New resource available for GPU-accelerated research computing
A new computing resource is available for Purdue researchers running applications that can take advantage of graphics processing unit (GPU) accelerators.
The system, known as Halstead-GPU, is a newly GPU-equipped portion of Halstead, Purdue’s newest community cluster research supercomputer. Halstead-GPU nodes consist of two 10-core Intel Xeon E5 CPUs per node, 256 GB of RAM, EDR Infiniband interconnects and two NVIDIA Tesla P100 GPUs. The GPU nodes have the same high-speed scratch storage as the main Halstead cluster.
A GPU-enabled application will often run many times faster on a GPU node than the same application could run on a general-purpose node. The new cluster’s P100 GPUs are well-suited to accelerate applications in molecular dynamics, deep learning, quantum chemistry and many other domains.
Interested researchers can now buy space in Halstead-GPU through ITaP Research Computing’s cluster orders website. To make the GPUs more affordable, ITaP is offering users a new subscription-based purchase option.
For $1,200 per year, researchers can purchase access to a shared queue available only to those who have purchased access. Researchers who buy shared access will be able to add their students to their queues the same way they can on other community clusters. Those who want a dedicated queue still have the option to purchase a node, with a five-year term of service.
To learn more about Halstead-GPU, contact Preston Smith, director of research services and support for ITaP, psmith@purdue.edu or 49-49729.