During the recent decades, the US was able to attract the best scientists from all over the world in every field, including computing, by providing them with competitive salaries and benefits. But now research budgets are being cut, and the former rate of progress cannot longer be maintained for the United States.
In 2013, the US will receive modest funding in their quest for exascale, while the European Commission, on the contrary, doubled their supercomputer spending to 1.2 billion Euro up to year 2020. (source)
The Japanese parliament also was investigating if spending for the “K” supercomputer had been excessive.
Insufficient funding means inability to recruit the best talents. Now, the “human capital” that, figuratively speaking, knows no borders can flow to other countries.
So, with the current funding situation, which countries are likely to make rapid progress in the supercomputing field? I believe the US and Japan have slowed down. We shall not also forget an unpleasant situation when IBM withdrew from the Blue Waters project in August 2011.
India has lots of brilliant computer scientists, but they successfully work in various countries (much of them in the US), and will not return back home unless the government of India creates advantageous conditions for them — of which we don’t have much news. If the US has a funding crisis, they might relocate, but not necessarily back to India.
Russia can make big machines (and has money for it), but the country is not developing its own CPUs or other components.
The European Union, despite abundant funding, has lost its momentum. It would be nice to see the EU research community presenting a new CPU microarchitecture, but it appears that currently Europe is mostly relying on hardware developed in the US and manufactured in Asia.
Now let’s consider China who is going to join the club of leaders. Compared to its peers, China looks like a young and vigorous player. The Chinese are brave and bold, and are not afraid to innovate.
What happens if they bring to market a new CPU architecture which is suitable for various workloads, ranging from generic to compute-intensive to data-intensive? I think it’s the worst nightmare of the Western world computer industry. Combined with China’s ability to cheaply manufacture those components, the appearance of such a CPU could, with time, shrink the market share of traditional CPU makers like Intel and AMD. That would lead to reduced R&D spending for those companies, eventually making their products less competitive. Finally, that would result in redivision of the global computer market — not just high-performance computing, but the richer and higher-volume server, desktop and notebook segments.
Of course, having a new CPU architecture that would suit both high-performance computing and desktop markets is more like a dream for now. Besides, China doesn’t have access to modern semiconductor technology: it’s leading company, SMIC, uses only 45nm process, while Intel is already using 22nm.
But China is moving in giant steps. They already have nice 16-core locally made CPUs. The CPU is called ShenWei, and Dr. Hung-Sheng Tsao explains it’s similar to Alpha 21164.
Some may think it is just a Chinese “copy”, a random success. But China has more talents to offer. The ICube Corporation, “founded by Silicon Valley veterans”, as the website tells, developed a mobile CPU with a completely new instruction set architecture (ISA).
I am sure that with time China can come up with new and original server and desktop CPUs, if they find it necessary.
Unfortunately, in case of the redivision of the global CPU market (if it ever happens), there will be lay-offs at current market leaders such as Intel or AMD. This will hurt common people and their families. But that’s what globalisation is all about. Salaries in China are currently an order of magnitude less than in the Western world, so I suppose that a certain equalisation would be fair. After all, they earned it by hard work!
