This article is more than 1 year old
Pimp my HPC cluster: Luscious cores, GPU gravy, optimised apps...
But DON'T FORGET to defang angry end users
HPC blog You’ve pimped out your fancy new HPC cluster with the latest technology, each node jammed to the gills with shiny new CPUs, loads of memory, and maybe even a double brace of GPUs or other accelerators.
The installation work is done, and users are getting their first shot at the system. The performance boost on the new box is so profound that you’re halfway expecting a mob of ecstatic users to hoist you on their shoulders and parade you around the building. While you’re dreaming away in your office, a mob of users are gathering….but they’re carrying weapons crudely fashioned from their desktops.
As you’re being beaten about the head by a postdoc wielding a used toner cartridge, you find yourself wondering if this situation could have been avoided.
Parallelising user code and, better yet, helping users to better parallelise their own code, is the answer. Code that’s well-behaved and scales well on smaller clusters can have a nervous breakdown on larger systems.
This is where Allinea comes in. It claims its tool suite makes it faster and easier to develop, debug and optimise code that will scale from the smallest closet cluster to the largest supercomputers in the world.
At GTC’14, I spent a few minutes with Allinea’s Beau Paisley. With real-world experience in large-scale applied physics, climate modelling, and image processing, Paisley knows what he’s talking about when it comes to making big systems perform.
In the video, we talk about the challenges of parallelising code and making the jump from traditional clustering to hybrid systems. It’s a brief but meaty conversation that’s worth a look.
The firm also hosted a couple of sessions at GTC14 – here's a look at the abstracts:
You can find these Allinea sessions here. And if you’re looking for even more details (aren't you ever satisfied?), you can check out Allinea here. ®
Biting the hand that feeds IT
