Chips and Cheese takes a very detailed look at the latest processor design from Zhaoxin, the Chinese company that inherited VIA’s x86 license and has been making new x86 chips ever since. Their latest design, 世纪大道 (Century Avenue), tries to take yet another step closer to current designs chips form Intel and AMD, and while falling way short, that’s not really the point here.
Ultimately performance is what matters to an end-user. In that respect, the KX-7000 sometimes falls behind Bulldozer in multithreaded workloads. It’s disappointing from the perspective that Bulldozer is a 2011-era design, with pairs of hardware thread sharing a frontend and floating point unit. Single-threaded performance is similarly unimpressive. It roughly matches Bulldozer there, but the FX-8150’s single-threaded performance was one of its greatest weaknesses even back in 2011. But of course, the KX-7000 isn’t trying to impress western consumers. It’s trying to provide a usable experience without relying on foreign companies. In that respect, Bulldozer-level single-threaded performance is plenty. And while Century Avenue lacks the balance and sophistication that a modern AMD, Arm, or Intel core is likely to display, it’s a good step in Zhaoxin’s effort to break into higher performance targets.
↫ Chester Lam at Chips and Cheese
I find Chinese processors, like the x86-based ones from Zhaoxin or the recent LoongArch processors (which you can buy on AliExpress), incredibly fascinating, and would absolutely love to get my hands on one. A board with two of the most recent LoongArch processors – the 3c6000 – goes for about €4000 at the moment, and I’m keeping my eye on that price to see if there’s ever going to be a sharp drop. This is prime OSNews material, after all.
No, they’re not competitive with the latest offerings from Intel, AMD, or ARM, but I don’t really care – they interest me as a computer enthusiast, and since it’s highly unlikely we’re going to see anyone seriously threaten Intel, AMD, and ARM here in the west, you’re going to have to look at China if you’re interested in weird architectures and unique processors.
The problem with Zhaoxin CPUs is that they aren’t Windows 11-compatible (and by that I mean they aren’t in the magic list of Windows 11-approved GPUs, who knows if they can boot Windows 11), so if you are a Windows user, you are buying a CPU with an expiration date.
So sad to see VIA’s CPU division fading away by becoming an obscure Chinese-government CPU, VIA and Transmeta were early pioneers of low-power x86 CPUs (though only VIA’s were native x86). When devices like OQO’s UMPCs came out, they looked so unreal for considering they were full-fledged PCs.
With all due respect, anyone interested in one of these CPUs probably doesn’t give a shit about Windows compatibility. I know I certainly don’t; I’d love to get my hands on one and throw all the obscure OSes at it (OpenBSD, Haiku, Sculpt, Redox…) to see if any of them boot and if so, how well they run. It’s a given that Linux will work on them but the intersection of an obscure CPU running an obscure OS is the stuff OS nerd dreams are made of!
If you’re not planning to run windows then there is very little reason to be using x86 at all. Pretty much everything else will run on arm, risc-v or loongarch (or sparc, power, but china would be less interested in these).
I can imagine the primary use case of these CPUs will be for running legacy workloads (dos, old windows etc) so the lower performance and incompatibility with win11 won’t be an issue.
bert64,
A lot of linux users want to run windows games under proton. You’re not going to get as good a result from x86 emulation than x86 hardware.
Also even for FOSS software that can be recompiled natively, Tons of ARM hardware has a longstanding issue with support that doesn’t exist with x86. I’ve been playing around with ARM devices for a long time now and this is my #1 gripe with them. I have never had this concern purchasing x86 devices, nearly all of which are easily supported by tons of ubiquitous distros as well as mainline linux. I *wish* ARM would come to this point too but on the whole ARM has evolved to be a different beast. With respect to RISC-V, I cross my fingers that it will be more like x86 than ARM, but I fear that x86 may be the exception and that vendor locked implementations are the real norm now.
Yes, absolutely there are tons of applications for random old hardware sitting around…but whether you can run your own OS on them however is much less certain if it’s ARM. Just this month I was given a three or so year old ARM chromebook, and I thought that it would be a cinch to install a vanilla version of linux on it. Boy was I naive. It turns out everything is model specific and it’s very hit or miss. Unless someone else has mapped out the road and viability first, you’re on your own. It has roadblocks in terms of google’s security chip, which I was expecting. and I thought it would just be a matter of buying a “suzy-Q cable” to reconfigure it. Turns out mine is a dud that only works in chromeos. Maybe I could spend several hours/days trying to hack the thing more, but I wish I had been given an x86 laptop instead 🙁 I can’t really complain because it was free, but nevertheless I think the point is relevant to whenever you talk about the relative usability of x86 and ARM IRL.
Linux on desktop still works significantly better on x86.
Which is a big focus for these CPUs,
OpenBSD was tested on a slightly older Zhaoxin KX-6640MA based laptop last year, which needed a few fixes, landing just in time for the 7.6 release, so there’s at least some chance OpenBSD will boot on these newer chips, if anyone gets hardware to test on.
https://marc.info/?l=openbsd-cvs&m=171854650113628&w=2
https://marc.info/?l=openbsd-cvs&m=171856082619245&w=2
I am a Windows user and I am concerned that the market for Windows-compatible x86 CPUs became officially a duopoly with Windows 11. When you have a certain amount of Windows x86 software like me, you prefer a third option to exist, even if it’s an insurance you hope you won’t have to cash. This is especially true given the trade sanctions and tariff drama happening right now.
kurkosdr,
So many of us have been looking for alternatives in the FOSS space as well. Diversifying away from mono-culture helps to give us more options and be less dependent. I wish ARM could get on board with ubiquitous standards. But as much as I wish this, they probably never will. Honestly I’ve lost my patience with ARM. Too many consumer devices requrie custom firmware/operating systems and Interoperability isn’t in their DNA. It makes ARM a poor alternative to x86. So I think we need more competition still.
All this talk about “monoculture” doesn’t make sense from a commercial standpoint. For a given form-factor, developers usually develop apps for the major ecosystems (and for desktops and laptops those are Windows-x86 and MacOS-whatever-ISA-Apple-uses-currently). This is where Desktop Linux and Windows Phone went wrong: The assumption that users desperately want to move to another much smaller software ecosystem for the sake of “diversity”. Τhey don’t, they want more choices within the existing major ecosystems, or in other words, they want more compatible choices. This is what Valve understood when the first to major versions of SteamOS were complete failures in the market and reversed course in the third version of SteamOS, making compatibility with Windows a big priority (though in my opinion they are still not there to a satisfactory degree for mainstream success).
I agree that plenty of consumers fall in line with the monoculture. But sometimes “you prefer a third option to exist, even if it’s an insurance you hope you won’t have to cash.” Clearly there are some of us who are more proactive about it.
You don’t really need to explain it to me, I’m well aware of how difficult it is to get support if you’re not running windows desktop or android/ios duopoly on mobile and even linux for hosting, etc. The market doesn’t reward going against the grain. Market share can be more influential than merit.
There are millions of us who do actually use alternatives for the sake of diversity, whether it’s linux, firefox, etc. But obviously we only make up a tiny fraction of the overall market and many vendors focus only on the top and ignore the rest. 🙁
Most users want native software support above things like proton (you can find many gamers asking for native software in steam reviews), however if there is no choice we have to tolerate windows-only software via a compatibility layer over having nothing at all. It’s far from ideal but this has always been the reality for niche platforms.
There is a positive feedback loop that keeps windows popular because windows is popular.
I don’t know how to fix this. It’s fairly easy for most gaming toolkits to compile native versions of software across platforms. Too many publishers can’t be bothered to support linux or even macos for that matter and this ends up reinforcing the windows monopoly.
Yes, they do that because they have hundreds or thousands of dollars invested in software for “the monoculture” or they need software that only works reliably in “the monoculture” to do their jobs. All this talk about “monoculture” is silly and it was invented the defenders of by loser OSes (be it Desktop Linux or Windows Phone) to blame consumers for acting in their best interests and buying compatible hardware and software.
I want a third compatible option to exist (it should have been clear I was talking about x86-compatible CPUs), and in the absence of that, I prefer keeping my existing x86 hardware and buying more x86 hardware from eBay until any trade restriction or tariff drama stabilizes or a third option emerges than leave x86.
No, it’s not, Valve found out the hard way when they tried to port their entire catalog to SteamOS. It was easy for John Carmarck because he was programming the games on Unix or Linux systems (and as a result, the DOS or Windows version was the port), but that’s not how most Windows games are made, in fact, most Windows games use Direct3D and other DirectX APIs, not OpenGL and SDL.
This is what made Valve realise that porting won’t happen and SteamOS 3.x should support Windows applications as best as it can.
kurkosdr,
Does “they” here refer to the users or publishers? If it’s user’s then I’m going to say this view is wrong. Users on alt-platforms very much want native software.
If it’s publishers, then you are right, but so many of them are already using a framework that’s already been ported to linux. They just need to output the appropriate binaries. It’s so frustrating when the publishers won’t use the portability options already available to them to support alternatives, but it’s always been this way with monopolies.
Your view seems very inconsistent: “I am a Windows user and I am concerned that the market for Windows-compatible x86 CPUs became officially a duopoly with Windows 11.”. With all due respect to me this reads like a hypocritical case of “other people are wrong when they complain about mono-cultures, but when I do it I am right”.
I’d say most publishers are not writing direct-x code, they use game engines and the most popular engine is unity. It certainly is in my collection anyway. Unity can output opengl code as easily as directx.
https://www.youtube.com/watch?v=U5HtU58e52s&t=1s
Valve were right, proton played a pivotal role in making linux gaming viable. However things really haven’t stopped there and an interesting side effect is that this actually helped make a strong case that there is demand for linux gaming. More publishers are taking notice even providing official linux support. There are way more native linux games than there used to be. Linux is still niche of course, but regardless the momentum has been going in the right direction.
Not to blow things out of proportion, but I am kind of surprised to see linux starting to make bigger splashes in the mainstream.
“I installed Linux (so should you)”
https://www.youtube.com/watch?v=pVI_smLgTY0
The sentence of yours that I quoted was:
“I agree that plenty of consumers fall in line with the monoculture.”
So, I clearly meant users/consumers. And yes, users have collections of paid software of either Windows or MacOS or need certain Windows or MacOS software to do their jobs.
Sure, but those users on alt-platforms are, like, 1% of the market and publishers don’t care, that’s the eternal chicken-and-egg problem of alt-platforms.
No, they can’t, some game engines are in-house and hence Windows-only, and even game engines that do support Desktop Linux (for example UE and Unity) don’t guarantee that middleware and DRM used by a game will also support Desktop Linux, and generally, publishers don’t want the support burden of a third ecosystem to serve a tiny minority of users. Even getting their games to work with Proton is too much of a support burden for them (which is why Proton must become more complete to become fully mainstream instead of relying on publishers).
See, this is what you don’t understand: I don’t mind the “monoculture” (or “duo-culture”, since MacOS is also a thing), I want more choices within the “monoculture”. For example, VIA and Transmenta were additional choices within the Windows-x86 “monoculture”. Similarly, Proton, if it ever works well-enough for the mainstream, will be an additional choice within the Windows-x86 “monoculture” (and yes, I find it ironic the plumbing underneath will be Desktop Linux, and no, I don’t care).
In other words, I want choice that respects the investment I’ve made in Windows-x86 software, any choice that does not respect that investment is irrelevant to me (and so is for any user invested in the Windows-x86 ecosystem).
See, this is what you don’t understand: People who buy the Steam Deck don’t buy it because it’s Desktop Linux (the average Steam Deck buyer really doesn’t care), they buy the Steam Deck because it runs some of their Windows-x86 games in a cool new form-factor and is very good VFM, which certain PC gamers find useful. In other words, it’s an additional choice within the Windows-x86 “monoculture” (an incomplete one, but useful for a significant minority because it allows them to bring their investment in Windows-x86 software to a new form factor for a very good price).
kurkosdr,
Well, the reason I asked is because I felt that linux users wanting linux software was self-evident. We’ll consider wine solutions if there’s no choice, but most of us consider it a crutch.
Yes some of them use in house engines, but most of them do not. I did check and a majority of the windows titles in my collection are built on a prebuilt engine like UE and I wouldn’t be surprised if this holds true in general. I agree that many publishers can’t be bothered to support a small market share, but this was always my point.
I agree that DRM can be a problem, but as you pointed out it is a problem for steam/proton too. I wouldn’t call it an advantage to windows software via proton.
Proton was undoubtedly the answer to “how do we get from a state of no linux gaming to a state of linux being taken seriously for gaming”. As it becomes more mainstream we’ll see more publishers supporting linux natively too.
Fundamentally it’s the same thing: we all want more choices that help us be less dependent. The way you go about this is up to do. You get to make and value choices differently than I do and this is fine! The hypocrisy stems from placing importance on choices that cater to your needs while dismissing the importance of choices that caters to someone else’s needs. I don’t know that we’re going to come to an agreement, but I thought it would be a good opportunity to highlight how the way you crave more competition within x86 is the same way other people crave more competition outside of x86 too.
Why do you assume that’s my view? It doesn’t matter why something becomes mainstream, the mere fact that it does makes it that much more likely for executives to agree to support it. It really is as simple as that.
The motives you want to attach to steam deck buyers are fine, but don’t overlook the big picture long term: the increase in market share for whatever reason will translate to executives green-lighting more native software. The amount of native games available on linux has already been picking up momentum for several years now, thousands of titles and growing. I know you hate to concede any wins to desktop linux, but frankly as a desktop linux user I actually do see it as a win regardless.
Wrong. These CPUs are Windows 11 compatible, and have been benchmarked by 3rd parties using Windows 11.
https://wccftech.com/zhaoxin-kx-7000-8-core-cpu-intel-core-i3-8100-equivalent-multi-core-performance-inefficient-design/
The problem with these CPUs is not compatibility, but rather their poor performance and efficiency. They exist merely as a placeholder for x86 within China, especially for large deployments that still require windows compatibility, to navigate all the uncertainty of the US-China trade relations and blockades.
Thom Holwerda,
I don’t have a use for one, but it would be neat to see what it could do. Certainly not at that price point though. Probably doesn’t even include the 150% tariffs coming to the US. I was complaining about inflation before, oh boy this is going to be economically stressful.
That’s exactly how I feel about it. I love playing with obscure hardware and seeing what I can make it do.
Yeah and I’m glad I have a Micro Center half an hour’s drive from me, so far they haven’t raised prices on anything in their “maker” section but I have a feeling in a few weeks the hammer will fall. I’m trying to get up there every weekend to get odds and ends for projects before that happens.
I may be the only one that is impressed that these chips are as fast as Intel chips from a decade ago. That is faster than the fastest designs you can license off ARM I believe.
However, as somebody that routinely uses older hardware, the key is that this is “fast enough” for the core mission here. These chips exist to ensure that China could self-supply if they lost access to x86-64 from Western countries. You typical office worker could be given a PC with these chips in them and not miss a beat. They will run a web browser and Office apps just fine. And when you consider that even web browsers offload to the GPU these days, pairing these chips with a reasonable GPU makes them even more capable of hosting modern workflows. Not everybody is editing video or running Triple A games. In fact, most people are not. And tools like “AI” can be handled server side through a browser or through an API and DeepSeek has already shown that China is doing just fine on that front.
I also have no doubt that these chips could get quite a lot faster very quickly if they needed to. These are being manufactured at something like 21 nm. China is already pushing 7 nm capability elsewhere. Shrink the process, add a few cores and suddenly you have caught up another 5 years or so with the west. Focus more R&D on them for the next 5 and you may find yourself only a year or two behind. That may mean that Western consumers still do not want to buy them but they are certainly more than good enough for those that cannot buy Western tech. They are probably also more than good enough for many industrial uses globally.
In my view, these chips show that China could absolutely keep their internal PC market and their economy tracking without disruption even if they entirely lost access to products from companies like Intel and AMD. It also shows that, if they have to build internal capability further., they are not so many years behind competing for the same customers as western chip suppliers are today.
Absolutely not. You can license an ARM cortex X/A core hat will mop the floor with this design, and be on par or slightly better than any current x86 core.
First, ARM is better now than I gave it credit for in my description. However, let’s not benchmark against “announced” designs. There are amazing RISC-V “designs” as well that will never see silicon, even though they are available to be licensed. Meanwhile, what you can buy is stuff “announced” 4 years ago.
Where can I actually get an X295?
In this article about “The World’s Fastest ARM PC”, it says “the single-core performance of the Altra Max (at 3.0 GHz) isn’t stellar – just above the old Threadripper 1950X (2017) in Cinebench 2024”. That is still a fair bit faster than the Zhaoxin of course.
https://www.tomshardware.com/software/linux/worlds-fastest-arm-pc-put-to-the-test-128-core-ampere-altra-max-cpu-claims-128-core-cinebench-2024-world-record
Anyway, I am not here to bash on ARM. Both Qualcomm X Elite and Apple Silicon are amazing and it sounds like ARM themselves are catching up.
My point was simply that the current generation of AMD or Intel is a misleading benchmark when assessing how viable it is to rely on other processors.
You can get an ARM-X925 on any phone that uses Mediatek’s Dimensity 9400 SoC, for example, of which there are a bunch of in the market right now depending on where in the world you live.
I have no clue what your point is.