from the my-processor-can-beat-up-your-processor dept.
According to VR World and HPCwire, the U.S. government has blacklisted "high technology" shipments to the National Supercomputing Center Changsha (NSCC-CS), National Supercomputing Center Guangzhou (NSCC-GZ), National Supercomputing Center Tianjin (NSCC-TJ), and the National University of Defense Technology (NUDT) in China. This effectively means that these major supercomputing facilities can no longer purchase Intel Xeon chips. Tianhe-2, the world's fastest supercomputer since June 2013 according to Top500, is located at the NUDT in Guangzhou and uses a total of 32,000 Intel Xeon and 48,000 Xeon Phi chips.
The main claim of the Bureau of Industry and Security's End-User Review Committee (ERC) is that NUDT, which used US-manufactured parts to produce the Tianhe-1A and Tianhe-2 supercomputers located at the National Supercomputing Centers in Changsha, Guangzhou, and Tianjin, is believed to be engaged in activities related to nuclear explosives.
The U.S. also uses supercomputers for nuclear weapons research.
The news coincides with the Intel Developer Forum 2015 in Shenzhen, China, at which the company announced new Braswell, SoFIA, and Cherry Trail chips, among other products. VR World speculates that the move could cost Intel $1 billion on lost Broadwell-EP Xeon E5v4 sales and accelerate the development of homegrown Chinese processors.
The Platform reports that CPU export restrictions to Chinese supercomputing centers may have backfired. Tianhe-2 has remained the world's top supercomputer for the last five iterations of the TOP500 list using a heterogeneous architecture that mixes Intel's Xeon and Xeon Phi chips. Tianhe-2 will likely be upgraded to Tianhe-2A within the next year (rather than by the end of 2015 as originally planned), nearly doubling its peak performance from 54.9 petaflops to around 100 petaflops, while barely raising peak power usage. However, instead of using a new Intel Xeon Phi chip, a homegrown "China Accelerator" and novel architecture will be used.
A few details about the accelerator are known:
Unlike other [digital signal processor (DSP)] efforts that were aimed at snapping into supercomputing systems, this one is not a 32-bit part, but is capable of supporting 64-bit and further, it can also support both single (as others do) and double-precision. As seen below, the performance for both single and double precision is worth remarking upon (around 2.4 single, 4.8 double teraflops for one card) in a rather tiny power envelope. It will support high bandwidth memory as well as PCIe 3.0. In other words, it gives GPUs and Xeon Phi a run for the money—but the big question has far less to do with hardware capability and more to do with how the team at NUDT will be able to build out the required software stack to support applications that can gobble millions of cores on what is already by far the most core-dense machine on the planet.
In the name of national security, China is restricting export of certain drones and computers:
From August 15, manufacturers of certain powerful drones and computers will have to give technical details to the authorities to obtain a licence prior to export, Xinhua news agency says.
[...] In the first five months of 2015, China exported some 160,000 civilian drones, a jump of 70 per cent year-on-year, worth more than $120 million, the official China Daily newspaper reported in July.
[...] The tightening of regulations comes two weeks after an incident in disputed Kashmir in which the Pakistani army claimed to have shot down an Indian "spy drone", reportedly Chinese-made.
China is also likely tightening controls on exports of powerful computers as it looks to maintain its edge in the global supercomputer battle long dominated by US-Japanese rivalry.
Starting August 15th, drone and supercomputer manufacturers will have to present technical details to the authorities in order to get a license to export.
Chinese-designed "Dhyana" x86 processors based on AMD's Zen microarchitecture are beginning to surface from Chinese chip producer Hygon. The processors come as the fruit of AMD's x86 IP licensing agreements with its China-based partners and break the decades-long stranglehold on x86 held by the triumvirate of Intel, AMD and VIA Technologies. Details are also emerging that outline how AMD has managed to stay within the boundaries of the x86 licensing agreements but still allow Chinese-controlled interests to design and sell processors based on the Zen design.
AMD's official statements indicate the company does not sell its final chip designs to its China-based partners. Instead, AMD allows them to design their own processors tailored for the Chinese server market. But the China-produced Hygon "Dhyana" processors are so similar to AMD's EPYC processors that Linux kernel developers have listed vendor IDs and family series numbers as the only difference. In fact, Linux maintainers have simply ported over the EPYC support codes to the Dhyana processor and note that they have successfully run the same patches on AMD's EPYC processors, implying there is little to no differentiation between the chips.
The new chips are surfacing against the backdrop of the trade war between the US and China that could escalate quickly, likely reinforcing China's long-held opinion that a lack of native processor production could be a strategic liability. Today's wars are won with chips, and their strategic importance certainly isn't lost on those in the halls of power. In fact, the Obama administration blocked Intel from selling Xeon processors to China in 2015 over concerns the chips were fueling the country's nuclear programs, and subsequent actions by the US have largely prevented China from achieving the technical know-how and equipment to develop its own chips through acquisitions and mergers.
That makes it even more surprising that AMD has managed to establish a franchise that allows Chinese processor vendors to develop and sell x86 processors in spite of US regulations and the licensing restrictions with Intel, but now more information is coming to light about how AMD pulled off the feat.
Related: Intel Launches New Chips in China as US Bans Sales to Supercomputing Centers
Intel Hints at Patent Fight With Microsoft and Qualcomm Over x86 Emulation
Data Centers Consider Intel's Rivals
Tencent Chairman Pledges to Advance China Chip Industry After ZTE "Wake-Up" Call
Chinese companies are manufacturing chips nearly identical to AMD's Epyc server CPUs, using two joint ventures with AMD. This move comes after the US blacklisted certain Chinese supercomputing centers in 2015 in an attempt to prevent them from using Intel Xeon chips, and more recently, Chinese telecom equipment maker ZTE was banned from buying components from US companies. China's Sunway TaihuLight supercomputer (formerly #1 on the TOP500 list) also uses domestically designed Sunway SW26010 manycore chips.
China isn't eager to embrace another American chipmaker like AMD. In response, AMD established two joint ventures with Chinese holding company THATIC -- one with Chengdu Haiguang Microelectronics Technology (CHMT), and another with Haiguang IC Design, also known as Hygon.
AMD owns a majority stake in CHMT, which ensures that its IP isn't transferred to THATIC. THATIC owns a majority stake in Hygon, which licenses AMD's IP from CHMT. Hygon designs the chips, and CHMT produces the chips through a suitable foundry and then sends them back to Hygon for packaging, marketing, and sales.
This arrangement seemingly placates American and Chinese regulators -- AMD's IP isn't being passed to a Chinese company, and a Chinese chipmaker gains access to superior data center CPU designs. AMD generates less revenues through these JVs than it would through direct sales, but it still gains a foothold in China's massive data center market. But more importantly, this move could wound Intel.
Good luck maintaining control of your "IP". As for the pain?
Many big companies, including Microsoft and Baidu, started installing AMD's cheaper chips in their data centers. In a meeting with Nomura Instinet analyst Romit Shah in June, then-CEO Brian Krzanich admitted that AMD was gaining ground, and Intel was trying to prevent it from gaining a "15% to 20%" share of the data center market. That admission was stunning, since Intel traditionally controlled more than 99% of the data center market with its Xeon chips. Intel's data center group grew its revenues by 11% to $19.1 billion last year, and accounted for 30% of its top line. Epyc was already a thorn in Intel's side, but AMD's sponsorship of Chinese clones could throttle its sales in mainland China, which accounted for 24% of its sales last year. Its total sales in the region only rose 6% in 2017, compared to 20% growth in 2016.