This article adds good background. Also what the announcement seems to be about is a design for a chip yet to be put into production. How this will compare to arm and intel cpus in terms of cost and performance is unclear (to me).
Reminds me of Adapteva's Parallela, which I bought and was really excited about, but was obviously a commercial failure. I work on software that has an above-average (but far less than HPC) need for concurrency, but even our most demanding customers seem to be content with off-the-shelf Intel, either because they don't need THAT many cores, or because they value the more sophisticated instructions (like encryption acceleration, etc.) I'm curious to know what kind of use cases y'all work on that might crave something like this.
Modern x64 is very difficult to beat and takes a lot more than just raw core count. The out of order execution, prefetching, multiple execution units, SIMD, cache, cache synchronization etc. are all very strong. I think the space where more weak cores can be utilized but not GPUs is pretty slim. Intel themselves even had that briefly with their high core count out of order atom core chips that didn't take off (or weren't given a chance).
I think software that can take advantage of more than a few cores without just using brute force fork-join parallelism in some of the heavy loops is rare. I don't think it has to be that way, but the problem comes down to software architecture which isn't going to be solved by leaving each application programmer to their own devices. It will take libraries that give them the means to do it without having to reason about low level synchronization.
>> the space where more weak cores can be utilized but not GPUs is pretty slim
Yeah I think that's what it is. As you said that I realized that I do actually have a few customers that can benefit massively from many cores, but they're just using MANY, MANY cores and it's all CUDA. I don't see many people going for the compromised approach of say, Xeon Phi.
Based on benchmarks (take them at face value) [1], Apple's A12/A12x processors are roughly 80% the performance of current-gen desktop CPUs (and on par with some mobile ones) in single threaded performance. I think a bit of that is power concerns, even with the 8 core chips.
There have been rumors floating around that Apple will switch to their own ARM cores (based on the A12, probably) for a future generation of Mac computers. I think they'd have to meet or beat x86 for that to happen.
Interestingly they're manufactured with TSMC's 7nm process, just like the new AMD chips.
It wasn't really open sourced. It's open in the sense that if you pay a bunch of money, and sign an NDA, you'll be given access to the source... which has always been the case. Some sales bro bought MIPS and slapped an "open" label on it. IMO, it's current steward has done more to kill it then ARM and RISC-V ever could.
When they first did their kickstarter they were talking about doing a follow up with 64 cores and being able to connect them together with those fancy pants connectors they had going on. I was all set to buy four... and none of that ever materialised.
I got the impression that a lot of folks had unrealistic expectations about how Linux would support the chip right from the start, which honestly wasn't all that surprising recalling the way they worded some of their promotional material.
One presenter claimed the board would pay for itself if you used it for mining...
I assume something like that would be much easier to implement than a real multicore system like linux but I never heard of anyone mining on their boards.
Maybe back in 2013 that might have been possible, but it seems like nobody figured out how to effectively use those boards back during that timeframe. And since then GPUs and then custom ASICs took over.
RIP. I have one of those boards from Kickstarter that's dead due to thermal destruction of the chip. They ran incredibly hot and didn't have sufficient throttling. Not that it was very useful anyway, but I liked the company's mission and attempt.
Nobody wants this particular chip but there isn't money to pay for a RISC-V Manhattan Project so there will have to be an incremental evolution of several generations of RISC-V chips before they produce a usable one.
> either because they don't need THAT many cores, or because they value the more sophisticated instructions (like encryption acceleration, etc.)
The parallela was much more complex, had its own bus architecture which required an FPGA to glue the chip to the Arm cores in the FPGA SoC which ran Linux. Compare that to a simple x86 or Arm system where the tooling and software is mature and you don't have to do any extra fancy work. Just write a program and execute it. Same thing that killed the IBM/Sony Cell processor. Neat idea, PITA to program.
It might've done well as a PCIe (or thunderbolt) thing. There are definitely large scale users of scalar math that could make use of it but you'd need to build it in 5nm and dozens of cores to be even remotely competitive.
Alibaba doesn't seem to have any press releases written about this, not even in Chinese. All reporting seems to be based on what CTO Zhang Jianfeng said at the currently ongoing Alibaba Cloud Summit ( https://summit.aliyun.com/ ). Maybe they'll have more information on the specifics once the summit is over.
> How long before Android runs properly on RISC-V?
It's mostly up to OpenJDK and V8 porting. When it comes to porting V8, I've found out the hard way that there's more to it than a generic understanding of compilers. Somebody still has to step up to the plate for a real JVM.
> And will they be able to persuade app-makers to rebuild all apps for the platform - x86 Android shows how hard that can be.
Real x86 Android devices had libhoudini (DBT), which was pretty decent. If RISC-V starts with low-end devices, or with a high-visibility model that developers are pretty sure will sell well, then the native compilation target will become more popular over time.
If China's bizarre local app market is involved, chances are they'll start supporting the target as soon as it's available.
Slightly longer answer is that _some parts_ of the OpenJDK are used. Namely, LUNI (implementation of java.lang, java.util, java.net, java.io) run on the device, but those are high-level enough that they will require either none or minimal changes for risc-v.
Compiling applications also uses OpenJDK (or the Jack compiler in platform builds) but that is an intermediate step on the host. The produced JVM bytecode gets converted to Dex bytecode during the build process.
I think the effort for Android will be two-fold:
On the software side, Bionic (android's libc), Dalvik (runtime), LLVM, external dependencies (i.e. boringssl, lib{vpx,hecv,mpeg4,etc},) definitely need work.
On the hardware side it all lies on the practical availability of a SoC. While I have no doubt we'll have something similar to android-x86 running on discrete chips, commercial devices will need a more complete cpu+gpu+dsp+modem package.
> A small correction, Android does not use OpenJDK. It uses the ART runtime
Oops! I got confused after there was uncertainty in 2015 because of the Oracle lawsuits. I believe it was reported at that time that they would switch to OpenJDK, or something similar enough was reported that that's what I remembered.
> On the software side, Bionic (android's libc), Dalvik (runtime), LLVM, external dependencies (i.e. boringssl, lib{vpx,hecv,mpeg4,etc},) definitely need work.
Not all of those need to be in software, but yes, it all needs to come together.
> commercial devices will need a more complete cpu+gpu+dsp+modem package.
Totally, though the non-CPU bits have little to do with RISC-V either way, and the drivers for them tend to be (relatively) portable C code, so that bodes well.
It has a lot of idiosyncrasies, and the upstream is a moving target.
When I began, it was easy to write an assembler (especially for RISC-V), but it was not easy figuring out which macros and stubs you needed to fill out before you could test them. It's hard to tell if you're making progress.
Each of the existing backends has its own signature, which betrays whichever backend was copied as a starting point. Some of them have comments on a given chunk, while the same piece of code in another backend might have no comments.
The other thing is the lack of a portable interpreter. This choice makes great sense from a performance perspective, but it's a pain when you just want to make incremental progress.
I have other engaging work now, so my port is abandoned, and I hadn't committed anything worth pushing for a while when I stopped.
I think it may make more sense if you ever worked on a hotspot target.
Though all things considered, I'm fairly certain I could do it in not much time if it was my full time job (which is what I said in a couple of interviews last year, when I was looking for a company willing to make that a thing).
Many apps will work with minimal if any changes as they don't use native stuff at all. There's also some interesting solutions for this like libhoudini that emulates ARM to make many native apps work without any effort on the app developers part.
I haven't thought about it much, but first reaction - I get that it might be a good candidate to get volumes up but isn't it kind of self-defeating? Why would you want to run android, of all things, on an open CPU?
Android is the leading OS for mobile phones and tablets. It is also quite popular in embedded for devices that have a screen, mostly as a side-effect of its dominance in mobile. It can often be easier to get graphics and touch drivers/support for Android compared to Linux. It is also easier to find developers who can write a custom Android app for the UI than any of the typical embedded/Linux UI frameworks, and the tooling is better.
Besides high quality hardware support and device availability. The fact of the matter is the core of Android is OPEN SOURCE. I can and have built every last bit of software on my Moto X4 thanks to LineageOS, which is based on AOSP.
ah right. the point still stands, though. i was also under the impression that many of the device drivers that ship with with modern android phones are closed source. see the step in building lineage "Extract proprietary blobs"[0]
Those are the firmware pieces running on the cores that aren't even running Android as their OS for the most part, so it's not really applicable to the root discussion.
Not the parent commenter that you are replying to, but yes LineageOS even gets the radios working. Its a full mobile android distribution that works great
i know, it's what i use daily :)
i'm fairly confident they have no builds that are "blobless" though - part of the build process is to extract proprietary code blobs from manufacturer images which are not open source, see [0]
It's extensible, so they could in theory introduce something to the ISA which would give them a competitive advantage, maybe even a publicly available extension that their competitors have simply left out, e.g. one of the crypto extensions.
There's also the benefit of being more flexible with teams who work on your SoC. With RISC-V, everyone has an ISA license, so they can contract, team-up, collaborate, or do basically anything they want, in order to work on their chips.
This also means that acquisitions are a much richer market than they are with other ISAs; in part because you don't have to justify the cost in time, money, and uncertainty of licensing or planning to license ARM.
Would they not be so enthusiastic about sanctions, Chinese would've been way more cautious jumping on RISC-V boat.
Now instead of having RISC-V as one off research projects, which would have probably died off silently, now they are all treated as survival essential projects.
Not many other companies or developers are involved in WDs effort, thats mostly an internal project. There are tens of obscure ISAs and CPUs that are used in similar applications, and no-one is particularly excited about those. But RISC-V seems to have serious momentum in both academia, and as base for upcoming platforms from new companies.
The best thing for China is that this is happening at a time where the US is insistent that the brightest Chinese minds in the US go back to China, and the ones in China, stay there.
It's genius. Genius, I tell ya. But the current administration doesn't care, because the negative impacts will only be felt a decade from now.
The same thing happened when China wanted to join the ISS[1] - the US Congress blocked this to avoid "tech transfer". Instead of being dejected and sad, China kickstarted its space station program, and are now collaborating with Europe.
Europe could have stepped up and become competitive long before, but they are too complacent with the status quo. It's good that finally someone steps up.
I wonder if most of what we are doing is being done for political purposes? Sort of, just to look good and get more votes. And if so, you kind of have to admit that from that perspective, the calculation works out even if China surges ahead even faster.
(In fact, you might even be able to use that surge, to do more grandstanding, to get more votes.)
Not saying this is 100% what's going on, but it's undoubtedly a factor in the equation here.
For sure. Third party countries may also decide less risk in buying from China (ie, China less likely to cut them off because of real or perceived issues).
It is probably no coincidence that ARM was sold to Softbank right when RISC-V was starting to make a name for itself. They saw the writing on the wall.
"RISC-V" isn't an organization that can "move to multi-chip modules". RISC-V is a foundation which essentially publishes a couple of PDF files (the user spec and priv spec) and organizes a bunch of other stuff around reference hardware designs, simulators and the software toolchain.
All of this is BSD licensed, so anyone can pick it up and manufacture chips based on the specification or by modifying one of the reference designs. Note that the largest part of designing and building a chip is not the choice of ISA (that's probably well under 10% of it).
What you may want to know is whether any manufacturers are going to make modular RISC-V chips. None of them to my knowledge. But most of them are currently focusing on the ARM space (embedded, IoT, AI, etc.) including this particular design.
The nice thing is that software and work on a RISC-V toolchain can now be much more common for vendors that elect to use RISC-V, though it remains to be seen how well the actual manufactured chip variation gets supported like instruction set extensions, peripherals, and configurations for many cores. ARM had some motivation to help ease the customization pain for tooling for their IP licensees, so I'm curious if RISC-V support will be more like ARM or more like early UNIX vendors.
Like a lot of chip desicions, this one is based on the effects of the process node in question. So the answer is probably when a company gets to TSMC 7nm while being capable of the high end bonding on the package substrate between chips, at a per chip gate count probably close to the ryzen cpu chips. Before then, it doesn't really make sense.
MCM is a old trick in the toolkit, eg Intel x86 and POWER chips have used it on and off over the years, as have various game consoles. The downside is the cost and complexity of packaging, and the power required for inter-module communication. Many of the past usages have been to closely couple a die from a DRAM process and another die from a logic process. It's not obvious that MCMs are becoming more popular on average in the future.
This is both awesome and terrifying. I'll never forget this research(https://www.wired.com/2016/06/demonically-clever-backdoor-hi...) which showed how an attacker can insert a backdoor into a chip during the manufacturing process in a way that is practically impossible to catch. That makes me very concerned about using a chip completely manufactured in China.
But at the same time, RISC-V needs to happen. It's a critical step to move computing forward. It's just a damn shame that this is the way it's going to happen. Hopefully this will push RISC-V manufacturing efforts in the US before the market becomes flooded with cheap open source and most likely backdoored chips manufactured in China.
Considering that this was announced by Alibaba Cloud, there's no need to worry about hardware backdoors since you probably won't be able to buy the chips anyway. If you are hosting something in their cloud, a backdoor is unnecessary, since the authorities can just come knocking on the front door if they need anything.
Is there a way to know that the present chips are not backdoored? Or, backdoored by US? Why should the world be concern about Chinese backdoors and not US ones?
Put it this way— as Americans, we’re all pretty confident that our government doesn’t pay any substantial number of people to go online and propagandize our human rights abuses under the guise of just being reasonable world citizens.
(Maybe that’s just because we know that right now our government can’t scratch its ass and elbow at the same time, that we won’t face any penalty for our human rights abuses anyway, and there’s no point in paying bureaucrats to do something that ignorant yokels will go out and do for free... but that’s beside the point :)
As a non-american, I am reading this in awe. You obviously believe this is true. I would suggest there is a counterfactual quality here: Maybe you just do it better
I'm not by nature a conspiracy theorist. I don't think the grassy knoll is here. But, I do think that the US government propagandizes inside, and outside all the time and has done so continuously, since the birth of the modern communication era.
You write as if the Vietnam war just didn't happen. That routine abuses by US armed forces in the middle east just didn't happen. As if Snowden just didn't happen.
Maybe I missed the /s in this. Btw, if I got a whack from the site I wouldn't deny this may not be a useful discussion, but I really do mean it: I think this cannot be read, except as a counterfactual non-reality.
> I do think that the US government propagandizes inside, and outside all the time and has done so continuously, since the birth of the modern communication era.
This is a valid point, and one that deserves focused thought. Because it's both very right and very wrong.
The US does generate constant propaganda on most issues of consequence. BUT. The US also generates constant propaganda on multiple sides of every issue.
The propaganda ecosystem is far more robust and distributed in the US. Whether it's straight journalism or spambots, or individuals expressing their heartfelt beliefs.
Maybe that's doing it better? Maybe it's just one mess after another?
One thing I know - it's quite a different from the sort of propaganda you get from a country where the national government maintains a list of forbidden words, events, and attitudes.
You seem to have missed the second paragraph, which tempered the one-sidedness of the first one.
But even then, I think gp is referring to a difference in kind, both in behavior and communication. China's record (present included) on human rights is truly abysmal and incomparable to any nation with individual legal rights.
Propaganda comes in soft forms, which its easy to claim everyone does (I'd agree with that). But there's also propaganda along the lines of "our 'education centers' in Xinjiang are teaching the 'pupils' valuable new job skills!"
Many schools in the USA are indistinguishable from minimum security prisons and detention centers. The only major physical feature that separates them is that schools tend to have larger parking lots and lack guard towers.
Whether you are talking about Xinjiang education centers or compulsory government schools in the USA the excuse for them is going to be along the lines of 'We need to ensure that these people become productive members of society'.
All in all they are compulsory miserable dehumanizing places that are designed to condition and manipulate people into what 'society' demands from them; as defined by the ruling authorities. In both cases anybody attempting to avoid conditioning and years in "detention" is guilty of a criminal act.
My point being that these things are not differences of types, but of degrees. And while it's easy to point fingers and laugh and pride yourself on the superiority of your own country, the take home lesson shouldn't be how much better America is then China, but to use them as a mirror to examine how and why things are better here and how to improve on them.
Reductio ad absurdum. Why don't we say society itself is a miserable dehumanizing construct, the very first layer of expectation foisted upon our free natural selves? A country has borders... what are those but abstract prison walls?
If you are saying elementary schools are reasonably comparable to concentration camps, you have lost the thread of the conversation.
Hand-waving and semantic games don't change the original discussions: who would you rather hold power over you?
1. A democratic government that enshrines human and political rights
2. An authoritarian ethnonationalist oligarchy that enshrines the primacy of the Party
Whoever powers your chips will at least potentially have that power over you. We're in an era where we can decide, but anything can change. Our choices and the values we act on now will set the stage for what is available to us in the future.
Very few Americans demonstrate the level of ignorance displayed by previous poster. In fact the level of paranoia and distrust can be almost palpable.
Things like Vietnam and Snowden are why so many people in our country treat things like gun rigts as a quasi-religion. Fear of the citizenry is one of the few remaining serious checks on government power they believe we have.
However, based on personal experience, only 20-30% of people really realize the level of propaganda we have been subjected to our entire lives. One of the big reasons for this is because USA was the pioneer in government and corporate propaganda. They have been doing it so long and spent so much money on it that it's practically a art form.
In fact I doubt 1 in 10 people every really thought about the fact that the entire economy of the WWW is almost exclusively based on propaganda in the form of advertisements. Public Relations Departments used to be called Propaganda departments before the Nazis came along and turned it into a bad word.
But even then it's not lost on a great deal of Americans that the only people in the past 40 years who have ever gone to jail for war crimes, torturing, and other forms of criminal malfeasance on the Federal level are the ones that exposed it to the public.
Why would we be pretty confident about that? Edward Snowden himself revealed to what depths the US would stoop, and it's got to be a thousand times worse now than what he reported. There are, in my opinion, definitely Americans out there spreading propaganda about the US on taxpayers money. You yourself could very well be one of those people.
The idea that the US doesn’t covertly spend money on propaganda, when it overtly does and has for decades, is hilarious.
I mean, VoA is actually a pretty good, so this isn’t intended to be criticism of it. However, it is intended as propaganda, although a fairly light version of it.
> Put it this way— as Americans, we’re all pretty confident that our government doesn’t pay any substantial number of people to go online and propagandize our human rights abuses under the guise of just being reasonable world citizens.
You would be desperately wrong in that assumption.
I think the astonishing thing is how incompetent our foreign service is today; even in 1979 the U.S. Embassy had hardly anyone who speaks Persian and got completely blindsided.
In 1954 the US was capable of kneecapping Iran and other countries, but today it is too blinded by an ideology that refuses to accept it is an ideology to do the same.
Surely you're aware that the 1953 coup d'etat overthrowing the democratically elected government in Iran, and installing the Shah in its place, and hence setting the stage for Khomeini later, was orchestrated by a certain country that you mention?
I posit that if the service in question had been _really_ competent, they'd have let sleeping dogs lie and _not_ meddled with other governments. Pretty much the entire world is paying for their incompetence today.
Then again, incompetence is not absolute, and I guess you're lamenting the slide in competence (or the lack thereof, depending on how you look at it) of the service in question.
Right, every (x86) CPU sold since 2008 is already backdoored by the US. Honestly as US citizen who will never go to China, I'd rather be backdoored by China, since they can't throw me in jail, seize all my assets, or generally ruin my life nearly as easily as the US could for breaking some future law.
> After all this research we definitely found the latest AMD CPUs which do not include any PSP functionality. According to a table from CPU World we know now that these CPUs are those we searched for:
x86 and its descendants are a no-go for me because of that. Neither Intel nor AMD are willing to produce x86 chips where the end user has full access to the highest privilege levels.
Lets not jump into politics! Backdoors are bad and it doesn't matter who does it.
Good to see development from any country.
Obviously they have enough reason to do it. Just consider sanctions. They also have to protect their own market.
Anyone that can afford it, should do it. Ultimately it is a good thing from technology prespective.
> On Aug. 24, Chinese state media announced a proposed change in the Criminal Procedure Law which would allow police to legally detain individuals and hold them incommunicado in secret detention for up to six months without contact with either their families or legal counsel.
Up to six months! That's not bad, considering the fact that some prisoners have been held in Guantanamo for 10 years+ and can be detained there indefinitely without trial and subjected to torture [1].
The Chinese don't intend this for terrorists who are blowing up rice markets filled with women and children in Iraq, they intend this for journalists and protestors.
Who are allegedly blowing up rice markets filled with women and children in Iraq. Remember, there is no trial.
Also, according to the HRW article:
> The Chinese government is pitching the proposed change as merely an extension of the conditions of the existing practice of residential surveillance, or “soft arrest,” to suspects in state security, terrorism or major corruption cases. “Soft arrest” allows police to confine criminal suspects to their homes for up to six months without trial or due legal process.
A different Khalid al-Masri is a suspected member of al Qaeda. So they tortured the wrong Khalid. Oops.
This is why basic things like the right to a trial, open court proceedings that journalists can attend, the Geneva convention prohibiting torture, etc etc are so important.
To trust a government to decide guilt in secret is to be ignorant of history.
Enough with the whataboutism. Everybody knows about the dirty business the US gets up to. It doesn't excuse Chinese authoritarianism and human rights abuses.
Have you ever spoken to a Chinese? They have their own good intentions too. Besides that, why do you always trust that our governments’s label of “terrorist” is never self-serving or opportunistic? The civilians we bomb every week might have a different perspective
Or for guys who had an opportunistic neighbor who wanted his land and was willing to "rat him out" to the US forces desperate to capture some bad guys.
According to your link about one thousand people were sent to Guantanamo. Right now in the region of Xinjiang about 1 million Uyghurs are being held in re-education camps (about 10% of the Uyghur population of the region).
Even if the USA compare terribly to the (rest of?) civilized world regarding human rights, as a non-American and non-Chinese I fear less the US government than the Chinese. (The fact that I'm not from a Middle East country probably helps...)
America sent all its US citizens of Japanese descent to Manzanar at one time. Many died there. The CIA overthrew the democratically elected government in Chile on September 11, 1973 leading to the deaths of 11,000+ people. Iranian-Contra, Bay of Pigs, United Fruit Massacre, The School of the Americas ... I mean, America isn't at the same scale because it's 300 million people vs over a billion.
Anyone who's done research would understand China is the one country that has no reason to make the effort of developing a backdoor because it's required that all companies will hand out their data through the front door, and the vast majority of Chinese people seem to be okay with the security vs privacy trade off.
It will give you true picture of how Hong Kong lost to mainland cities in terms of quality of life and innovation. So economy is the main root of all this frustration and protest. Extradition bill is just the spark.
Lived long enough in Hong Kong before 1997 and after 1997 and being a Hong Kong family can understand it from inside/out. So sorry to break your bubble they can not take care of their economic situation without China, it was true before 1997 as it is today. Hong Kong's survival is China, which did not tax it yet like other cities of China with better GDP, but spend on it. Hong Kong got rich when China was closed acting as middle man and earning US$ 5-20k, doing nothing just some paper work on each container shipped from China. Easy money, I was one of such beneficiary.
If you want to talk about human rights, please read through the first instance of re-interpretation by China of Basic law of Hong Kong, when Hong Kong people requested China to re-interpret the basic law against basic Human Rights not to allow mainland spouses to enter Hong Kong. So Hong Kong is still denying the most basic human right of a family, but at that time it was in the interest of Hong Kong people and bad for Chinese spouses, so no protest against it neither any backlash. Indeed overall society was very happy at the cost of infringing human rights. You do something wrong once and open the doors, it will come to bite later. If at that time Hong Kong people would not have involved China for re-interpretation, it would have brought 150,000 people from across the border, along with it some more hardships, but would not have eroded the freedom. Hong Kong people chose prosperity over freedom and now will pay for it.
So fighting for human rights now is just a farce, main reason is, get rich quick does not work, majority live in prison cell size apartment(again by choice) and have no way to change it.
Let's not pretend how they are governed internally is what concerns the rest of the world.
The US has been in constant wars, invasions, meddling with foreign politics, running protectorates, etc. for over a century, including having bases all around the world. Far more countries have battle and regime change scars by the US than by China. Including it being the only country that ever dropped atomic bombs (and to civilians, none the less).
And it's not just the Orange Baboon in charge that's worrying (though of course there's that too) -- the "Nobel Peace Price" winner Obama had a worse track record.
(Of course for Americans anything that happened over 20 or 40 years ago is ancient history and the world should have forgotten it and totally trust them now -- still there were like 4 wars and one regime change they assisted in the last 20 years, so there's that).
Such a shame to insult Americans lack of historical context while also seeming to insinuate the world was some sort of idyllic utopia prior to American engagement.
How about we don’t forget that the entire world has pretty much been a gigantic warzone ever since civilization started and how about we remember that American hegemony has been the single most peaceful period of world history.
I’m not even trying to insinuate America should be some great arbiter of what is good and holy, but the fact of the matter is, people have been slaughtering eachother since communities became a thing, and only recently has it become so infrequent that people can go “oh! Look how America fucked up by bombing someone here!”.
As if, without America, there wouldn’t be regional conflicts pretty much everywhere, just like there literally always have been.
>Such a shame to insult Americans lack of historical context while also seeming to insinuate the world was some sort of idyllic utopia prior to American engagement.
The world doesn't have to be an "idyllic utopia" not to want another player meddling on other countries business -- much less a 10000 pound gorilla one...
>How about we don’t forget that the entire world has pretty much been a gigantic warzone ever since civilization started and how about we remember that American hegemony has been the single most peaceful period of world history.
If we forget the Cold war, its proxy hot wars, meddling and support for dictatorships (Iran, Banana republics, Chile, etc) and the associated death toll, and so on. Including a few possible close calls for hundreds of millions (Bay of Pigs for one).
And the relative peace is not because of some protective attribute of "american hegemony" (and e.g. its benevolence), but because the few other major players (basically colonial opressors like UK, France, etc) got burned in the 2 world wars and got off the main game, so no great fights between them anymore...
I think you are both correct. But that only highlights the question about whether we can trust chip manufacturers. I currently tend to say yes, but I would still prefer to buy the American one.
Well, I'm of the smaller European countries that were fucked around by those Europeans colonial powers, so there's that...
But if you think it's some badge of honor for your country to "fuck around much of the world", and you think only bitterness for not being able to fuck colonies etc, as much as before is why anyone would oppose it, sure, more power to you...
You seem to forget that Pax Americana is the most peaceful time period in the history of mankind. Is american hegemony perfect? No. but would you prefer a Soviet or Chinese hegemony instead?
All things considered, america has been pretty decent considering it has been the uncontested most powerful force in the world for the past 80 years.
Getting back on the topic of backdooring chips, the US has a strong track-record lately of leaking secrets. You can take some solace in the fact that anything the US would try in that arena has a much higher chance of being leaked compared to authoritarian regimes.
...are you actually trying to say that China is "morally superior" to the USA? Because, without bothering to delve into the tedium of enumerating all the reasons, that is definitely a losing argument.
To rebut your specific two points:
-more countries have battle scars from the USA than North Korea too. It more strongly indicates economic might than disregard for human life - wars cost money. Watch this space.
-Orange Baboon is in large part worrying because of the nonzero chance of making a grab for dictator-for-life. China already has one of those, who happily imprisons rivals and vanishes critics. I'd take Orange Baboon over God-emperor Xi any day.
>...are you actually trying to say that China is "morally superior" to the USA?
Well, counting people they've killed globally, wars they've started, foreign meddling, atomic bombs dropped to other nations, etc, they surely are. Call me when the numbers change..
But who cares for those when it's their side doing them?
Yeah, and China does equally disgusting things to its own citizens. Literally right now they are enacting cultural genocide on millions of their own people (see: the uyghurs)
Do you really have faith that China is going to act any differently to outsiders?
>Yeah, and China does equally disgusting things to its own citizens.
Which I'm against, but that's for them to oppose.
Several Eastern and Western European countries had dictatorships, torture, executions, etc, in 20th century as well, didn't see anybody coming from outside to fix things (and don't think their population would welcome an invasion to "fix" things -- like those "democracy bringing" wars in the middle east).
And those pointing the finger at China today, were doing "disgusting things to their own citizens" (Jim Crow? Seggregation till the 70s? western world's worst prison system, 90% titled to blacks + death penalty still in use in 2019?), or were (and are) assisting regimes doing "disgusting things to their own citizens" then and now, so there's some hypocrisy involved too.
>Do you really have faith that China is going to act any differently to outsiders?
No, but there's a player with a known track record on this matters (a bad one) and one that is yet to build one. It's kinda like what Muhammad Ali allegedly said once: “No Viet Cong ever called me n....r”. Sure, they might. But thus far...
But we have due process and most of these offenses are due to the failed drug war. Whether or not people should be locked up for drug offenses is one thing, but the laws being broken are pretty clear and generally aren't being used as punishment for expressing anti-government sentiment.
>Americans imprison and torture children at the border
This angle is pretty ridiculous. The border facilities are totally overwhelmed by an influx of people who by law are not allowed entry to the U.S. The vast majority are economic migrants, not assylum seekers. Moreover, they made the choice to come here knowing full well the illegality of their actions, and they are not being "imprisoned," they are being detained for a few weeks while the overloaded system processes their claims. Calling it torture is an exaggeration. Also the miserable conditions at the contentious facilities recently in the news are not representative of ICE facilities in general.
Only when people are not just executed by cops because they dared look too black or homeless or whatever (so many more cases of police killings that any European country when adjusted for population that it's amazing).
And even then, the due process is full of batshit-crazy aspects, from prosecutorial blackmail deals, to "three-strikes" BS.
And let's not even get into prison conditions, use of solitary confinement, rape as "joke", private prisons and prison labor, and so on...
Or the fact that it's the only (or close) western country to still have the death penalty...
And that's for official prisons. Now let's add the various "sites"...
>Only when people are not just executed by cops because they dared look too black or homeless or whatever
These atrocities are an exceedingly tiny minority out of millions of yearly police interactions.
>And even then, the due process is full of batshit-crazy aspects, from prosecutorial blackmail deals, to "three-strikes" BS
The system is not perfect, but the point is that in contrast to authoritarian China, it is generally not abused to suppress political dissent and U.S. citizens have rights to due process and representation in the legal system that do not exist in China.
Why don't you talk to the parents who are knowingly taking the risk to entry the country illegally?
>Save you speech for the children being tortured
I don't think you know what the word torture means. You've been misled by propaganda. And, again, the problems are happening limited to a select number of overwhelmed facilities.
" The severe mental suffering that officials have intentionally inflicted on these families for coercive purposes, means that these acts meet the definitions of torture under both US and international law,” said Erika Guevara-Rosas, Amnesty International’s Americas Director."
>The alleged conditions documented at the facilities follow a Homeland Security inspector general report that found "dangerous overcrowding" and unsanitary conditions at a different CBP facility in El Paso, Texas, where hundreds more migrants were being housed than the center was designed to hold.
>The El Paso Del Norte Processing Center housed as many as 900 migrant detainees earlier this month despite only having a recommended capacity for 125.
9x over capacity. What exactly are you expecting? The implication that this treatment is intentional is pure political bias. In any case these conditions are not a secret and people continue to voluntarily make the illegal journey. Who is responsible here?
That said, the actual ACLU CBP report was enlightening, detailing abuses going back to 2014 when the CBP program lowered standards during rapid expansion. It's hard to trust other modern news sources like those you provided because of their heavy political slant and propensity for sensationalism.
You're missing the point. No, it isn't "fine" to torture children, but these facilities are hugely overburdened and have procedures to follow. Aside from the corrupt agents who do mistreat migrants, it's hard to ensure humane conditions for detainees when your facilities have nine times more people to process than they're designed to handle. What's the alternative right now, release them into the U.S.?
We need to stop conflating abuses by individual agents with poor conditions caused by underprovisioning of resources due to a general increase/spike in migrant activity.
Also, there is some merit to the idea that poor conditions at these facilities act as a deterrent to voluntary migration. No one is forcing these people to come here illegally and, again, the vast majority are economic migrants, not legitimate asylum seekers.
Let me reword what you said so that I better understand it. Is this what you're saying: "Also, there is some merit to the idea to torture children, these facilities are hugely overburdened and have procedures to follow"
I think that's the gist of your argument, even if it is putting words in your mouth.
Well, at least you admit you're not arguing in good faith. That's not what I'm saying at all. Apparently the involvement of children precludes you from considering this discussion rationally.
This is an example of what I mean by lack of scrutiny or criticism. If you look at the study by CHRD [1] cited in the article you’ve linked, you’ll see that their methodology is flawed in many ways. The most worrying part is that this study in particular is the basis of many articles similar to the one you've posted.
What they did was they interviewed eight ethnic Uyghurs from eight different villages in Kashgar Prefecture in Xinjiang (the poorest and most problematic area in terms of violence and extremist ideology). They asked them questions like “how many people do you estimate live in your village” and “how many people do you think were taken from your village”. They then calculated the ratio of people detained to total inhabitants in these eight villages and extrapolated this ratio to the entire Uyghur population in the province of Xinjiang (population 23 million, many parts of which are prosperous and violence-free) in order to obtain the million+ estimate. The 1 million number (or 3 million, it’s hard to keep track of these days) is always thrown around like it’s a confirmed fact, which it decidedly is not.
Furthermore, as the article mentions, CHRD is a group based in Washington, D.C that is almost wholly funded by U.S government grants [2, p.9]. There is a clear conflict of interest here especially in the context of growing rivalry and geopolitical posturing between China and the U.S. The article also mentions Radio Free Asia, an organization literally founded by the CIA [3][4] as an anti-Communist propaganda arm. They claim they are no longer affiliated. Again, another source wholly funded by the U.S government.
Most research in America is funded by U.S. government grants at least partially if not mostly. What's your point? Am I to believe that any research agency that receives US grant money is secretly a propaganda arm of the government?
I don't really know what you expect me or the average layperson to do. I can't get on a plane and go to Xianjing, but I can read and listen to the experiences and stories of the many people that have fled the region and entrust the credibility that (certain) journalists have to do investigative journalism. The source of the article I linked to is one of many.
Also consider that you don't have to be sitting in a detainment camp to be a victim of a cultural genocide.
> Most research in America is funded by U.S. government grants at least partially if not mostly. What's your point?
My point, as I’ve stated earlier, is that there is a conflict of interest here. It's not that much of a stretch to think that the U.S government would like to delegitimize the CCP to further their own geopolitical and strategic interests.
> I don't really know what you expect me or the average layperson to do.
I expect you as a HN reader (but perhaps not the average layperson) to not view news reported in the mass media as infallible, but to scrutinize everything that is said with a high degree of criticality. Whose interests do these stories represent? Who benefits from fomenting mass outrage against an external boogeyman? Who is paying for the experiences and stories which you are consuming and why?
Knowledge of history can be of help here. One example is Rebiya Kadeer’s testimony to the U.S Congress about the Uygher situation [1] which many people accept verbatim. We can see an exact, historical parallel in the form of the Nayirah testimony [2], which was a contributor to the U.S decision to engage in the Gulf War. We all know how that turned out.
> The source of the article I linked to is one of many.
Look at the primary sources. There barely a handful of them which 99% of articles like these trace their journalism back to.
> Also consider that you don't have to be sitting in a detainment camp to be a victim of a cultural genocide.
This is true. I’d like to add that asking Uyghers to speak Mandarin in these “education camps” does not constitute cultural genocide and neither is banning elements of Wahhabi Islam which is not native to their culture in the first place. It’s a simple fact of life that if you want to do business in China or have any kind of economic opportunity, you need to learn Mandarin. This does not preclude the practice of your own native customs or languages. People think China is all Han when in fact there are many, many ethnic groups there [3] practicing their own languages and customs, and getting along just fine.
I'll end by saying that Uyghers are definitely being detained and "re-educated", I've never contested that. I'm merely questioning the scale and characterization of these camps as well as the motivations of some of the mainstream publications.
> My point, as I’ve stated earlier, is that there is a conflict of interest here. It's not that much of a stretch to think that the U.S government would like to delegitimize the CCP to further their own geopolitical and strategic interests.
And it's also not a stretch to suggest that a government that defines separatism and religion as two of it's three defined evils is committing cultural genocide on a large (millions+) section of it's country that doesn't conform to the rest of it's society.
> I expect you as a HN reader (but perhaps not the average layperson) to not view news reported in the mass media as infallible, but to scrutinize everything that is said with a high degree of criticality.
There is only so much time in the day, and today I'll spend most of mine scrutinizing the layout of virtual networking devices. In order to receive news and information about topics I am less knowledgeable about I unfortunately must entrust others with the credibility to perform this scrutiny on my behalf. The ones I entrust to do this may fail, but that doesn't mean that I'd necessarily do better than them.
Perhaps my original statement doesn't withstand scrutiny when put into certain context but I don't think it's hyperbole in this discussion.
> And it's also not a stretch to suggest that a government that defines separatism and religion as two of it's three defined evils is committing cultural genocide on a large (millions+) section of it's country that doesn't conform to the rest of it's society.
I’ve already addressed the million+ figure and cultural genocide point above. As far as the Three Evils, religion is not one of them. Religious extremism is (for example, Wahhabism and Salafism). Many religions are practiced openly in China today. There are 39,000 mosques in China, many of them paid for by the state. Separatism is considered an evil for obvious reasons. Namely, it undermines the stability of society which can lead to large scale consequences in a country of 1.4 billion that is barely four decades removed from third world status, like China.
> There is only so much time in the day, and today I'll spend most of mine scrutinizing the layout of virtual networking devices. In order to receive news and information about topics I am less knowledgeable about I unfortunately must entrust others with the credibility to perform this scrutiny on my behalf. The ones I entrust to do this may fail, but that doesn't mean that I'd necessarily do better than them.
This is all fine and understandable. But if that is the case, don’t you think it would be better not to perpetuate conjecture masquerading as facts which can’t withstand basic scrutiny?
I’ve merely said that there exists a conflict of interest due to funding sources, which is a legitimate concern in any context let alone geopolitics.
I’ve never advocated for automatically dismissing anything wherein a conflict of interest exists, only to regard it with a critical eye and a grain of salt.
It's not necessary to be outright 'authoritarian'. The 'national security' excuse goes a loooong way in the USA when the government wants something from a company. Also, financial sanctions can be equally effective for competing companies, or even just the carrot for the company that plays ball, like the alleged payments for Skype, Google, Microsoft, and others in PRISM.
This has very little to do with politics. I don't think there can be any serious debate about who is more authoritarian and who has a worse record in terms of human rights in our times. But:
- There is no guarantee that a backdoor introduced by the US will only be used by the US, or by someone with whom the US has willingly and knowingly shared knowledge about the backdoor. For all we know the MSS might be spying on some citizens using US backdoors in US chips.
- There is no guarantee that the US will always use a backdoor introduced by it for morally infallible reasons. That's a risk with any weapon, of course, and it's super easy to write it up under the cost of doing business -- right until the cost of doing business involves your own device and your own life.
- Even if you trust the checks and balances of US government completely, there is no guarantee that all governments with whom the US would share the backdoor have the same checks and balances in place.
It's not like the backdoor will ask an attacker for their US passport and a search warrant before allowing access.
There are a lot more limitations and checks on what the US can do. So yes, it's way worse if China does it. Because there not beyond imprisoning or killing people over information gained. Secondly we have no reason to believe the US had implemented backdoors in major processors.
The problem with the OP is it's whataboutism, it's a logical fallacy to say well A does something bad too, so B doing something bad is equivalent. That's a false equivalency.
>There are a lot more limitations and checks on what the US can do.
I am a German citizen living in the UK and I own a Huawei phone. A while ago the American president issued a domestic law and I still don't know whether my phone will be usable in the future or not.
I am not aware of China having any ability or willingness to do the same. As a European citizen I am accustomed to feel the effects of US politics, I can't say the same thing about China.
If anything the loss of US control in the future gives me modest hope for more alternatives and competition.
>Secondly we have no reason to believe the US had implemented backdoors in major processors.
Intel Management Engine has the capability to be a backdoor[1]. Whether the US has keys to it is in question. But it is known that Intel has included features on chips for the NSA[2].
How many US officials were tried, let alone convicted, for their role in torturing people from 2001 to 2014?
Hint: zero.
(The 11 soldiers who were torturing detainees in Abu Ghraib were not tried for torture. Only watered down charges like dereliction of duty, assault, and battery.)
> There are a lot more limitations and checks on what the US can do.
I don't care about the US promising they won't be bad. The fact that their companies can be forced to install surveillance mechanisms is already the worst case scenario for a hardware buyer. Beyond that, I don't care if it's the Americans, the Chinese or the fucking aliens doing the spying.
> There are a lot more limitations and checks on what the US can do.
These limitations and checks are written into laws. These laws can be easily ignored. Remind me as to how many of the 11 soldiers actively torturing people in Abu Gharib and Bagram were actually tried for their crimes?
>There are a lot more limitations and checks on what the US can do
If only those "limitations and checks" were actually followed by the US government and intelligence agencies. Here are some examples I collected during a previous discussion of US vs China surveillance states
>Because there not beyond imprisoning or killing people over information gained
The US government also certainly uses the information gained to kill and/or imprison people, although you could perhaps argue that the scale is not the same
>we have no reason to believe the US had implemented backdoors in major processors.
There is evidence that US intelligence agencies have backdoored products in the past. Personally, that's reason enough to believe that they may have backdoored processors as well
The United States has been executing people overseas for decades without any oversight. Predator drones during the Bush/Obama/Trump eras have only increased in use with no sign of decreasing.
If you read The Dictators Handbook, the author talks about how more democratic nations use foreign aid as a means to prop up autocratic nations in order to extract their resources (autocracy overseas to support the standard of living of those who are your selectorate).
It's like we totally forgot about the 2013 NSA spying revelations, or the 1973 CIA backed coupe in Chile, or the US overthrowing a democratically elected government during the Iranian-Contra crisis...
We should absolutely be concerned that the United States could (and does) inject backdoors to spy on other nations, because causing them to destabilize helps with resource extraction and to keep prices low, pleasing the domestic constituency.
Yeah, because the US is not "equally authoritarian", so the rest of the world shouldn't mind!
E.g. Chile where they helped establish Pinochet to power should feel secure. Italy where they collaborated with far right groups and deep state against democracy should also feel OK. And so on...
It's not OK. It's bad if either do it. So if your response to "China is bad" is not "That is bad" but "The US is bad too!", then it's a big red flag that the discussion isn't about security, but something else.
The entire discussion started not with whether "China is bad", but whether US is better:
> Hopefully this will push RISC-V manufacturing efforts in the US before the market becomes flooded with cheap open source and most likely backdoored chips manufactured in China
China oppresses its own people, while US oppresses people in Iraq, Yemen and elsewhere. US being perceived as a lesser evil doesn't make its chips more secure than Chinese chips.
Security comes from rigorous research and evidence, not paranoia or prejudice.
>Is there a way to know that the present chips are not backdoored?...
As a practical matter, that's largely irrelevant to us in the West. The world that's shaping up is a world where we won't be able to get those processors in any case.
Latin America absolutely would. Yes, there were crimes that we shouldn't forget, but the whole hemisphere is vastly wealthier than it was a century ago, owing almost exclusively to global peace, world trade, and their existence within the US sphere of influence that limited conflict to just a few hot spots.
The middle east didn't see the same benefit, and why? Because they sat in the middle of a cold war zone of contention that stoked every existing bit of strife into a hot war.
In fact I had this very discussion with a group of Africans recently. Yeah, they weren't shy about "educating" me. Some of what they believe is, in my view, not fact based. (At least, it can't be proven in court so to speak.) But, yeah, a lot of what they were saying was definitely true.
If you are worried about human rights, freedoms, you should worry about USA as well. There are enough ways to circumvent or interpret the law to suite the need, like border search and seizure few meters outside the constitutional zone applied to American citizen.
I have seen enough instances when first amendment rights are also bulldozed (still there is hope as someone emerges to preserve it). Arresting assange in UK trying to extradite him to USA, Snowden are just the tip of the iceberg.
Constitution is not absolute but a moving target with time. It requires a good faith constant effort to keep it and it's spirit alive.
I believe having a democracy does not guarantee it will remain same. The best example is India, where the current govt is bulldozing the constitutional spirit with amendments to turn it into Hindu nation and make it more like China where govt has absolute control. Anyone Who criticise the govt will be termed terrorist.
Both USA and China have their own set of good and bad, can't paint either with a broad brush.
I'm not. I'm alluding to what a government can do with this data.
This data/backdoor/power in the hands of a communist autocratic government with complete disregard for territorial and human rights is an infinite times worse, on a completely different scale, than in the hands of a democratic government.
Fair point as applied to the wider government.
But a little less convincing in the context of the Snowden leaks which also showed that this kind of data was then used to extrajudicially kidnap people of any nationality on any territory and subject them to any number of human rights abuses.
There is of course a difference in scale, I'll hand you that. But it's unclear if that difference in quantity is big enough to make it a difference I quality.
Sure, but I assume that even those who were kidnapped were, in majority, plausibly interested in terrorism or something in that regard. So even though NSA snooping you is bad, at least they're mostly interested in terrorists or I guess spying other governments, I don't know. What I do know, that in China a big reason for government to spy you is to stop you from criticizing it.
I just can't buy that China is better than US regarding human rights. Or do people actually think the opposite?
Is it whataboutism? I think it is more just an acknowledgement that there is literally no escaping this kind of thing, so it is an irrelevant concern in this context. The point isn't "other entities are already doing it, so it's ok", it's "is there any hardware for which this isn't true? Then why is that a reason to dismiss this hardware?".
US dragnet snooping has nothing particular to do with law enforcement. Nobody is allowed to know what is done with the stolen information. When we find out one use, that doesn't preclude any other use.
The use that nets the greatest return for any snooped information is extortion. Spooks are very familiar with this use, as it is their go-to method to "turn" people to recruit informants. Of course turning is only one of many uses for the same information. Coercion applied to public officials creates corruption, to judges miscarriages of justice, to police persucution of private enemies. It poisons the public well everywhere it touches.
Prosecutors are also very familiar with extortionate use of secrets. We see it in "plea bargains", where people are made to plead guilty to charges prosecutors could not prove in court, but that only scratches the surface. "Constitutional protection" is no shield against any use that doesn't introduce the facts in testimony.
Note that having "nothing to hide" doesn't help you if the secret is about somebody else, and used to coerce their actions against you.
Not sure how to best say it, but I think in the end it's better to not put any trust into any hardware- (or software-) provider anyway, it shouldn't matter whether your chips are built by a "good" or "bad" actor, just assume it's a bad actor when designing the systems around it. Maybe this leads to more robust systems overall, also for cases when there was no "evil intent" (see Spectre/Meltdown).
That's not practical. If bytes anywhere in memory can trigger a backdoor, webservers won't be possible because they have to listen to arbitrary requests.
It's maybe not practical, but it is not impossible. You can build a four-component system that does the job. You build two computers with different ISAs from vendors you assume to not share a backdoor, and let all I/O happen through a simple verifier, that copies the input for both machines, compares byte for byte their output. For cryptographic purposes, the random number generators of the machines must be linked, which would be our forth component. The software for each computer should be derived from different sources, which must follow the same I/O specification.
Just build it into an FPGA. Make it simple enough to verify. Done.
"Everything has vulnerabilities!" - Bullshit. Connect your Ethernet MII to a decoder that sets an LED if a certain pattern comes by. Can you hack it, to make it turn on when it isn't supposed to, with crafted network packets alone?
I don't get it - NSA has inserted many backdoors in different devices like routers and even hard disks, I don't recall China has such a record. Why are you so concerned about China but the U.S. ?
I don't get it - CPC has inserted many backdoors in different devices like edge routers and even cell phones, I don't recall U.S. has such a record. Why are you so concerned about U.S. but the China ?
Yeah, because China has not been run by jackwads the whole time. They're just good guys trying to understand the world. Oh, they commited a couple of genocides here and there but they have good intentions.
You have done nothing but reinforce the point. If anything, that's even more reason to make sure China is not run by jackwads in the future. I mean, what are you disagreeing with?
That the US is run by jackwads?
or that China should not be run by jackwads when it becomes the superpower?
Or are you just off on some other tangent altogether? (more whataboutism?)
Because this subthread is about the idea that jackwads in China should be able to do something simply because jackwads in the US do it. Which is a patently ridiculous notion.
This is nonsense whataboutism backed by...poor memory??
Also, just because there are concerns about Chinese back-doors does not mean the author is entirely unconcerned about US back-doors, nor does it even tangentially relate to his argument. (In fact, I think the author would agree that both are equally bad.)
At least, however, in the US there are enforceable State and Federal Constitutional protections for citizens and residents that limit the ability of the government to spy on its citizens. Even if these protections are not perfect (they are not) or sometimes abused (they are), they are objectively better than the clusterfk of privacy and rights violations the Chinese state inflicts on its citizens every day, with the consent and complicity of national technology firms.
It may not mean much if you're not a US Citizen (it shouldn't) but like it or not this is a new front of national security, and to "not get it" is to miss the point entirely.
"That makes me very concerned about using a chip completely manufactured in China."
I understand the risk, but isn't that cat "already out of the bag" for ARM licensees and other ISA implementations? Is there something particularly special about this one?
Can you explain why the RISC-V needs to happen? I am not familiar with the chip industry, so I don't understand why the RISC-V will move computing forward and x86/ARM won't move it forward.
Any industry needs competition! Sadly the consolidation of the [server] processor market over the last decade from half a dozen or more very different designs (which had collectively driven computing for the last 30+ years) has left us with the duopoly of Intel who's not really innovating and ARM who kind of are, but mostly in mobile.
Obviously, RISC-V also needs software support; gcc supports it, Golang has added RISC-V as a target (although it's not yet implemented) and there are several projects working on LLVM (e.g. lowRISC and SiFive's port). Having played around with the Fedora-RISCV port under QEMU, it's actually surprisingly good thus far.
Having an open-source processor design with good software support that anyone can manufacture without royalties is a good thing and will hopefully spur both Intel and ARM into innovation.
Removal of "percentage of revenue" based royalties is probably the big one. Others may cite the open source ISA, but I think royalties is the big deal. See Western Digital's plans for RISCV. Taking a 1.5 to 2 percent "off the top revenue/not-profit skim" away is a big deal.
Edit: Fwiw, I hope the open source ISA has secondary societal benefits. I'm not skeptical, just pragmatic. Follow the money.
Given the manufacturing profit margin, it’s unimaginable if any big player will take the lead(unless the government chime in, but wait that’s how China works!)
Nobody can afford to keep the compilers and subroutine libraries to themselves, so the more people use the ISA, the more useful it becomes. Even if you wouldn't in a million years deploy such a chip at your financial institution or on your email server; or, without a separate network, in your home.
IIRC in the wake of Spectre, Google had a security blog that noted that even at Google with all the layers of software involved they felt that there really was no way to ensure that something in those layers isn't running something unexpected / doing something it shouldn't.
IIRC The general topic was speculative execution and such security, but I think it also made it clear that more generally all the chips, software involved ... you just can't know.
To be clear, I don't say that to dispute anything you're saying, quite the opposite.
I would really liked to know what extentions they will support. Will the support the Vector extention? How about some of the once that are not quite finished yet, like Bit manipulations.
This chip exists entirely as a file on a computer, and maybe a few FPGAs, so I doubt even the company announcing this has very much idea. It seems as if they're hoping to get SMIC to manufacture this, so it won't exactly be a competitive node (although China is pouring vast amounts of resources into making their semiconductor manufacturing competitive, so give it a few years).
Although it sounds radical (16 cores!) it's a rather low end chip. So months to a year, and millions to low 10s of millions of dollars. However the Chinese govt is massively investing in semiconductors at the moment so they will regard that as valuable research and investment rather than a cost.
It's not just a normal market - China may see this as essential to its ability to develop its technology. It's Made in Chine 2025 policy.
That's taken on new urgency as the west has started cutting China off from western tech - so it may be normal companies wanting some insurance in case intel / arm cut them off (trade disputes etc) AND the govt itself wanting to product its industrial base from cutoff during trade disputes.
I don't have any special knowledge on this, but from the announcement: According to reports, Black Iron [Xuan Tie] 910 can be used to design and produce high-performance end-to-end 5G chips, artificial intelligence and automatic driving.
So I guess they want to sell it as an alternative to ARM/nVidia/Qualcomm in those roles.
What i don‘t get is how optimizations should work for Risc-V. For e.g. intel chips you can do optimizations for every generation because some things have been optimized in the architecture so you can do different tricks. But with risc-v only the instruction set is the same and every cpu could implement it differently with different performance characteristics. How should someone optimize for this?
There's two ways: Firstly different chips will have different sets of extensions.
Secondly macro-op fusion - you can optimize particular sequences of instructions in order to try to get certain optimized micro ops, and it's expected in future that different sequences will be needed on different cores. A "wrong" sequence will run slowly but still run.
Of course all of this is a PITA from our (software) point of view.
Yeah because ARM sells the complete implementation for the ARM core. But with Risc-V there may be hundreds of different implementations, you can‘t really add optimizations for everyone.
Not everyone buys the implementations, I specifically listed multiple non-Cortex ones. Cavium/Marvell ThunderX, ThunderX2 (Broadcom Vulcan), Qualcomm Centriq (press F for Falkor), Ampere eMAG (Skylark), all current Apple stuff, are all very different custom implementations.
For all my .NET developers, if you want your .NET Core app to target RISC-V ISA, you would need to use Mono AOT and to specify the LLVM target is RISC-V.
