CULA Library Speeds GPUs on TSUBAME 2.0 Super

Print Friendly, PDF & Email

This week EM Photonics announced that the Tokyo Institute of Technology is using the company’s GPU-accelerated linear algebra library CULA tools for the TSUBAME 2.0 supercomputer. As home to Japan’s first petaflop performance supercomputer, Tokyo Tech is considered a leader in the deployment of GPUs at scale, with the only TSUBAME 2.0 being the only supercomputer ranking in the upper ranks of the TOP500 and GREEN500 supercomputer lists.

The majority of the achievable FLOPS in TSUBAME 2.0 is due to the power of the GPUs, so it is essential that we provide as comprehensive a software stack to utilize them to their fullest potential as possible,” indicated Professor Satoshi Matsuoka, the TSUBAME 2.0 leader. “CULA will be an extremely valuable part of the portfolio, allowing our scientists to conduct large scale simulations at unprecedented speeds,” added Prof. Matsuoka.

CULA is EM Photonics’ GPU-accelerated linear algebra library. According to a company spokesperson, CULA is based upon the LAPACK interface and has solvers capable of up to 7x speedups over competitor’s tuned linear algebra routines.