China aims to redraw the petaflop graph with 100 Pflops in 2015 China shocked the supercomputing world in late 2010 with a chart-topping 2.56 Petaflop/sec Tianhe-1A. It was a surprising system on several levels: 1) it topped the incumbent number one box (Oak Ridge’s Jaguar) by almost 50%; 2) it was the first (and, so far, only) hybrid commodity combo of Intel Xeon and NVIDIA Tesla processors …
They rely on western technology to make this happen. NVIDIA is a Californian company.
The UK has ARM - very low power (heat and computational ability). We had INMOS, but sold it to the French who effectively abandoned the Transputer.
The US has Intel, NVIDIA, AMD, IBM and a host of startup companies.
Israel has companies we know little about, innovating away.
China doesn't innovate. India doesn't innovate. Now they both are cash rich perhaps they will, but it's hard to shake off decades of risk averse business practice.
The article specifically mentioned there is the option of using their home grown silicon and the suggestion that the Chinese don't innovate is borderline insulting. I'm not keen on some bits of China, what with my wife being from Taiwan. But when it comes to their ambition, you cannot simply say that they don't innovate.
I reckon for every degree qualified engineer in the UK or US, there is at least one in China, many of whom have most likely studied at a western university and are at least as smart as we are, but with a much better work ethic (9-5 being a half day for them!) And they can spend all day looking at what the likes of Intel, NVidia and IBM produce then working out how to make it a bit faster or use a bit less power and before you know it, they'll have done 100PFlop and will be getting to Exascale computing!
No intention to insult. I don't see many Chinese companies who's products are truly innovative. The Koreans have taken on large Japanese companies (e.g. Samsung taking on Sony / Panasonic in TVs).
I'd rate Chinese global ambitions higher than the USA (as is evidenced by their significant investment in Africa to ensure they have the raw materials they required to be able to continue to be the leading manufacturing nation in the world).
Israel innovates in the areas of military equipment (inc. avionics / radar) and silicon, not in mass manufacturing.
The CPUs were American-made, the GPUs were American-made; the operating system, file system and job scheduler are off-the-shelf open source software. So how exactly was Tianhe-I Chinese innovation?
Compare that with the K computer, in which every single significant hardware component (and some of the software stack) was developed in Japan.
This post has been deleted by its author
100 pfs in 2015 doesn't seem like such a big deal. For example, the ORNL Titan system can theoretically be upgraded to 50 pfs - it will be on-line in late 2012 at about 20 pfs and with more budget may be upgraded to 30 pfs. So, with the huge focus by US HPC component players like Intel and Nvidia, is a mere doubling from 50 now to 100 such a stretch? Of course, power and reliability will be gate keepers. But, should we bet that a "black" 100+ pfs system won't exist at some US government location in 2014?
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2024
