The Incredible Power of the Amazon EC2 Cluster GPU Instances

I'm one of the developers of CUDA at NVIDIA, so if anyone has any questions about that or GPUs in general that they'd like answered, feel free to post it here and I'll do my best.<p>However, answers might be a bit delayed--I'm at SC10 this week (or if you're at SC10, I'm an easy person to find at the NVIDIA booth).

This is nice. I work at a University and I spoke to the guy who runs our High Performance Computing system about Amazon EC2 the other day. He made a good point which is worth noting:<p>You can bring up a lot of EC2 instances to run large jobs in parallel and get a lot of CPU horsepower, but the main drawback of EC2 is the network. Most of the jobs that run on our local "super computer" involve processing terrabytes of data. Transferring terrabytes of data to the cloud is painful, and will continue to be painful for a long time to come. I suspect even transferring terrabytes of data between EC2 instances wont be smooth either.<p>Still, EC2 is great and I love how fast they're bringing out new features.

javacl, <a href="http://code.google.com/p/javacl/" rel="nofollow">http://code.google.com/p/javacl/</a><p>idiomatic Clojure wrapper from the developer of Aleph - <a href="https://github.com/ztellman/calx" rel="nofollow">https://github.com/ztellman/calx</a><p>I have to say, running Clojure on instances like these for a couple hours at a time to get a sense of what Clojure offers in terms of concurrency and parallelism on a 8-core machine with gobs of RAM is great fun - <a href="http://dosync.posterous.com/clojure-multi-core-amazon-cluster-compute-lat" rel="nofollow">http://dosync.posterous.com/clojure-multi-core-amazon-cluste...</a>.<p>It's the kind of computing excitement I imagine Lisp Machine users had.

Interesting move. We just bought two Tesla Cards for a university project so I know how much people could save by just prototyping on a "small" card that can do ?OpenCL? and then using a Quadruple Extra Large instance at $2.10 per hour for the actual computation instead of buying a 5 digit Euro workstation

This offering blows away the few people that were already trying to compete in the GPU cloud space simply by being AWS.<p>Our research is just moving into GPU-based processing, and we can probably adapt our current EC2 based framework to work with this relatively easily.

GPU Programming with PyCUDA:<p><a href="http://pdf.cx/6gc94" rel="nofollow">http://pdf.cx/6gc94</a><p><a href="http://mathema.tician.de/software/pycuda" rel="nofollow">http://mathema.tician.de/software/pycuda</a>

<i>Tianhe-1A with a peak performance of 4.701 PetaFLOPS</i><p>Nit: Linpack is an astoundingly easy benchmark to optimize, and they only attained 2.57 Pflop/s there. Most real science runs at much lower efficiency than Linpack (often more than an order of magnitude), primarily due to architectural reasons, so the theoretical peak number is even less meaningful than Linpack.

If you are looking at writing gpu code checkout <a href="http://www.tidepowerd.com/" rel="nofollow">http://www.tidepowerd.com/</a> startup in this area, just released their first beta, a .net gpu library/compiler<p>edit: gnu-&#62;gpu, for a harsh downvote, iPad auto correct :)

Also interesting is that Amazon's CTO takes time out to post to HN.

I predict Amazon's rank in the top500 is about to improve.

That's a nigh-trump-card. Wow. AWS: say hello to massive purchases from researchers.

I am not qualified to do the calculations, but I wonder if these things can profitably generate bitcoins?

So, I just put together my task list for this week, distributed it and now this.<p>This. This day I've been waiting for.<p>There goes my task list. :\

for completeness, here's the official announcement from amazon.<p><a href="http://aws.amazon.com/about-aws/whats-new/2010/11/15/announcing-cluster-gpu-instances-for-amazon-ec2/" rel="nofollow">http://aws.amazon.com/about-aws/whats-new/2010/11/15/announc...</a>

How many of you dear readers are really going to need such processing power?