GPU Scheduling

GPU Scheduling with PBS Professional


PBS Professional supports both basic and advanced GPU scheduling. Basic scheduling consists of prioritizing jobs based on site policies, controlling access to nodes with GPUs, ensuring that GPUs are not over-subscribed, and tracking use of GPUs in accounting logs. Configuring PBS to perform basic scheduling of GPUs is relatively simple, and only requires defining and configuring a single custom resource to represent the number of GPUs on each node.

With the advent of the Graphical Processing Unit (GPU) as a general-purpose computing unit, more and more high-performance computing (HPC) users are moving toward GPU-based clusters to run their scientific and engineering applications. This model allows users to use a CPU and GPU together in a heterogeneous computing model, where the sequential part of the application runs on the CPU and the computationally intensive part runs on the GPU. By exploiting the massive parallelism in GPUs, users can run applications almost forty percent faster, compared to the traditional CPU-based mode.

Although basic GPU scheduling will meet the needs of 95 percent of customers, PBS Professional also supports more advanced GPU scheduling: the ability for a job to separately allocate (request and/or identify) each individual GPU on a node. This capability is useful for sharing a single node among multiple jobs, where each job requires its own GPUs.

Read about Altair and NVIDIA GPU performance testing.

In addition to support for GPUs and other accelerators, PBS Professional also supports the Intel® Xeon Phi™ coprocessor.

For an in-depth view, download our technical paper on "Scheduling Jobs onto NVIDIA Tesla GPU Computing Processors using PBS Professional."

"PBS Professional offers the best scalability in the industry with a rich set of technical features, advanced reservation capabilities and proven success in Japan's National Research Grid Initiative, giving us great confidence in its ability to manage the exceptional architecture of TSUBAME 2.0."

– Fumihiko Hisamitsu, General Manager, IT Hardware Operations Unit, HPC Division, NEC Corporation

Bookmark and Share

Subscribe to join our Newsletter
Learn about product training, news, events and more.