Overview of Halstead-GPU
Halstead-GPU is a new type of addition to Purdue's Community Clusters, designed specifically for applications which are able to take advantage of GPU accelerators. While applications must be specially-crafted to use GPUs, a GPU-enabled application can often run many times faster than the same application could on general-purpose CPUs. Due to the increased cost of GPU-equipped nodes, Halstead-GPU is being offered with some new purchase options to allow for shared access at a lower price point than the full cost of a node.
Halstead-GPU was built through a partnership with HP and Intel in May 2017. Halstead-GPU consists of HP compute nodes with two 10-core Intel Xeon-E5 processors (20 cores per node), 256 GB of memory, and two Tesla P100 GPUs. All nodes have 100 Gbps Infiniband interconnect and a 5-year warranty.
To purchase access to Halstead-GPU today, go to the Cluster Access Purchase page. Please subscribe to our Community Cluster Program Mailing List to stay informed on the latest purchasing developments or contact us via email at rcac-cluster-purchase@lists.purdue.edu if you have any questions.
Halstead-GPU Namesake
Halstead-GPU is named in honor of Maurice H. Halstead, Professor of Computer Science and Software Science pioneer. More information about his life and impact on Purdue is available in an ITaP Biography of Halstead.
Halstead-GPU Specifications
| Front-Ends | Number of Nodes | Processors per Node | Cores per Node | Memory per Node | Retires in |
|---|---|---|---|---|---|
| With GPU | 2 | Two Haswell CPUs @ 2.60GHZ with Two Tesla P100 GPUs | 20 | 256 GB | 2022 |
| Sub-Cluster | Number of Nodes | Processors per Node | Cores per Node | Memory per Node | Retires in |
|---|---|---|---|---|---|
| A | 4 | Two Haswell CPUs @ 2.60GHz with Two Tesla P100 GPUs | 20 | 256 GB | 2022 |
Halstead-GPU nodes run CentOS 7 and use Moab Workload Manager 8 and TORQUE Resource Manager 5 as the portable batch system (PBS) for resource and job management. The application of operating system patches occurs as security needs dictate. All nodes allow for unlimited stack usage, as well as unlimited core dump size (though disk space and server quotas may still be a limiting factor).
On Halstead-GPU, ITaP recommends the following set of compiler, math library, and message-passing library for parallel code:
- Intel 17.0.1.132
- MKL
- Intel MPI
This compiler and these libraries are loaded by default. To load the recommended set again:
$ module load rcacTo verify what you loaded:
$ module list
