Amazon Web Services is adding more raw horsepower to its high performance computing instances with the launch of its Cluster Compute Eight Extra Large, a service that ranked No. 42 on the Top 500 list of supercomputers. The new instance can hit 240.09 teraflops.
In a blog post, AWS outlined the following:
- The new HPC instance includes 2 Intel Xeon processors with 8 cores each. The Cluster Compute Eight Extra Large also has 60.5 GB of RAM and 3.37 TB of instance storage.
- AWS cut the price of current Cluster Compute instances to $1.30 an hour.
- The operating systems in the HPC instance can be either Amazon’s Linux or Windows 2008 R2.
- An array of 290 Cluster Compute Eight Extra Large instances would yield a 63.7 teraflops system for less than $1,000 an hour.
- Reserved instances are supported.
The company stepped up its high performance computing (HPC) game last year with two instances designed things like analyzing the human genome and processing images. The HPC services, part of Amazon’s EC2 compute clusters, were designed for low latency. The Cluster Compute Eight Extra Large now features more powerful hardware.
As noted in previous supercomputer stories, there isn’t a massive amount of revenue in HPC but there’s a certain vanity appeal that can drive business elsewhere. Amazon Web Services is making sure it has its HPC services covered and rest assured it will try to move up the supercomputer rankings going forward. Here’s a look at spot pricing for instances. HPC prices are at bottom.
