Amazon Cluster GPU: A New Word in the World of Cloud Hosting

Amazon announced the Cluster GPU instance - a HPC instance with two NVIDIA Tesla “Fermi” M2050 GPU - for fans of really “fast” computing.

NVIDIA, has long been promoting its idea of ​​transferring heavy computing from the main processor to the GPU, the benefit of modern GPU power is not inferior - and sometimes exceeds the standard CPU. There is iron, there is a platform for development, there was (at least, I did not know) hosting where all this could be launched.

And here one of the main players in the Cloud hosting market - Amazon announced the availability of the Amazon Cluster GPU instance. According to the link, the instance specification is as follows:
* 22 GB of memory
* 33.5 EC2 Compute Units (2 x Intel Xeon X5570, quad-core “Nehalem” architecture)
* 2 x NVIDIA Tesla “Fermi” M2050 GPUs
* 1690 GB of instance storage
* 64-bit platform
* I / O Performance: Very High (10 Gigabit Ethernet)
* API name: cg1.4xlarge

Promised peak performance: one trillion double-precision FLOPS (specially copied from the announcement not to be mistaken in zeros :))
')
Issue price: $ 2.10 per hour