News for hackers. Stuff that matters.
Re:Zhaoxin KaiXian x86 CPU Tested by mkup on Friday January 01, 2021 @06:23PM (#25608406)
This article mentions Windows 10 China Government Edition:

Windows 10 China Government Edition that is designed to meet the country's security and privacy standards (Microsoft basically stripped out all of its spyware for this special version of Windows for the Chinese government, and we don't have access to a copy).

Did anyone here heard more about it?

Re:Zhaoxin KaiXian x86 CPU Tested by webmobdev on Friday January 01, 2021 @07:16PM (#25608812)
Apparently it's an 'Enterprise' edition available only for the Chinese government - https://answers.microsoft.com/en-us/windows/forum/windows_10...

It's funny that the Chinese, who have created a vast network of online and offline spywares, gets to have a Windows stripped of all spyware! Shows how much Microsofts cares about us ordinary users.

Re:Zhaoxin KaiXian x86 CPU Tested by zinekeller on Friday January 01, 2021 @07:24PM (#25608866)
> It's funny that the Chinese, who have created a vast network of online and offline spywares, gets to have a Windows stripped of all spyware! Shows how much Microsofts cares about us ordinary users.

Your point is correct consumer-wise, but saying that only the Chinese Government benefits from this is misleading. It is certainly possible to disable every telemetry - if you use the enterprise version.

(By the way, Windows 10 Home China is also like Windiws 10 Home - except that telemetry is stored domestically in China rather than transmitted to US.)

Re:Zhaoxin KaiXian x86 CPU Tested by webmobdev on Sunday January 03, 2021 @06:06PM (#25626144)
> but saying that only the Chinese Government benefits from this is misleading.

No, the Chinese version has all the telemetry code removed. There's nothing to turn off, unlike in the Enterprise edition.

Re:Zhaoxin KaiXian x86 CPU Tested by djsumdog on Friday January 01, 2021 @08:23PM (#25609260)
So Microsoft still has full access to telemetry, they just have to generate the reports inside their China divisions?
Re:Zhaoxin KaiXian x86 CPU Tested by webmobdev on Sunday January 03, 2021 @06:06PM (#25626148)
Not for the Chinese government. Their Windows version has no telemetry code at all.
Re:Zhaoxin KaiXian x86 CPU Tested by Scoundreller on Friday January 01, 2021 @08:13PM (#25609205)
> Windiws 10 Home

I had to do a search to see if that was a knockoff or just your typo!

(It's a typo)

Re:Zhaoxin KaiXian x86 CPU Tested by zinekeller on Friday January 01, 2021 @07:27PM (#25608894)
That relevant quote is misleading to say the least and possibly sensational: the real reason such a modified version exists is to replace the cryptographic standards from US FIPS to Chinese-approved ones. Of course, they (like every company and government lucky enough to have access to Enterprise versions) can turn off that telemetry (there is no detail if it was indeed removed, but I wouldn't be shocked if it instead collects telemetry to be sent off to Chinese government-controlled servers).
Re:Zhaoxin KaiXian x86 CPU Tested by trm42 on Friday January 01, 2021 @08:19PM (#25609231)
And we're talking about whole country which already censors and blocks internet traffic so the telemetry being sent to the US wouldn't be a real problem anyway except for saving internal bandwidth / firewall capacity.
Re:Zhaoxin KaiXian x86 CPU Tested by netheril96 on Friday January 01, 2021 @10:47PM (#25610202)
But Microsoft's servers are not on GFW's blocklist.
Re:Zhaoxin KaiXian x86 CPU Tested by userbinator on Friday January 01, 2021 @05:36PM (#25607962)
What's the equivalent level of performance in terms of AMD/Intel CPUs? Sandy Bridge or thereabouts? Because for the past few years, PCs have basically been "fast enough" for most things, super-inefficient monstrosities like Electron excluded. In fact my "daily driver" is probably slower than this Chinese CPU.

Incidentally, there is an interesting and active market in Chinese motherboards with X79/X99 chipsets, which are from that era, along with used Xeons and such.

Obviously those who place performance above all else will not be buying these.

Re:Zhaoxin KaiXian x86 CPU Tested by mhh__ on Friday January 01, 2021 @06:09PM (#25608278)
Seems to be roughly in line with https://ark.intel.com/content/www/us/en/ark/products/129945/... in some benchmarks as per the article

Some Reddit searching suggests a geekbench score comparable with a FX-8100

It's also worth saying that these comparisons can be overly favourable as (from the architectural design perspective) you can effectively get some performance for free from the roughly 4 process doublings in between this and the Bulldozer.

Re:Zhaoxin KaiXian x86 CPU Tested by sudosysgen on Friday January 01, 2021 @06:36PM (#25608514)
Only one process doubling. This is 16nm and Bulldozer was 32nm.

Its obviously still quite a bit behind architecturally, though.

Re:Zhaoxin KaiXian x86 CPU Tested by mhh__ on Friday January 01, 2021 @07:04PM (#25608718)
Density has gone from about 7.5 on GF to 28 on TSMC if I read correctly.
Re:Zhaoxin KaiXian x86 CPU Tested by sudosysgen on Friday January 01, 2021 @07:24PM (#25608863)
Ah, so around 1.9 doublings. Seems like I was wrong :)
Re:Zhaoxin KaiXian x86 CPU Tested by drzaiusapelord on Friday January 01, 2021 @04:46PM (#25607514)
I wonder if Windows ARM will be so mature that by the time these groups get x86 parity with Intel, it won't matter outside of specialized high-performance applications like gaming because we'll be migrating to ARM by then anyway.
Re:Zhaoxin KaiXian x86 CPU Tested by faitswulff on Friday January 01, 2021 @06:56PM (#25608647)
Other than Apple silicon, is there any other competitive ARM desktop processor? If not, and if, as the article states, the Chinese government is funding a 100% domestic silicon stack by 2022, then Zhaoxin might be competitive sooner rather than later.
Re:Zhaoxin KaiXian x86 CPU Tested by ksec on Saturday January 02, 2021 @01:33AM (#25610895)
>is there any other competitive ARM desktop processor?

Depends. There are ARM Server Processor, there are also ARM High Performance CPU-Core Design ready to be fabbed. So the question here isn't if there is one, it is who has the right incentive to fab one.

Re:Zhaoxin KaiXian x86 CPU Tested by yjftsjthsd-h on Friday January 01, 2021 @09:57PM (#25609910)
> any other competitive ARM desktop processor

Perhaps better to say any other competitive ARM machine, since the actual CPUs are generally fine (not amazing, mind, but good enough) but crippled by unfortunate firmware/motherboards/etc.

Re:Zhaoxin KaiXian x86 CPU Tested by k_sze on Friday January 01, 2021 @08:25PM (#25609276)
I’m thinking Microsoft’s SQ2 or other Qualcomm chips might end up in a laptop first (Surface Pro X doesn’t count as a proper laptop), then desktop soon after. (Complete speculation though)
Re:Zhaoxin KaiXian x86 CPU Tested by djsumdog on Friday January 01, 2021 @08:28PM (#25609291)
I'm really afraid of Linux getting left behind in the dust. It will be interesting to see if we will get RISC-V devices that can stay competitive with Apple's proprietary ARM based chips.
Re:Zhaoxin KaiXian x86 CPU Tested by solarkraft on Friday January 01, 2021 @09:35PM (#25609763)
But Linux has been working perfectly on ARM for many years (decades at this point?).

We just need general purpose computing to stay alive (and it's not actually being killed with the M1 yet).

RISC-V doesn't make much of a difference here. You can still build highly locked-down proprietary systems using it.

Re:Zhaoxin KaiXian x86 CPU Tested by yjftsjthsd-h on Friday January 01, 2021 @09:59PM (#25609925)
In fact, Apple occupies a fairly tiny fraction of ARM that's not shipping Linux out of the box - most ARM chips are running Android or some other embedded Linux spin. Inferior hacked-up Linux installs on locked-down hardware, yes, but just running Linux on ARM isn't anything like a difficulty.
Re:Zhaoxin KaiXian x86 CPU Tested by majewsky on Saturday January 02, 2021 @01:45PM (#25615209)
Running a non-ancient Linux on ARM is what's difficult. Especially when connecting stuff to the internet, it's completely unacceptable to have to rely on some 3.x or 4.x kernel that the vendor patched to death to make it run, then abandoned to move on to the next SKU.
Re:Zhaoxin KaiXian x86 CPU Tested by mhh__ on Friday January 01, 2021 @10:04PM (#25609962)
> You can still build highly locked-down proprietary systems using it

M1 is oh-so close to that

Re:Zhaoxin KaiXian x86 CPU Tested by djsumdog on Saturday January 02, 2021 @02:21AM (#25611054)
Linux doesn't really work with ARM boards. It's more of a clusterfuck. You can run mainline Linux on any PC. When it comes to mobile ARM, there is no unified device management system (UEFI/BIOS). Devicetree is a joke and no one uses it, and every device take a Linux kernel and patches the hell out of it in totally non-upstreamable ways.

PostmarketOS has made great strides in fixing this issue, but Linux doesn't just work on ARM. You have to port each distro you want to use to each and every device. With a PC, you can buy any Lenovo, Dell, HP, etc off the self and you will always be able to boot to a console; usually graphics mode too. That's not the same case at all with ARM boards.

I wrote more about this a few years back:

https://battlepenguin.com/tech/android-fragmentation/

Re:Zhaoxin KaiXian x86 CPU Tested by mhh__ on Friday January 01, 2021 @10:15PM (#25610025)
> It will be interesting to see if we will get RISC-V devices that can stay competitive with Apple's proprietary ARM based chips.

Just to really make this obvious in case of any misunderstanding:

RISC-V is an open ISA. You are free to use the instruction set. A microarchitecture implementing the RISC-V ISA can be just as proprietary as the developers please.

Re:Zhaoxin KaiXian x86 CPU Tested by mandelken on Friday January 01, 2021 @05:04PM (#25607683)
(This is from April 2020 -> could use (2020) in the title)
Re:Zhaoxin KaiXian x86 CPU Tested by zaroth on Friday January 01, 2021 @10:41PM (#25610169)
Too soon!
Re:Zhaoxin KaiXian x86 CPU Tested by 11thEarlOfMar on Friday January 01, 2021 @03:51PM (#25606941)
Would be interested to know what design software they use for these chips.
Re:Zhaoxin KaiXian x86 CPU Tested by mhh__ on Friday January 01, 2021 @06:11PM (#25608300)
In general, I'm very curious what a modern EDA flow looks like for an Intel or AMD - of those two AMD are apparently much better at automating their designs. I'm guessing not much of the HDL is ever actually touched by a human any more (i.e. generated from a specification by some in house program).
Re:Zhaoxin KaiXian x86 CPU Tested by colejohnson66 on Friday January 01, 2021 @10:21PM (#25610072)
Tools like nMigen do exist where you write a Python program that generates the Verilog for you. You can also run formal verification on it. I’m sure Intel and AMD have their own in house software systems.
Re:Zhaoxin KaiXian x86 CPU Tested by ai_ja_nai on Sunday January 03, 2021 @06:52AM (#25621261)
Is this of any relevance, given the first stabs of Apple and AWS at pivoting respectively desktop (m1) and server (Graviton) markets toward ARM?
Re:Zhaoxin KaiXian x86 CPU Tested by branon on Friday January 01, 2021 @06:26PM (#25608436)
Interesting. Makes me want to get my hands on one. Feasible for a hobbyist or no? Tried searching on eBay but didn't get many results.
Re:Zhaoxin KaiXian x86 CPU Tested by xony on Friday January 01, 2021 @07:25PM (#25608878)
Atleast someone is checking Intel's monopoly. thank you china
Re:Zhaoxin KaiXian x86 CPU Tested by yjftsjthsd-h on Friday January 01, 2021 @10:01PM (#25609941)
Surely AMD or non-x86 is in a far better position to check Intel's position? Given that they're shipping something today that's a match for current-gen Intel...
Re:Zhaoxin KaiXian x86 CPU Tested by blackrock on Saturday January 02, 2021 @03:24AM (#25611285)
Meh. My opinion is that they (China) should abandon this approach.

This was planned back in the early 2010s, before the rise of Trump and MAGA.

The reality and situation on the ground has changed. Now, the powers that be of America, is pursuing persecution and containment of all Chinese advancements in technology.

China should abandon Intel x86 and all software from Microsoft. In fact, they should ban it and make it illegal to use.

They should instead focus on Huawei’s new Harmony OS and a new RISC-V processor instruction set architecture.

They must also abandon ARM, since that’s under western control, and nVidia is trying to acquire it, which means it’ll also be under American control.

America played its card. They are willing to use sanctions for everything. The only way for China to counter, is to not play America’s game. And instead, to make their own game.

The benefit to this is multipronged. It allows domestic manufacturers to build expertise in software and hardware systems for operating systems, databases, application software, and hardware components.

But first, they must crack the photolithography barrier, and build a domestic competitor to ASML.

For Americans, you don’t have to worry about using Chinese technology, since you’re not the target market for this. You can’t anyways, since Huawei has been banned in America. So you don’t have to worry about China mysteriously spying on your Facebook posts (as if you’re really that important to them anyways).

[dead]
Re:Zhaoxin KaiXian x86 CPU Tested by ur-whale on Friday January 01, 2021 @05:15PM (#25607775)
>and almost certainly backdoored by the CCP. No thank you.

Not that I disagree with you, but to be fair, there's no way to tell any modern CPU that's in your computer hasn't been backdoored.

Intel CPU's most certainly do not have a glowing reputation in that department.

Re:Zhaoxin KaiXian x86 CPU Tested by garmaine on Friday January 01, 2021 @06:08PM (#25608266)
Backdoored by whom is an important real-world distinction.
Re:Zhaoxin KaiXian x86 CPU Tested by buran77 on Friday January 01, 2021 @06:48PM (#25608597)
People usually start off by debating the principles related to spying, surveilance, backdoors, especially in China related articles. Once the fact that most likely Intel CPUs or Windows can be safely considered "backdoored" is brought up, those same people pivot to "who backdoored it matters".

No, it does not. A backdoor is a backdoor and once it's out there it will sooner or later be used by everyone with enough resources, making the "who" irrelevant. Any Windows or Intel backdoors are very likely now in the hands of all major intelligence services around the world.

But let me frame it in a different, albeit slightly crass way: if you have a turd floating in your soup can you honestly say there's a good answer for "who put it there?" that will make you like that soup more?

[dead]
Re:Zhaoxin KaiXian x86 CPU Tested by buran77 on Saturday January 02, 2021 @11:55AM (#25614117)
> a cpu backdoor is not akin to a crouton

Of course not, croutons are generally appreciated, while backdoors are only appreciated by ignorant people, hackers, spies, and dictators.

> your stupid example

That's harsh given that you don't even seem to have processed what the example was conveying:

> you ate it right up until the moment you realized it

So then let's try again: once you realize the turd is there will knowing it's Russian, Chinese, or from the US make a difference on whether to keep eating or not? On whether to complain or not? Because GP claiming "Backdoored by whom is an important real-world distinction" suggests you, the customer, keep eating and don't complain if you like the answer.

I find the analogy to be reasonably valid.

Re:Zhaoxin KaiXian x86 CPU Tested by cesarb on Friday January 01, 2021 @06:34PM (#25608504)
> Backdoored by whom is an important real-world distinction.

Not really. Once something is backdoored by X, it can be accessed not only by X, but also by anyone who managed to hack into X. This is one of the main objections against adding backdoors (no matter how they are named, a backdoor doesn't stop being a backdoor just because you named it a "front door") to cryptographic protocols; even if you try to somehow make it so only "good guys" can use the backdoor, the "good guys" can be compromised (through technical or social means) by the "bad guys".

As an example, many telephony systems around the world have backdoors meant to be used only by local police. In a well-known case in Europe (https://en.wikipedia.org/wiki/Greek_wiretapping_case_2004%E2...), these backdoors were exploited by agents of a foreign country.

Re:Zhaoxin KaiXian x86 CPU Tested by garmaine on Friday January 01, 2021 @06:37PM (#25608521)
I would rather have a device with zero back doors, obviously.

But given that my present choice can only be between devices that are backdoored by the intelligence services of democratic nations with strong rule of law, or authoritarian regimes, and there is no third option, the choice is clear.

Re:Zhaoxin KaiXian x86 CPU Tested by seniorivn on Friday January 01, 2021 @06:47PM (#25608591)
if you are a regular person of no interest for politicians, your been choice is a backdoor from the least influential agency

so if you are Chinese you better stick with American security, if you are American you might consider chinese/Russian security etc

Re:Zhaoxin KaiXian x86 CPU Tested by BelenusMordred on Saturday January 02, 2021 @03:00AM (#25611191)
Given the choice should a western civil rights activist or whistleblower prefer a device "backdoored by the intelligence services of democratic nations" or one backdoored by despots perpetually at odds with the west?

Even a slight glance at modern history will quickly tell you what would be in the best interest of these heroes.

Re:Zhaoxin KaiXian x86 CPU Tested by klyrs on Friday January 01, 2021 @06:53PM (#25608628)
China's draconian laws can't touch me unless I leave the soil of this authoritarian democracy, which I'm quite critical of and has a long history of persecution. To tell the truth, I prefer Chinese backdoors to domestic ones -- as China will be reluctant to share keys to those backdoors with the government that has physical access to me.
Re:Zhaoxin KaiXian x86 CPU Tested by garmaine on Friday January 01, 2021 @10:03PM (#25609957)
China's security services have no qualms about extorting foreigners on foreign soil. You don't know how you're vulnerable.
Re:Zhaoxin KaiXian x86 CPU Tested by colejohnson66 on Friday January 01, 2021 @10:23PM (#25610080)
If China is able to snatch me up off US soil, that’d be quite an achievement.
Re:Zhaoxin KaiXian x86 CPU Tested by garmaine on Friday January 01, 2021 @10:52PM (#25610223)
I said extortion, not kidnapping. And it has happened.
Re:Zhaoxin KaiXian x86 CPU Tested by roenxi on Friday January 01, 2021 @09:07PM (#25609551)
In the real-world on a 5 year time horizon it is better to be backdoored by foreign governments than local governments. There is a much higher range of things law enforcement can do to abuse the backdoor than some group of foreigners.
Re:Zhaoxin KaiXian x86 CPU Tested by eloff on Friday January 01, 2021 @05:10PM (#25607741)
There's no way to trust Chinese hardware or software thanks to the lack of freedom in China for both individuals and companies. But the domestic market there is huge, as is the global market that doesn't care about security from Chinese state hackers.

Apple's CPU wasn't that interesting at the start either, it's early days yet.

Re:Zhaoxin KaiXian x86 CPU Tested by qwerty456127 on Friday January 01, 2021 @06:30PM (#25608471)
Seems yet another return of Cyrix, this time under a communist flag.
Re:Zhaoxin KaiXian x86 CPU Tested by rasz on Saturday January 02, 2021 @06:36AM (#25612098)
Centaur.
Re:Zhaoxin KaiXian x86 CPU Tested by xony on Friday January 01, 2021 @07:28PM (#25608896)
no cares US's patent troll crp , We are waiting for chinese chip to be sold in outside china
Re:Zhaoxin KaiXian x86 CPU Tested by djsumdog on Friday January 01, 2021 @08:27PM (#25609281)
If it's build on the VIA intellectual property, VIA bought Cyrix and have access to some of the licensing around x86. It will be interesting to see if they try to push these in the US and if they're legally allowed to. (They probably are).
Re:Zhaoxin KaiXian x86 CPU Tested by mhh__ on Friday January 01, 2021 @10:03PM (#25609954)
Unless they were somehow subsidized by the Chinese government I don't understand what market there would be for a slow, undocumented, probably buggy, power-hungry chip like this
Re:Zhaoxin KaiXian x86 CPU Tested by hyperman1 on Saturday January 02, 2021 @03:27AM (#25611293)
Do not underestimate the power of the word 'cheap', especially in the enterprise where some division buys the hardware for the other divisions and is only reviewed on cost.
Re:Zhaoxin KaiXian x86 CPU Tested by xony on Friday January 01, 2021 @07:29PM (#25608904)
They are able to match the performance of i5 , woooow
Re:Zhaoxin KaiXian x86 CPU Tested by neogodless on Friday January 01, 2021 @10:20PM (#25610064)
On Page 1, there's a chart comparing low resolution gaming to a Core i5, and the Zhaoxin chips are running DoTA at just 20% the framerate of the Core i5 integrated chips.

On Page 2, the Zhaoxin KX-U6780A is compared to a Core i3 and Ryzen 3, and underperforms both of them.

On the Page 3 gaming benchmarks, it often performs around AMD A10-9700[0].

The one place is does seem reasonably competitive with a 7th gen Core i5-7400 is in the 7-zip compression benchmark on Page 4, where the eight cores are put to use against a four core Intel chip that doesn't have hyperthreading enabled.

[0] https://www.amd.com/en/products/apu/7th-gen-a10-9700-apu

[1] https://ark.intel.com/content/www/us/en/ark/products/97147/i...