SoylentNews is people
SoylentNews
Submitted via IRC for guy_
Tencent chairman pledges to advance China chip industry after ZTE 'wake-up' call: reports
While the U.S. administration said on Friday it had reached a deal to put ZTE back in business after the company pays a $1.3 billion fine and makes management changes, the plan has run into resistance in Congress, indicating ZTE was still far from out of the woods. Also, ZTE is yet to confirm the deal.
"The recent ZTE incident made everyone more clearly realize that however advanced one may be in mobile payment, without the mobile, the chips and the operating system, you still cannot compete," Chinese media reports cited Tecent's Pony Ma as saying at a forum in Shenzhen on Saturday.
[...] Tencent is looking into ways it could help advance China's domestic chip industry, which could include leveraging its huge data demand to urge domestic chip suppliers to come up with better solutions, or using its WeChat platform to support application developments based on Chinese chips, Ma said.
"It would probably be better if we could get in to support semiconductor R&D, but that is perhaps admittedly not our strong suit and may need the help of others in the supply chain."
(Score: 3, Interesting) by MichaelDavidCrawford on Monday May 28 2018, @08:39PM (4 children)
A former employer of mine owns the US rights to Zeni 4. It was developed in the PRC and competes directly with Cadence and Mentor Graphics.
I chatted with a chip designer about it, he said it was just as good as its US competitors but is not widely known outside of China.
In 1981 a UCD physics grad student told me that shrinking US chip feature sizes had the unintended effect of stimulating a real honest to goodness Soviet chip industry.
This because they were no longer able to reverse the smaller chips.
Just about the only thing the US could do at this point is to build fabs for smaller sizes faster than the PRC can. I expect that will work just fine for five or ten years.
Yes I Have No Bananas. [gofundme.com]
(Score: 2, Insightful) by Anonymous Coward on Monday May 28 2018, @08:45PM
Just about the only thing the US could do at this point is to build fabs for smaller sizes faster than the PRC can. I expect that will work just fine for five or ten years.
Or just long enough for the current shortsighted administration to think that they won their trade war with China, while all they really did was wake up the Chinese to their biggest weakness. so they can start addressing it. Well done...
(Score: 2) by Snotnose on Monday May 28 2018, @10:46PM (2 children)
Nah. The Chinese don't really care about IP. How to make speedy chips is in the literature and patents, there are 3-4 times as many Chinese as Americans, and the Chinese are not complacent. In 10 years they'll be making chips that are 5 years behind Intel/AMD. In 20 they'll be ahead, both in design and fab. For once I'm glad I expect to be dead in 20 years, I suspect this country is in for a world of hurt.
As an added bonus they won't bother to emulate the x86 architecture, they'll come up with their own. Followed by their own version of Windows, probably based off Linux.
When the dust settled America realized it was saved by a porn star.
(Score: 2) by MichaelDavidCrawford on Monday May 28 2018, @11:18PM
IIRC they already have one but I don't think it's widely used
Yet.
Yes I Have No Bananas. [gofundme.com]
(Score: 4, Insightful) by takyon on Tuesday May 29 2018, @12:51AM
This timeline doesn't make much sense to me...
1. Intel has been stagnating at the 14 nm node for years due to low yield for 10 nm.
2. Pretty much every fab is now dependent on a third party, ASML, delivering the extreme ultraviolet lithography tools that they need to scale further.
3. What does it mean to be 5 years behind Intel? Haswell chips were released starting in 2013 and the performance gains have been modest since then, and some years have seen no gain.
4. We're running up towards the atomic limits. In ten years time we will probably be around the 0.5 to 3 nm mark. The only real way forward for classical computing is some kind of massive boost to 100+ gigahertz or terahertz clock rates by using a new material or design that doesn't get as hot, or vertical stacking, again with the need to address heat issues.
5. SMIC [wikipedia.org], at as low as 28 nm, is not so far behind Intel and AMD, and could be considered "5 years behind". Remember, AMD was making 28 nm chips right up until 2017 when it skipped to 14 nm with Ryzen. That's another thing China's fabs could try: acquiring the IP they need, and then skipping nodes.
6. It was a homegrown manycore chip [wikipedia.org] that put China's Sunway TaihuLight [wikipedia.org] at #1 supercomputer in the world spot. It uses a RISC architecture instead of x86.
You're too tied to the fate and fortune of the U.S. Maybe retire overseas?
[SIG] 10/28/2017: Soylent Upgrade v14 [soylentnews.org]