Nvidia announces better ARM support and new ARM CPU - x86 not the main player anymore

[leadeater]

1 hour ago, Kisai said:

You are not a chip engineer, Red is not a chip engineer, I am not a chip engineer.

That's actually not that big of a deal for the discussion, you can have a well informed and educated discussion on the matter without literally being a chip engineer. As long as you stick to actual facts, data and evidence there is a foundation for a common grounds of discussion. The issues at hand is basically exactly as you describe however, opinion and gaps in knowledge being treated as ground truth which make a discussion in the realm of reality near impossible.

 

You, I or anyone else doesn't need to be a chip engineer to actually analyze a given processor architecture because there is a clear difference between designing something and being able to understand it enough to discuss it. So my actual experience is not chip design but that doesn't mean I don't actually understand it and I think you can too. I might have a little more understanding of some deeper technical aspects, or maybe not, but this having to be a chip engineer is just a fallacy. Such an argument almost always gets thrown back in ones face when the other has gotten out of their depth and cannot form a counter argument, not saying that is what you are doing here but that's how it often plays out.

 

Someone doesn't agree with someone else, they don't value the other sides experience or opinion and then just end up entirely dismissing it. I didn't go get a formal qualification, with distinction and an industry award (from Fujitsu) at my graduation, for nothing. I haven't spent years in the industry for no reason or gained zero knowledge during that time.

 

The discrediting and slow death of professionalism is nothing but disheartening, something happening across many different fields around the world with the most troublesome one in my opinion being health care. 

[Bombastinator]

29 minutes ago, leadeater said:

That's actually not that big of a deal for the discussion, you can have a well informed and educated discussion on the matter without literally being a chip engineer. As long as you stick to actual facts, data and evidence there is a foundation for a common grounds of discussion. The issues at hand is basically exactly as you describe however, opinion and gaps in knowledge being treated as ground truth which make a discussion in the realm of reality near impossible.

 

You, I or anyone else doesn't need to be a chip engineer to actually analyze a given processor architecture because there is a clear difference between designing something and being able to understand it enough to discuss it. So my actual experience is not chip design but that doesn't mean I don't actually understand it and I think you can too. I might have a little more understanding of some deeper technical aspects, or maybe not, but this having to be a chip engineer is just a fallacy. Such an argument almost always gets thrown back in ones face when the other has gotten out of their depth and cannot form a counter argument, not saying that is what you are doing here but that's how it often plays out.

 

Someone doesn't agree with someone else, they don't value the other sides experience or opinion and then just end up entirely dismissing it. I didn't go get a formal qualification, with distinction and an industry award (from Fujitsu) at my graduation, for nothing. I haven't spent years in the industry for no reason or gained zero knowledge during that time.

 

The discrediting and slow death of professionalism is nothing but disheartening, something happening across many different fields around the world with the most troublesome one in my opinion being health care. 

testing my understanding of your statement.  To use an analogy, while there is a fundamental difference between knowing how to design a gun and how to use a gun, while it may affect the nature of discussion of guns it doesn’t prevent experienced users from discussing the pros and cons of different models of gun.  The scary people are the ones who don’t even know basic gun safety and do things like point them at their foot and look down the barrels yet attempt to speak as experienced users.  and there seems to be more and more of them.

[Kisai]

1 hour ago, leadeater said:

I might have a little more understanding of some deeper technical aspects, or maybe not, but this having to be a chip engineer is just a fallacy. Such an argument almost always gets thrown back in ones face when the other has gotten out of their depth and cannot form a counter argument, not saying that is what you are doing here but that's how it often plays out.

 

It's just the circular problem on this forum and elsewhere, if two people start talking past each other, they're not listening and then you get like 5 pages of threads of just two people repeating the same thing at each other while nobody else can get a word in. We see this a lot with the crypto threads and we see this every time an apple topic comes up. This thread turned into that.

 

I consider two responses, one with the opinion, facts and sources for the opinion, and one clarification if someone responds to it. Past that point, I'm not going to keep re-arguing the same point with everyone, either I was right the first time, or I generalized something and someone tries to "Correct" the generalization rather than let it go. Past performance is not indicative of future results.

 

ARM performance in the M1 doesn't mean it will scale up to a Mac Pro, and among this conversation in the thread, really, we've all seen what Windows RT was. 

 

https://www.zdnet.com/article/microsoft-is-designing-its-own-arm-chips-for-datacenter-servers-report/

 

The desire to put ARM in servers is from a energy density point of view. Microsoft is experimenting with it, Amazon is already doing it, but they're not doing it for GPGPU. Yet.

 

Considering that many Linux VM's can be spun up on ARM chips and don't require specialized instructions (eg web servers, edge servers, email servers, dns, etc) means that purchasing large quantities of expensive x86-64 processors is unnecessary, and to be honest I would be surprised if the ARM instances had GPGPU capability (and none do that I see)

AWS Arm cpu EC2 t4g, A1, and M6g: https://aws.amazon.com/ec2/instance-types/

 

Can't get much info on Google's use, though it would seem like Google would be a likely customer for this new arm-gpgpu product since they make a big deal about deeplearning.

[igormp]

1 hour ago, Kisai said:

Amazon is already doing it, but they're not doing it for GPGPU. Yet.

Do you mean them making their own GPUs, or having ARM CPUs in their GPU instances? If the former, they have their own inferentia accelerators, and in the later they recently released it (it's in a closed preview, if your company is a big client you can request access).

 

1 hour ago, Kisai said:

Can't get much info on Google's use, though it would seem like Google would be a likely customer for this new arm-gpgpu product since they make a big deal about deeplearning.

They do have their own TPUs, but no ARM CPUs yet. Same goes for Azure.

[RedRound2]

Congratualtions for wasting my time again and you being the person I've had the msot unproductive conversation with on this forum. You keep the same damn thing by cherry picking quotes out of contexts and assuming your own version.

On 4/22/2021 at 12:20 PM, leadeater said:

Ahh no, not at all. Nowhere did I say it was "more powerful" at all. I said your sweeping statements about how much better M1 is and revolution to ARM is wrong when you exclusively reverence the performance increase from the last generation Intel Mac products to the current generation M1 Mac products.

LITERALLY WHERE? Where did I say that ARM was on top of the wolrd anywhere in this thread?? Please point it out to me. All I said that many people were never expecting ARM on laptop/desktop to be competitve. And Apple delivered more than that. That is a fact, no matter how much you want to deny it

 

Stop making shit in your head

On 4/22/2021 at 12:20 PM, leadeater said:

This issue still stands that doing that is deeply flawed, so flawed to the degree of only comparing basic sports cars to Tesla Model S P100D then saying that all petrol powered cars are slower than the Tesla. There are hundreds of cars faster than the Tesla, there aren't hundreds of CPUs faster than the M1 but the number is non zero. And for the mobile market the Ryzen 5000 U series in both 15W mode and 25W mode are vastly better than the Intel CPUs in the previous generation Intel Macs.

Again WHERE?? Why do you enjoy writing the same thing spanning multiple paragraphs. Do you really have nothing better to do?

Where did I say that ARM cores were the fastest in the world?? Please point it out to me.. Don't be a coward and just cherry pick things you can find some loophole in english language to perfectly take what i said out of context and twist it to conform your own narrative. Why are you comparing 5000 series to Intel? That's not the conversation here.

On 4/22/2021 at 12:20 PM, leadeater said:

BS numbers here because at this point I can't be bothered getting the real figures but if the M1 is 2x better than the last gen Intel MacBooks and the Ryzen Mobile 5000 series is 1.9x better then how much better is the M1 to the Ryzen Mobile 5000? Is it 2x like you've been saying? Nope.

Where did I say M1 is faster than Ryzen 5000 series by 2x?????

 

What is going on with you to make up all these random BS claims?? Seriously, I wouldn't even bother to reply to you but you really have a habit of making assumptions with some prenotional judgement you made.

On 4/22/2021 at 12:20 PM, leadeater said:

Sighhh, this is exceedingly tiring. Please stop making things up. Is it really that hard? You said M1 proved ARM can deliver powerful CPUs, I said no server ARM CPUs have already proven that. I didn't say it wasn't impressive, I've said it isn't as impressive as some people, such as you, are trying to make out. Not as != not. You seem to have consistent difficultly understanding the concept of degrees, the degree of how much better something is or isn't for example.

 

Ryzen Mobile 5000 U

 

Hopefully that was big enough for you to not miss it for like the 100th time.

You mentioned Ryzen 5000 series in this particular reply a lot. Do you think I can see your messages as you type?? That's not how this forum works

 

I've already addressed my thoughts about Ryzen in the reply above this quote. And from waht i found in my 5min google search, Ryzen 5000 series a 35W part beats the M1 is 75% of the tests But, I didnt go too deep, beacuse that system had a 35W chip, 32GB RAM, and dedicated graphic card. So the fact that M1 could keep up and beat it in some tests is itself a huge testament to M1's capabilities. Besides, can you also find me a good Ryzen 5000 series laptop in same performance and build quality as the MBA? Bcause the MBA at a 1000$ is a damn nice deal

On 4/22/2021 at 12:20 PM, leadeater said:

Your assumptions about how hard it is to scale down a server CPU doesn't make you correct. These server CPUs are in fact scaled up from reference ARM archecture with the cores themselves largely not changed at all, the one with the significant change is the A64FX which those changes are now in the ARM reference architecture. They are powerful because they have 80 cores, the same cores in your (well not your) Android phone. Yea sure they didn't put in to the SoC USB controllers or Wifi but that is only because it's not needed in a server, not because it's technically hard, not at all.

 

You're doing 1 +1 = 3, your math is wrong not mine.

Where did i do 1+1=3? I did a 1+1 is logic of why these companies didnt enter the consumer market and crush the competitors. You still cant find a reason for it? Sorry, whatever you claim is not as easy as you think it is.

On 4/22/2021 at 12:20 PM, leadeater said:

You accept that you are not well versed in the server market and server CPUs and yet I actually am, literally my job, you want to treat your assumptions as more correct than my real knowledge? I sincerely hope you don't go to WebMD rather than to a doctor.

Just because you deal with servers doesn't mean you're an expert with VLSI and chip design. Serioulsy , stop digging your own grave. I'm not and I have admitted it. It's you who claim to be a doctor and secretly searches WebMD for answers to my question, or tunr vastly more complicates things like chip design into a basic math problem

 

On 4/22/2021 at 12:29 PM, leadeater said:

So basically TL;DR still going to ignore Ryzen Mobile 5000 U series and comparing to desktop CPUs? Could you please actually listen to anything at all I've said and look at the exact series of CPUs I've pointed to you. Unless you talk about Ryzen Mobile this conversation is over.

You are refencing the same damn reply? 

On 4/22/2021 at 12:29 PM, leadeater said:

Because yes, you can do fanless Ryzen Mobile U, if one chooses to. The 15W TDP is lower than the M1 SoC package power is, or appears to be as measuring that as accurately as you can the Ryzen is not as good.

Show me some performance number. I predict fanless Ryzen 5000 series will be worse than M1, but it is a prediciton. I already wasted my time replying to you and I couldn't find anything relaibale in the short time frame I had 

On 4/22/2021 at 12:29 PM, leadeater said:

If you think I've been changing my argument or twisting words that's because I'm responding to things you bring up, unless you don't want to me actually address points you make, so if you're done making illogical leaps about things I've said then the discussion will get a whole lot simpler.

No, you're doing that

[leadeater]

15 hours ago, RedRound2 said:

Congratualtions for wasting my time again and you being the person I've had the msot unproductive conversation with on this forum. You keep the same damn thing by cherry picking quotes out of contexts and assuming your own version.

Why are you bringing this up again?

 

15 hours ago, RedRound2 said:

Where did I say M1 is faster than Ryzen 5000 series by 2x?????

Read the quoted sentence again and apply some better comprehension to it. There is a concept known as giving an exemplar to illustrate a point, my sentence literally says the numbers are not real so please stop that nonsense.

 

If you can't treat something that was directly stated as not real figures and you go on to treat it as if they were or pretend I'm claiming you said it then may I ask you have a deep think about how this conversation got to where it did. If something as basic as this is a problem then I personally suspect some other underlying issue, one I suspect being that you have made the subconscious or conscious decision to treat anything and everything I say as a lie or it's always about you or something you have said.

 

And in case it wasn't clear or you simply just did not want to take onboard what I was saying, it's an illustration of the performance differential of the competing best in class which matters if you are making the claim the M1 performance is so much better that it could drive an industry change. Sure I could agree if you only limit to looking at the Intel CPUs in the last gen Apple products but it's a simple fact that it's not as attractive when including Ryzen Mobile. The M1 is better, but not better enough to make any large changes happen due to all the other factors.

 

15 hours ago, RedRound2 said:

Where did I say that ARM was on top of the wolrd anywhere in this thread??

When you literally said it proved something yet products before it already had done so. Don't like me pointing that out? Fine, that's not my problem that is yours. If you don't actually care about those products then you're welcome to ignore them and continue on. I already said I don't blame you for not being aware of them as they just aren't applicable to you. However that doesn't change the fact that high performance ARM was on the market before hand and unless you think people in the silicon and chip design industry were not aware of those then you are dead wrong. If you think Microsoft wasn't aware then you are dead wrong, everyone that matters at all for some of things things you tried to argue knew.

 

You didn't know, so what? You could have just gone, "huh that's interesting but I don't care" but instead you wanted to argue and discredit. This conversation was unproductive due to you, I tried to have a constructive conversation but you made every attempt possible to make it not.

 

You lacked the knowledge from the outset to have this discussion that is why you think I'm cheery picking and I'll repeat anything that is correct if you try and deny or counter argue it. If you don't like something being repeated back at you either stop and A) Disagree but move on, or B) Consider what is being said might have truth behind it. I actually don't care if someone picks A or B, I care if they pick C, with C being continuing arguing in ignorance.

 

This topic is about Nvidia ARM CPUs for HPC and servers, with some on the side consumer devices for researchers and data scientist. All the ARM CPUs I talked about are directly applicable to the topic, the M1 is the least applicable.

 

FYI I pick option A, you should pick A now too.

 

15 hours ago, RedRound2 said:

Just because you deal with servers doesn't mean you're an expert with VLSI and chip design. Serioulsy , stop digging your own grave.

No, I do have enough knowledge to have the discussion. I will not simply let you pass by and discredit my education, qualifications and experience. Other people on this forum do as well, more than just a mere handful. That's a pathetic retort only made by those out of their depth, at no point did anyone need to be an employee of say TSMC/Intel to have this discussion.

 

15 hours ago, RedRound2 said:

Ryzen 5000 series a 35W part beats the M1 is 75% of the tests But, I didnt go too deep, beacuse that system had a 35W chip, 32GB RAM, and dedicated graphic card

That's because you were looking at Ryzen Mobile H not U series. Have a look at Ryzen 5800U reviews. And yes the M1 is faster, I don't remember saying the M1 wasn't faster. You wanted to have that argument not me, my point was AMD does have 15W CPUs that perform very well and therefore the power efficiency isn't as bad as one might want to think.

https://www.techspot.com/review/2235-amd-ryzen-5800u/

 

If you, as you keep doing, use Intel CPUs as the measuring bar then your argument and opinions appear far more favorable but like I have to keep pointing out there are better x86 options, much better, so any claims about market changes must (not optional) be measured against all the options on the market not ones you want to limit to because you only want to compare last gen Mac to M1 Mac. You are making claims to changes outside the Mac ecosystems so your imposed limits are irrelevant.

 

[Bombastinator]

11 hours ago, RedRound2 said:

Congratualtions for wasting my time again and you being the person I've had the msot unproductive conversation with on this forum. You keep the same damn thing by cherry picking quotes out of contexts and assuming your own version.

LITERALLY WHERE? Where did I say that ARM was on top of the wolrd anywhere in this thread?? Please point it out to me. All I said that many people were never expecting ARM on laptop/desktop to be competitve. And Apple delivered more than that. That is a fact, no matter how much you want to deny it

 

Stop making shit in your head

Again WHERE?? Why do you enjoy writing the same thing spanning multiple paragraphs. Do you really have nothing better to do?

Where did I say that ARM cores were the fastest in the world?? Please point it out to me.. Don't be a coward and just cherry pick things you can find some loophole in english language to perfectly take what i said out of context and twist it to conform your own narrative. Why are you comparing 5000 series to Intel? That's not the conversation here.

Where did I say M1 is faster than Ryzen 5000 series by 2x?????

 

What is going on with you to make up all these random BS claims?? Seriously, I wouldn't even bother to reply to you but you really have a habit of making assumptions with some prenotional judgement you made.

You mentioned Ryzen 5000 series in this particular reply a lot. Do you think I can see your messages as you type?? That's not how this forum works

 

I've already addressed my thoughts about Ryzen in the reply above this quote. And from waht i found in my 5min google search, Ryzen 5000 series a 35W part beats the M1 is 75% of the tests But, I didnt go too deep, beacuse that system had a 35W chip, 32GB RAM, and dedicated graphic card. So the fact that M1 could keep up and beat it in some tests is itself a huge testament to M1's capabilities. Besides, can you also find me a good Ryzen 5000 series laptop in same performance and build quality as the MBA? Bcause the MBA at a 1000$ is a damn nice deal

Where did i do 1+1=3? I did a 1+1 is logic of why these companies didnt enter the consumer market and crush the competitors. You still cant find a reason for it? Sorry, whatever you claim is not as easy as you think it is.

Just because you deal with servers doesn't mean you're an expert with VLSI and chip design. Serioulsy , stop digging your own grave. I'm not and I have admitted it. It's you who claim to be a doctor and secretly searches WebMD for answers to my question, or tunr vastly more complicates things like chip design into a basic math problem

 

You are refencing the same damn reply? 

Show me some performance number. I predict fanless Ryzen 5000 series will be worse than M1, but it is a prediciton. I already wasted my time replying to you and I couldn't find anything relaibale in the short time frame I had 

No, you're doing that

It may be the thread title that is what is making exaggerated claims. Specifically that x86 is “not the main player anymore”. It’s still the main player.  Apple still has a sub 15% market share. For them to be “the main player” it would have to be above 50%.  It is merely that it is now possible for that to change.  There are a lot of vague words being used and the result is they can be taken at many levels.  You refer to the M1 as “more than competitive” that can be anything from strongly competitive to absolutely dominant.  It depends on how it is being read.