Comments on: China Unveils Design of New Supercomputer Center in Changsha http://insidehpc.com/2010/12/31/china-unveils-design-of-new-supercomputer-center-in-changsha/ HPC News Without the Noise for Supercomputing Professionals | insideHPC Sun, 19 May 2013 10:46:54 +0000 hourly 1 http://wordpress.org/?v=3.1.1 By: Architekk http://insidehpc.com/2010/12/31/china-unveils-design-of-new-supercomputer-center-in-changsha/#comment-288915 Architekk Sat, 01 Jan 2011 14:46:10 +0000 http://insidehpc.com/?p=16329#comment-288915 It is not, ultimately, the fancy buildings, nor even the massive array of commodity components that make the potential of China's supercomputer impressive - it is what people DO with it that matters. Anyone can build buildings or assemble commodity components. This is where our own intellectual bankruptcy may cause our demise. There exists (existed) an extreme prejudice against heterogenous GPU computers (as merely gamer tech) that has not yet been discredited even in light of China's advancements. The primitive, existing computational fabrics and the software that runs on supercomputers are the real culprit. While known for decades, only now is this being acknowledged (and begrudgingly, at that). The hubris that thinks exascale computation lies on the current technological trajectory is probably wrong. That hubris is successfully keeping progress toward exascale computing at bay. It is not, ultimately, the fancy buildings, nor even the massive array of commodity components that make the potential of China’s supercomputer impressive – it is what people DO with it that matters. Anyone can build buildings or assemble commodity components. This is where our own intellectual bankruptcy may cause our demise. There exists (existed) an extreme prejudice against heterogenous GPU computers (as merely gamer tech) that has not yet been discredited even in light of China’s advancements. The primitive, existing computational fabrics and the software that runs on supercomputers are the real culprit. While known for decades, only now is this being acknowledged (and begrudgingly, at that).

The hubris that thinks exascale computation lies on the current technological trajectory is probably wrong. That hubris is successfully keeping progress toward exascale computing at bay.

]]>