Chinese chip designers hope to topple Arm's Cortex-A76 with XiangShan RISC-V design The Institute of Computing Technology at the Chinese Academy of Sciences (ICT CAS) has showcased progress on a fully open-source processor, designed around the RISC-V architecture, which it hopes will offer competition for Arm parts at the performance end of the market. Developed from the opening of a GitHub repository to …
This whole Entity List thing and its Cold War connotations is likely to do us far more harm than good. Before we went all in on this during the last administration China was happy to be a global citizen, albeit a successful and self-centered one (just like everyone else....), a country happy to buy as well as sell. The price they paid for this is that we sought to blunt their competitiveness by sanctions -- we didn't like the way that Huawei was cleaning up in the communications sector so took action to try to destroy them, for example. This is an open challenge for them to replace us, a goal that any and every Chinese person can get behind (because wouldn't you do the same if it was your country that was being attacked?). We might win a few battles with our trade restriction nonsense but we will lose the war (....and we might well have already; I was reading a business news article just yesterday about the problem the Chinese have with too many dollars, they're essentially becoming worthless to them because they can't put them to work without risking revaluing their currency).
Incidentally, China is only an 'untrustworthy, authoritarian regime' because our media keeps on telling us so. Embedded in these ongoing comments about China and the Chinese is a form of implicit racism, the notion that Chinese aren't real people capable of doing real things without the guiding hand of the Great White Master (and, of course, doing it our way for our benefit). This is nonsense and we need to get used to seeing these people as both equals and competitors, not some kind of primitive ant-like beings that 'work for a handful of rice a day'. (BTW -- we used to say the same thing about the Japanese.)
Yeah I cant help think this is a different time and situation!
The USA was very concerned at Japan's rise and they also brought sanctions and various laws to try and hobble Japan's economy which for the most part worked! (google US Japan Trade War)
However China isnt Japan and China has a bigger population base a larger knowledge pool to tap into and a 'whatever I do what I want' attitude.
Plus being an authoritarian regime it can redirect R&D spending quicker than I can nuke microwaveable popcorn, will that produce tangible results overnight nope but given the talent/people that China can amass I personally arent betting on the USA as being the winner here (unfortunately)
"was happy to be a global citizen"
were they though? China in general doesn't give a rats ass about IP and in many industries has gained a virtual monopoly simply by undercutting any and all competition simply by subsidizing those industries very heavily. The western world has let this happen because in many cases they're happy to let heavily polluting industries reduce or disappear or because they gain more from the lower prices for base materials or components than they lost from losing the manufacturing. But it IS slowly destroying western economies and handing the keys to basically world domination China.
China (Since CCP) has never acted as just a "self centered global citizen". It's never been interested in being just a peer. Its been backstabbing and scheming all it can to get even and then ahead. This isn't about a racist view of the Chinese people. Far from it. "We" can see what they are capable of as a people. Both good AND bad. And "we" see what the current Chinese regime might do with that capability. And a lot of people think we might not want to let them do certain things unchecked. This doesn't boil down to "Chinese people bad". It boils down to "Chinese communist regime bad". It's history repeating and we let it happen.
China in general doesn't give a rats ass about IP
this whole "IP" thing is a blatant imperial way to (try to) tell other people what they can and can't do. Conceptually, intellectual things aren't property, because you can't take them away.
The real concept behind this is "commercial exclusivity" meaning that if you have invented something, you have a temporary exclusivity to sell stuff based on this. But it doesn't roll off the tongue as nicely. And accusing someone of violating your exclusivity doesn't sound as bad as accusing someone of stealing your intellectual property.
Basically, we're all standing on shoulders of giants, all our inventions are built on other people's inventions.
To assume that most recentish computer hardware contains in it's innards some sort of deliberately embedded/tinkered with code that makes it easier for agencies/baddies/whoever somewhere, to have a snoop around.
I'd have thought without having to sneak some noughty code etc. in by the backdoor, there are already plenty enough handy exploits or whatnot for use by ne'er-do-wells, ready and waiting, just due to regular happenstance/mis-cut-and-paste/fookups etc.
You mean some kind of processor like an Intel ME or an AMD PSP that runs encrypted binary code, that can never be audited, with full access to all hardware and RAM that can be accessed by the normal default CPU cores.
ref: https://libreboot.org/faq.html#intel
ref: https://libreboot.org/faq.html#amd
If China can produce a powerful CPU, without the equivalent of a ME or PSP, at a reasonable price they will gain a massive share of the market very fast.
(Then in 5, 10 or 15 years time once they have gained a large enough market share they could add the Chinese equivalent of the FISA core as I call them.)
Yes and no. Really depends on your use case. It should be faster than many microcontrollers and if you implement certain operations in HDL that you would otherwise implement in software, you can get some advantage over off the shelf CPUs.
I would say that this is more to test drive the design before actually committing it natively to silicon or just for fun.
I've worked with both hard and soft versions of this architecture and the FPGA -- soft -- versions are a lot slower than the hard ones. Its a trade off -- the hard processor supports a much faster clock but that doesn't necessarily translate to an overall higher performance, it depends on the application -- my projects have invariably been dedicated embedded applications rather than general purpose computing applications. What the FPGA does give you -- it depends on the FPGA, of course -- is a lot of flexibility in being able to size internal memory and the processor caches to optimize performance (which is often offset by the FPGA's designer parsimony with I/O pins -- they do not like wide address and data buses, they have other plans for those pins!).
Incidentally, RISC-V is a standardization of a very common type of RISC processor, a type that's been around for decades starting with the MIPS. The ARM should just be another type of three address RISC but I believe its differences are fewer registers, more special registers and an instruction set that includes an inbuilt skip capability (this is the sort of instruction tweak that's one of those "good ideas at the time" but I don't know if it could be exploited efficiently by a compiler). The ARM was originally intended to use a small area of silicon in ASICs and was never a particularly fast processor, especially in its earlier versions. (I've been an eye-witness - but fortunately not a participant -- in several ARM based embedded disasters.) You got performance out of it by adding co-processors; I'm not sure how the modern incarnation looks but I suspect its like an x86, everyone uses it because everyone uses it rather than its inherent performance capabilities. (For me, I've always been either pure RISC or PowerPC at the detail level, with a compiler living between me and the others.)
Depends on what you want to do, with a FPGA you have the option to improve performance with hardware tweaks.
Lets pick something simple, lets say that you had 4 CPU cores and wanted to encrypt 4 blocks of data in parallel.
With a FPGA you might decide to implement 100 dedicated hardware circuits to encrypt 100 blocks of data in parallel.
Lets say that the CPU's are clocked at 1.4GHz and the FPGA's is clocked at 300 MHz.
I'll let you guess which will use a lot less power and process a million blocks of data fastest.
Try designing a new CPU without accidentally infringing someone else's patents, it's impossible.
There are millions of patents in CPU design and mostly they are trivial , overlapping and unavoidable.
Intel and AMD both have so many patents it's basically a MAD situation. (Mutually assured destruction). They both agree not to sue each other because otherwise they would end up wasting billions on lawyers fees. With a new entrant with small coffers, they can just wipe them out.
The ISA is only a small part of a processor's performance. As China has demonstrated with its MIPS and x86 processors so far, they just do not have CPU designs that are refined enough to compete with offerings from Arm, Intel or AMD. Slapping on another ISA on the same anemic core isn't going to make it magically run faster, especially when RISC-V and MIPS are essentially identical RISC ISAs.
This is just a Loongson-style CPU with a slightly different ISA.
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2024
