodes.
- It is much faster than Ethernet on the cluster and we strongly encourage you to use it whenever possible.
- The easiest way to utilize it is using MPI. If you can adjust your application to use MPI, you should do it.
- Another way to utilize Infiniband may involve directly relying on OFED libraries, such
libverbs or libfabric
- Please feel free to contact us if you need any help enabling Infiniband for your application.
- Ipoib (Ethernet over Infiniband)
- Running Standard Ethernet network over Infiniband hardware, it is slower than native infiniband protocol but still very faster. If using native Infiniband is not an option, use Ipoib.
- By default all tcp/ip traffic on Turing, Wahab & Hadoop are routed through Ipoib
- Ethernet
- Standard Ethernet network which is much slower than Infiniband. Only use it when talking to internet.