News Russian Company Taps China's Zhaoxin x86 CPU to Replace AMD, Intel CPUs

[Admin]

A big lost for Russia may mean a significant win for China's Zhaoxin.

Russian Company Taps China's Zhaoxin x86 CPU to Replace AMD, Intel CPUs : Read more

 

[BillyBuerger]

Is that really an 8-pin CPU power header for a 25W CPU?

 

[escksu]

Is that really an 8-pin CPU power header for a 25W CPU?

Yes it is. But 8pin atx 12v is standard these days. You can still connect 4pin connector to 8pin socket.

And if you are wondering why still need the connector. Its for 2 reasons:

1. Atx 24pin connector supplies power to pcie slots.

2. Trace routing. If you tap power from the 24pin, you have to route traces from the atx connector. As you can see, its very far away from the voltage regulators.

 

[Headbrother]

The way is "ARM" and "ARM-V"
Or Until something different is accomplished.

 

[BillyBuerger]

Yes it is. But 8pin atx 12v is standard these days. You can still connect 4pin connector to 8pin socket.

Yeah, that's a good point. Most PSUs expect an 8-pin CPU power. If it were just a 4-pin which would be sufficient, you'd have to split the now default 8-pin CPU power cable and have the second half hanging off looking weird.

 

[King_V]

I'm curious, given Russia's limited options, how much China is going to milk them on the price.

I imagine that China is going to enjoy being the partner with the much greater leverage between themselves and Russia, and will have no qualms over leveraging Russia's self-inflicted desperation.

 

[user2287]

Couldn't Intel just yank the licensing?

 

[thisisaname]

Couldn't Intel just yank the licensing?

If they could there is a few others they would like to un license before doing this one.

A license you can revoke on a whim is not much of a license.

 

[hotaru.hino]

Couldn't Intel just yank the licensing?

Instruction sets are not copyrightable. Basically it was ruled that instruction sets are considered an interface to the system. It'd be like someone trying to copyright a steering wheel, the keyboard, or a light switch. They are interfaces to something. If instruction sets were copyrightable and someone could be able to be taken down, then I'm surprised Wikipedia still has a page listing every x86 instruction.

What is copyrightable however is how these instructions are executed. That's the tech being licensed around the microprocessor world.

 

[TJ Hooker]

Instruction sets are not copyrightable. Basically it was ruled that instruction sets are considered an interface to the system. It'd be like someone trying to copyright a steering wheel, the keyboard, or a light switch. They are interfaces to something. If instruction sets were copyrightable and someone could be able to be taken down, then I'm surprised Wikipedia still has a page listing every x86 instruction.

What is copyrightable however is how these instructions are executed. That's the tech being licensed around the microprocessor world.

The x86 ISA is protected by patents rather than copyright. Patents only last 20 years, so the original x86 patent has expired, but in practice it keeps getting extended by new instruction set extensions. By the time older patents expire, the newer extensions (still protected by patent) have received wide enough adoption that a CPU without them would be handicapped, especially for a newcomer trying to break into the market. Wash, rinse, and repeat.

Implementation/execution details would also be protected by patent, or simply be trade secrets (as patenting them necessarily involves revealing the design details to the public, potentially allowing a competitor to copy your product/feature with a design modified to skirt the patent).

 

[Krotow]

By moving on such CPUs Russias IT sector is effectively moving back to begin of 2000-s. Well, it is pretty much usable for existing office and utility applications anyway. But any progress will be slow if even will happen.

 

[hotaru.hino]

The x86 ISA is protected by patents rather than copyright. Patents only last 20 years, so the original x86 patent has expired, but in practice it keeps getting extended by new instruction set extensions. By the time older patents expire, the newer extensions (still protected by patent) have received wide enough adoption that a CPU without them would be handicapped, especially for a newcomer trying to break into the market. Wash, rinse, and repeat.

In the article I posted it says:

Can you protect the instruction set of a processor? Case history suggests you cannot. There are obviously copies of Intel’s x86 instruction set as well as previous MIPS workalikes, off-brand ARM chips, 8051 replicas, plus other clones, doppelgängers, knockoffs, reproductions, and imitations of almost every CPU or MCU you can think of. Some of these ISAs are harder to duplicate than others, and some have been more successful than others, but that’s beside the point. Those are technical and commercial concerns, respectively. Creating a CPU that runs someone else’s code is usually on solid legal ground.

The exception is when you duplicate the internal workings of the processor. That’s not okay. More than a few ARM and MIPS knockoffs met their end because they followed too closely the circuitry required to implement certain operations, not because of the operations themselves. They ran afoul of patent law, not copyright law. Simply having the FOOBAR instruction in your ISA repertoire is generally okay, but you have to implement it with “clean room” hardware that doesn’t step on someone else’s patents. That’s where it gets tricky. Intel and AMD processors have nearly identical instruction sets, but their internal microarchitectures are very different.
...
Bottom line, copyright law has little to say about microprocessors. You can’t prevent someone from describing your instruction set, or even duplicating its essential functions. In both human- and machine-readable form, cloning an existing ISA is fair game. Or fair use.

The ISA is not protected by patents, it's the implementation that is. And if you think about it, all an ISA is doing is mapping instructions that a processor does to a series of bits that software can use to issue those instructions. A patent would protect you from someone using those same sequence of bits for an instruction. While sure, you could skirt around this by not having the same bits mean the same thing, it just sounds silly to say nobody else can use hex code DEADBEEF for "ADD" or something. Maybe by patenting an instruction, the actual algorithm behind it can be patented if it's novel, but you can't go after another company who made a processor to even acknowledge the bit code, even if all it returns is 0.

Besides that, if the ISA were protected, then every emulator would be illegal unlessed they licensed the ISA because they have to process it in a way that the software running on the emulator doesn't know any better. To to put in another way, emulators are software implementations of the CPU (well, technically the entire system, but it includes the CPU).

 

[TJ Hooker]

In the article I posted it says:


The ISA is not protected by patents, it's the implementation that is. And if you think about it, all an ISA is doing is mapping instructions that a processor does to a series of bits that software can use to issue those instructions. A patent would protect you from someone using those same sequence of bits for an instruction. While sure, you could skirt around this by not having the same bits mean the same thing, it just sounds silly to say nobody else can use hex code DEADBEEF for "ADD" or something. Maybe by patenting an instruction, the actual algorithm behind it can be patented if it's novel, but you can't go after another company who made a processor to even acknowledge the bit code, even if all it returns is 0.

Besides that, if the ISA were protected, then every emulator would be illegal unlessed they licensed the ISA because they have to process it in a way that the software running on the emulator doesn't know any better. To to put in another way, emulators are software implementations of the CPU (well, technically the entire system, but it includes the CPU).

Your article is some guy (who makes no claims of being a lawyer or legal expert) speculating on what the court ruling on API copyright means for ISAs. Whether ISAs, or instructions mnemonics, can be copyrighted has not been tested in court AFAIK. He talks about there having been various copies of x86 of the years, but Intel has a long history of suing over those copies.

Intel seems to think the x86 ISA (and extensions) itself is proprietary, and that that x86 emulation does in fact infringe on their IP:
"However, there have been reports that some companies may try to emulate Intel’s proprietary x86 ISA without Intel’s authorization. Emulation is not a new technology, and Transmeta was notably the last company to claim to have produced a compatible x86 processor using emulation (“code morphing”) techniques. Intel enforced patents relating to SIMD instruction set enhancements against Transmeta’s x86 implementation even though it used emulation. "
https://newsroom.intel.com/editorials/x86-approaching-40-still-going-strong/#gs.0s8p3u

To look at it another way, if ISAs are inherently open, why does RISC-V exist? Developing a new ISA for the purpose of having it free/open doesn't make much sense if all ISAs are free/open.

You could be technically correct though. Maybe it's not stricly the ISA itself that is patented, but rather Intel (or ARM, etc.) has so many patents relating to fundamental implementation details of ISA that it becomes practically impossible to create an implementation of the ISA that doesn't infringe on one bit of IP or another. Or at the very least, makes it such a legal minefield that no one wants to try, especially given Intel's litigious past.

 

[hotaru.hino]

Intel seems to think the x86 ISA (and extensions) itself is proprietary, and that that x86 emulation does in fact infringe on their IP:
"However, there have been reports that some companies may try to emulate Intel’s proprietary x86 ISA without Intel’s authorization. Emulation is not a new technology, and Transmeta was notably the last company to claim to have produced a compatible x86 processor using emulation (“code morphing”) techniques. Intel enforced patents relating to SIMD instruction set enhancements against Transmeta’s x86 implementation even though it used emulation. "
https://newsroom.intel.com/editorials/x86-approaching-40-still-going-strong/#gs.0s8p3u

Nintendo says the same thing about emulators, that emulators are illegal. And they've yet to successfully sue any emulator developer out of stopping their work.

I would also argue that DOSBox is a counterpoint to what Intel says, because it emulates something approximating a Pentium class processor. DOSBox has existed for about 20 years now emulating up to about a Pentium class processor, with builds on non x86 hardware. Granted you could say Intel is simply turning a blind eye because what DOSBox provides cannot threaten Intel's current sales (nobody buys a modern Core i9 to run DOS and Windows 95... I hope). But at the same time it's not in good form to be inconsistent with who you go after.

Also the prevailing reason I heard Intel went after Microsoft for trying to make an x86 emulator on ARM was because Microsoft had a deeper "insider" level knowledge of how Intel's processors worked and were using that to build the emulator.

You also have to consider that ARM and IBM should have a legal field day, considering that most consoles and portables made in the last 20 years have used one or the other and there's plenty of emulators for those systems.

To look at it another way, if ISAs are inherently open, why does RISC-V exist? Developing a new ISA for the purpose of having it free/open doesn't make much sense if all ISAs are free/open.

The term "open" is kind of vague. Some argue that it means that the thing is free for any one to use without licensing fees or whatever. And sure, RISC-V have some of that. But it can also mean that nobody really owns it, and RISC-V also has that.

You could be technically correct though. Maybe it's not stricly the ISA itself that is patented, but rather Intel (or ARM, etc.) has so many patents relating to fundamental implementation details of ISA that it becomes practically impossible to create an implementation of the ISA that doesn't infringe on one bit of IP or another. Or at the very least, makes it such a legal minefield that no one wants to try, especially given Intel's litigious past.

Well another thing is, what even are these patents that Intel holds? Well I did find at least one https://patents.google.com/patent/US7499962

It's for a Fused Multiply-Add instruction in the SSE extension, but it describes how an apparatus implements it. Just out of curiosity I wanted to see if there was a strict legalese definition, and well, nope: https://www.lawinsider.com/dictionary/apparatus?cursor=ClYSUGoVc35sYXdpbnNpZGVyY29udHJhY3RzcjILEhpEZWZpbml0aW9uU25pcHBldEdyb3VwX3YzMyISYXBwYXJhdHVzIzAwMDAwMDBhDKIBAmVuGAAgAA==. But at the same time, the commonality between definitions is that an apparatus is some sort of physical thing.

If anything, I believe Intel thinks they can sue based on patent infringement when they probably don't have firm case and hope that the defendant doesn't want to spend the money on a legal battle with a company willing to sink 9-10 figures to put you out of business.

Some other things to consider here that are related

• Apple in its early days was stomping on the clone market, but only because the clones contained ROM images from Apple's firmware. One company, VTech, managed to make firmware that was compatible with Apple's architecture and Apple couldn't do anything about it (https://en.wikipedia.org/wiki/Laser_128).
• Similarly, with the IBM 5150 PC, once companies developed a non-infringing firmware for the platform, clones started flooding the market.
• And then there was Nintendo with the GameBoy. Since they didn't include a lockout chip, Nintendo managed to curb unauthorized games by forcing developers to have the Nintendo logo in the ROM. Since the logo is trademarked and the boot ROM is specifically looking for it, anyone who isn't an authorized developer who has this logo can be sued by Nintendo for trademark infringement. People got around this by shoving the right bytes at the right time. They're not technically showing a logo and you can't really argue a stream of bytes is a trademark infringement because those bytes could mean anything
  View: https://www.youtube.com/watch?v=I1cUIGHZLGA

 

[drea.drechsler]

I think it interesting to think of how the "law of unintened consequences" may come to haunt the seemingly moral actions aimed at alienating Russia. Specifically: by forcing them to China these actions are giving China's nascent semiconductor industry another revenue stream to support future development. Whatever the current processors' weaknesses may be, future ones will only improve. Of that I have no doubts.

Also: one reason given for the west's banning of these Chinese products is fear of the 'back doors' they're known to put in their devices that the CCP may exploit to their own purposes. Perhaps ironic, but these chips are vital for Russia's "smart" weapons: does this suggest China will have less to fear should Russia/Chinese relations go south in the future as they so often have in the past?

 

[Krotow]

Business is trust, that's why i'm convinced the west made a big mistake by sanctioning a country like Russia, a country who trades with half of the globe even if it's energy and not high tech like china or usa. Those sanctions are making life difficult for many countries in the world, some begin to starv. They hurried into sanctions and once you put your hand on financial circuits, you don't control anything, the system is too complexe, can be out of control and everybody is loosing. Now they are buying their energy with rubles. Politics should stay away from finance. They are like kids playing with guns they don't understand
Europeans are doing things difficult to understand, for instance why germany want to commit industrial suicide.

Nevertheless mentioned sanctioned countries severely lag behind. I would not hope much on Russian engineers. Russia have special talent to cultivate gray mediocre folk and cut out and hinder talented people. Brain drain is real and will only increase. Remaining people are pretty much incapable to make scientific and engineering progress. China in general collect intellectual property of already developed things worldwide. Both ways are not very sustainable. Don't bother about European companies who are buying gas in rubles. Near end of year this process will begin to shrink fast. Russia now can getting used to taste their own oil and try to breathe their gas, while it still exist as whole country.