James Vincent, Circuit Breaker, Chinese supercomputer is the world’s fastest- and without using US chips, here. Starting to evaluate platforms for NIMo. In terms of selling automation efficiency to banks we might be able to get the Sparc M7 cloud. or one of these guys in the game. We can scale up NIMo to simulate all US assets and liabilities (call it 15 Tn USD all in ). Then you can back out from the simulation where individual banks did better or worse than expected. You could work with the Fed on this. I think this is interesting from the 3 Rs they taught in school back in the day Research, Regulation, and Revenue. Nobody wants to steer a rocket ship with a joystick in 2016 I’d imagine.
The TaihuLight is comprised of some 41,000 chips, each with 260 processor cores. This makes for a total of 10.65 million cores, compared to the 560,000 cores in America’s top machine. In terms of memory, it’s relatively light on its feet, with just 1.3 petabytes used for the entire machine. (By comparison, the much less powerful 10-petaflop K supercomputer uses 1.4 petabytes of RAM.) This means it’s unusually energy efficient, drawing just 15.3 megawatts of power — less than the 17.8 megawatts used by the 33-petaflop Tianhe-2.
More significantly than its specs, though, is the fact that the TaihuLight is built from Chinese semiconductors. “It’s not based on an existing architecture. They built it themselves,” Jack Dongarra, a professor at the University of Tennessee and creator of the measurement system used to rank the world’s supercomputers, told Bloomberg. “This is a system that has Chinese processors.”
SC16, Jack Dongarra Shares his Assessment of World’s Newest No. 1 Supercomputer, here.
Tianhe-2A was supposed to be upgraded with Intel’s Knights Landing processors, but last year the U.S. Department of Commerce blocked the export of Intel technology to some parts of China.
When the Commerce Department blocked the exports, China invested heavily in HPC research and development and they are replacing Intel with their own designs. This system is based on a Chinese processor with 260 cores. For comparison Intel’s Knights Landing has 72 cores. Both of the processors have about the same cycle time – 1.45 gigahertz for the Chinese processor and 1.40 GHz for Knights Landing. It means that China has continued in leapfrogging the U.S. by a considerable amount.
Stacey Higginbotham, MIT Technology Review, Supercomputer Powered by Mobile Chips Suggests New Threat to Intel, here. It is interesting that with NIMo as the focus I really do not care about general purpose SPEC style benchmarks or linear algebra. You will never willingly take a non compulsory miss in L2 while execution a stack of a quadrillion FLOPS. I just want expression evaluation at peak – probably could go to GPU.
Fujitsu said this week that it will use ARM-based processors to build a successor to an existing Japanese supercomputer called Project K. Fujitsu is building the “Post-K” machine for the Riken Advanced Institute for Computational Science, which plans to use it for biomedical, climate, and energy research. The computer is slated to be installed and turned on in 2020.
James Renders, HPC Wire, What Knights Landing is Not, here. Runs emacs.
One of the first things I did when I initially logged on to a Knights Landing machine was to type in “yum install emacs.” I’m sure that whoever built that emacs binary had never heard of Knights Landing. It worked and I was happy to have the power of emacs so as to no be slowed by the primitive “vi.” I am so happy that software just runs, without a recompilation needed. No need to do something weird with Knights Landing to use it with your favorite software. It’s just like any other processor from Intel in that respect! It can run anything you would expect a processor to run: C, C++, Fortran, Python, and much more. It really is a full processor!