SUPERCOMPUTER VENDOR Cray will bung Nvidia Tesla GPGPU accelerator boards into its next generation Cascade cluster.
Back at the International Supercomputing Conference, Cray was unveiled as the big name partner for Intel's Xeon Phi accelerator in its high-profile Cascade cluster. Now Cray has come out and said that it will offer Nvidia Tesla accelerator boards alongside the Xeon Phi option.
Cray said that it will be offering Tesla boards based on the current generation Kepler architecture. The firm said it will continue to offer Xeon Phi as an option, but given Nvidia's reputation in the HPC market, customers might want to go for the more established brand.
Peg Williams, Cray's SVP of HPC systems said, "Adaptive Supercomputing is about building a robust hardware and software environment that ultimately supports giving our customers choices. Adding the Nvidia Tesla Kepler-based GPUs, which are designed for computationally intensive HPC environments, into future versions of our Cascade system will give our customers the flexibility to choose from a variety of powerful accelerator options."
Nvidia launched its K10 Tesla board earlier this year and said its double precision floating point K20 board will be available by November. Intel's Xeon Phi is expected to tip up around the same time and it might become a battle over ease of software development rather than outright speed.
Cray said that its Cascade clusters will be widely available in 2013. µ