Penguin Computing, experts in high performance computing (HPC) solutions, today announced the immediate availability of PODtool, an interface that makes it easy to offload excess workload to Penguin’s on-demand environment Penguin Computing on Demand (POD).
With PODtool HPC in the cloud users can now:
- Easily and quickly tap Penguin Computing’s compute resources from any Linux system without having to allocate and de-allocate server instances
- Integrate ‘spill over’ job submission into an existing workflow and avoid delayed processing of workflows due to internal resource constraints
- Automatically transfer data and analysis results from and to POD
PODtool is a web-service based command line interface that drastically simplifies the management of compute jobs on Penguin Computing’s HPC on-demand offering POD. Compute job scripts for SGE and TORQUE can be submitted and will be implicitly routed to the corresponding execution environment. Data staging commands allow for easy data transfer of job input and output data to and from POD. Job control and monitoring features enable users to fully control compute jobs they are authorized to manage. PODtool’s user account mapping capabilities facilitate the management of POD compute jobs from any Linux system.
The development of PODtool has been driven by feedback provided by POD customers looking for easy access to Penguin Computing’s on-demand resources. The Information Management Systems & Services (IMSS) organization at the California Institute of Technology (Caltech) has been working with Penguin on making computational POD resources available to Caltech’s scientific community through podsh: “Our researchers and students make intensive use of the numerical computing environment Matlab. We have been increasingly ‘maxing out’ out our internal compute resources with Matlab jobs. With a limited budget for capital expenditures we started looking at making compute cycles available on-demand to the Caltech research community. PODtool was a real differentiator Penguin Computing offered, as it has been an absolute requirement for us to make it easy for our users to use remote resources, to lower our support burden”, says Chad Schmutzer, Associate Director of Academic Computing Solutions.
“Penguin Computing has always provided HPC solutions that work reliably and are easy to use. Just as Scyld ClusterWare simplifies the use and management of a Linux HPC cluster, PODtool makes it easy to tap into additional compute resources to spill over computational workload”, says Tom Coull, Senior Vice President and General Manager for Software and Services at Penguin Computing.