Nvidia and IBM have been sitting in a tree, K-I-S-S-I-N-G to ready the latest IBM iDataPlex servers – the first mainstream high performance computing systems built up from the nVidia Tesla 20-series GPUs.
Nvidia has been boasting how its GPUs are helping industries left right and centre. The Tesla M2050 is a GPU computing module which works with servers to boost performance, while also – says Nvidia – staying very reliable and keeping system monitoring and management tight and hassle free.
The company boasts that CUDA, which lets GPUs be used for supercomputing as well as visual computing tasks, has been adopted in over 350 universities and that 10 books have already been written on the subject by developers. It claims the CUDA toolkit has been downloaded over 200,000 times. All very well, but where does Nvidia see itself going with its apparently fantastique technology?
Nvidia has played down the use of Tesla and CUDA in its past clients. In a statement, the general manager for Tesla at Nvidia said that IBM’s adoption of Tesla for the HPC server line is the “most significant milestone” for Tesla so far. Is Nvidia trying to sidle up to IBM?