Graphics processing units (GPUs), which handle more streams of calculations than central processing units (CPUs), became integral to supercomputer design in the late 2000s, significantly boosting computational speeds for scientific and artificial intelligence workloads. The Japanese supercomputer TSUBAME 1.2, launched in 2008, was the first supercomputer to integrate GPUs in its design. Since then, powerful supercomputers using GPUs in their architecture have achieved high speeds, measured in TFLOPS, such as the Tianhe-1A supercomputer, launched in 2010 in China (2,556 TFLOPS), and the Titan, launched in the United States in 2012 (17,600 TFLOPS).
verifiedCite
While every effort has been made to follow citation style rules, there may be some discrepancies.
Please refer to the appropriate style manual or other sources if you have any questions.
Select Citation Style