Over at the Nvidia Blog, George Millington writes that, the fourth consecutive year, the Nvidia Tesla Accelerated Computing Platform helped set new milestones in the Asia Student Supercomputer Challenge, the world’s largest supercomputer competition.
Some 175 teams from universities in North America, South America, Africa, Asia, Europe and Oceania participated in ASC16. Of these, 16 advanced to the final round held last week at Huazhong University of Science and Technology in Wuhan, China. Armed with their custom-built systems, finalists competed on six different supercomputing application benchmarks within a 3,000W system power limit. The benchmarks included the surface wave numerical model, MASNUM; the material simulation software, ABINIT; the High Performance Conjugate Gradients benchmark; and ABySS, a de novo, parallel sequence assembler. Teams also had to train a deep neural network for speech recognition. The final test was LINPACK, the industry benchmark used to measure the performance of the world’s most powerful supercomputers, like the Titan system at Oak Ridge National Laboratory. Thanks to our Tesla K80 GPU accelerators, Huazhong University beat out the field of competitors, grabbing the highest overall score.
The team from Zhejiang University used a system with eight NVIDIA Tesla K80 GPUs to establish a new record on LINPACK.
“Clocking in at a remarkable 12.03 teraflops – 12 trillion floating-point operations per second – Zhejiang’s system overturned the previous record of 11.92 teraflops. Nanyang Technological University achieved that record at ASC15. Sun Yat-sen University set a record of 9.27 teraflops using Tesla K40 GPUs at ASC14. NVIDIA GPUs continue to help computer scientists, researchers and engineers around the world tackle massive computational challenges. Through competitions like ASC, they’re preparing new generations of experts to address tomorrow’s toughest problems.”