News Intel fires back at AMD's AI benchmarks, shares results claiming current-gen Xeon chips are faster at AI than next-gen EPYC Turin

[Admin]

Intel fired back at AMD's recent AI benchmarks that claimed AMD's EPYC Turin is 5.4 times faster than Intel's Xeon chips in AI workloads. Intel claims its existing chips are faster than AMD's upcoming flagship.

Intel fires back at AMD's AI benchmarks, shares results claiming current-gen Xeon chips are faster at AI than next-gen EPYC Turin : Read more

 

[peachpuff]

The ai benchmark wars have begun...

 

[redgarl]

I believe in Intel Power Points...

/SARCASM

 

[vanadiel007]

But can it run Crysis?

 

[Makaveli]

But can it run Crysis?

It can and its glorious!

 

[Tech0000]

I think Paul is right in his analysis of the discrepancy: "If this benchmark represents true performance, the likely disparity here is Intel's support for AMX (Advanced Matrix Extensions) math extensions. These math functions boost performance in AI workloads tremendously, and it isn't clear if AMD employed AMX when it tested Intel's chip."

The hugely dominating computational and data complexity in AI is the matrix multiplications that are performed repeatedly in both forward and backwards passes. Any direct silicon implementation of matrix multiplications reducing computational latency, increasing computational throughput, AND reduces data bandwidth requirement and latency, will greatly improve the AI performance. AMX does that and not by a small amount either. Given the fundamental laws in complexity theory, if you apply AMX machine instructions intelligently, there is no theoretical possibility a scalar or vector based solution can compete (all else equal # CPUs, #cores, etc, etc). If you want to dive deeper (i.e. completely geek out) into the understanding of AMX and how it can be applied to AI, I suggest reading Intel 64 and IA-32 Architectures Optimization Reference Manual: Volume 1 chapter 20.

 

[TinyFatMan]

Who remembers the “war” between Lotus 1-2-3 (Lotus Software) and Quattro Pro (Borland)?

And even more, who remembers the winner? A clue. You probably use it today. Long live the monopolies.
Smile

Intel vs AMD… And ARM Meanwhile…
Intel, AMD (and Microsoft) continuing to ignore 4 billion consumers, smartphone users…

Just because you have trillions doesn't mean you're more intelligent than everyone else...

We must not confuse “brilliance/giftedness”, which can be “measured”, finally, we must take these measurements with a grain of salt, with intelligence, which is too complex to be measured and will remain so. Intelligence being everything that makes an organism a living being.

 

[rluker5]

Since Intel is providing specific details into it's testing methods I find them more believable than some unverifiable claims made by AMD.

The differences in claims are huge. It would be like one brand of potato chips saying their new and improved one ounce bag has four times the weight in potato chips as their competitors one ounce bag and thus is four times the value when in fact they both have one ounce bags.

 

[rluker5]

Who remembers the “war” between Lotus 1-2-3 (Lotus Software) and Quattro Pro (Borland)?

And even more, who remembers the winner? A clue. You probably use it today. Long live the monopolies.
Smile

Intel vs AMD… And ARM Meanwhile…
Intel, AMD (and Microsoft) continuing to ignore 4 billion consumers, smartphone users…

Just because you have trillions doesn't mean you're more intelligent than everyone else...

We must not confuse “brilliance/giftedness”, which can be “measured”, finally, we must take these measurements with a grain of salt, with intelligence, which is too complex to be measured and will remain so. Intelligence being everything that makes an organism a living being.

I have an Intel CPU smartphone. Intel did try that market when they were giving out the 2w atoms and they were in all of those cheap tablets. If you look at the jumps Intel is making in efficiency in their mobile line, there might be a return to them. Hopefully running Windows. I have wanted a real Windows phone for so long, Windows Mobile was just a tease.

 

[Avro Arrow]

Intel is claiming that Xeon is superior to EPYC?

Yeah, I'll believe that when I see it (and I don't mean seeing it from Intel).

 

[purposelycryptic]

Who remembers the “war” between Lotus 1-2-3 (Lotus Software) and Quattro Pro (Borland)?

And even more, who remembers the winner? A clue. You probably use it today. Long live the monopolies.
Smile

Intel vs AMD… And ARM Meanwhile…
Intel, AMD (and Microsoft) continuing to ignore 4 billion consumers, smartphone users…

Just because you have trillions doesn't mean you're more intelligent than everyone else...

We must not confuse “brilliance/giftedness”, which can be “measured”, finally, we must take these measurements with a grain of salt, with intelligence, which is too complex to be measured and will remain so. Intelligence being everything that makes an organism a living being.

AMD and Intel specialize in x86 processors, which don't really lend themselves to phones. This is fine. In a world with finite resources, you have to make choices on what to allocate your resources towards, or risk spreading them too thin. In big business, a jack of all trades, master of none is going to lose to masters in their individual fields.

ARM doesn't make or sell processors, they only design them, as well as other chips, and develop and sell software and programming tools. They aren't really comparable to AMD or Intel: ARM just designs and licenses out IP, which their licensees then use as a base to manufacture processors and SoCs. There is plenty of competition worldwide in the field of ARM processors and SoCs.

ARM (the architecture, not the company) processors compete with x86 processors in the server space, and companies like Qualcomm are currently trying to make inroads in the Windows laptop space, as a result of Apple having moved their Mac OS computers entirely to the ARM architecture.

Microsoft's repeated attempts at the same over the years met with abject failure, so it will be interesting to see the results this time; both Microsoft and Qualcomm have already invested significant amounts of money in this attempt, as have various OEMs and several other ARM processor manufacturers.

Unlike with Apple, the ARM architecture will fracture the Windows software ecosystem if successful, with likely significant negative repercussions to the overall perceived stability and reliability of the Windows platform, as a PC can no longer be relied upon to "just work" with software, due to the split architecture.

Emulation offers a band-aid solution for the time being, but for the average consumer, and especially the large SMB market, who don't fully understand the difference and simply need their software to work, which for many companies can be positively ancient and barely able to run on modern x86 systems, this is likely to result in significant frustration which, if not handled just right, may collapse majority opinion on Windows ARM machines, resulting in a death spiral they cannot recover from.

Which, honestly, is likely to be the best result for both end users and developers - the power efficiency benefits ARM offers can't really outweigh the chaos adding a second system architecture to the Windows ecosystem will result in.

Obviously, Microsoft, Qualcomm, and everybody else who has invested significant money and resources in the idea is likely to disagree. Just one of the reasons I've sold my stock in them before the launch.

 

[Avro Arrow]

Since Intel is providing specific details into it's testing methods I find them more believable than some unverifiable claims made by AMD.

Considering the marketing diarrhea that Intel has produced lately, I wouldn't believe a word they say:

I mean, come on... They might appear to be more transparent but they've been quite unhinged lately because over the past 30 years, they have only been the "also-ran" CPU producer for a year or two when the first Athlon 64 came out and they were never threatened in the data centre space by Opteron the way that EPYC has just been so dominant.

They don't know how to handle not being on top so beware what words you choose to believe from them.

The differences in claims are huge. It would be like one brand of potato chips saying their new and improved one ounce bag has four times the weight in potato chips as their competitors one ounce bag and thus is four times the value when in fact they both have one ounce bags.

I look at it this way... AMD has been spanking Intel all over the place in every sector of the industry and EPYC has been spanking Xeon since it was introduced, relegating Xeon to almost not existing.

I find the idea that a Xeon could out-perform a 128-core/256-thread Turin-based EPYC CPU in anything to be extremely far-fetched. If it turns out that it's true, then great, tech advances and competition continues.

I just don't see it being true based on everything else we've been seeing.

 

[purposelycryptic]

Intel is claiming that Xeon is superior to EPYC?

Yeah, I'll believe that when I see it (and I don't mean seeing it from Intel).

...in specific AI applications, according to Intel.

So yeah, definitively take that information with a big salt chaser.

I keep wanting to invest in Intel just because their stock is so damned low right now, but, at the same time, I know it's low for a reason, and they have yet to really show me anything to convince me it won't fall further.

 

[Avro Arrow]

...in specific AI applications, according to Intel.

So yeah, definitively take that information with a big salt chaser.

I keep wanting to invest in Intel just because their stock is so damned low right now, but, at the same time, I know it's low for a reason, and they have yet to really show me anything to convince me it won't fall further.

Yeah, they have a terrible corporate culture there. They tried copying AMD's tactics by hiring Jim Keller after the massive success of Ryzen. I don't think that Jim stayed there for even three months which means that he hated it there because they had to have been paying him a fortune.

Jim hasn't said anything about his time there, probably because he's too classy to drag them through the mud. I don't think that Intel is a good investment, they're just trying to pretend that they are. It's the Intel corporate culture that resulted in their stock being so low.

If you want to invest, nVidia is definitely the safest and AMD has the most room to grow because of their dominant CPU presence and significant GPU presence. Only AMD could create the Frontier and El Capitan supercomputers on their own.

 

[jlake3]

Since Intel is providing specific details into it's testing methods I find them more believable than some unverifiable claims made by AMD.

The differences in claims are huge. It would be like one brand of potato chips saying their new and improved one ounce bag has four times the weight in potato chips as their competitors one ounce bag and thus is four times the value when in fact they both have one ounce bags.

I mean... I feel like both of them are somewhat suspect? Intel might be transparent, but not being involved in the AI space I'm not sure how universally supported their open-source plugin is and if things being compiled with AMX is the norm or not. If you're buying this kind of hardware at scale you can almost certainly afford to do that optimizing, but it's not clear that AMD was optimizing the test to their hardware the same way Intel was.

I also don't see either of them saying how much power they used. Intel is getting a reputation for brute-forcing benchmark wins at disproportionate power. (Edit: Specifically in consumer desktop, at least. Less sure about Xeon)

Also Zen5 "Turin" looks like it supports more PCIe lanes and more memory channels per socket than 5th-gen "Emerald Rapids". If you can run your project on accellerators, that would be an advantage to AMD. If you have an absolute behemoth of a project that can't fit on accellerators and needs all the memory and bandwidth you can throw at it, that might also give AMD an advantage.

Just because Intel comes out ahead at their tested conditions and is transparent about what those are doesn't necessarilly mean they've got the better AI chip all around.

 

[rluker5]

I mean... I feel like both of them are somewhat suspect? Intel might be transparent, but not being involved in the AI space I'm not sure how universally supported their open-source plugin is and if things being compiled with AMX is the norm or not. If you're buying this kind of hardware at scale you can almost certainly afford to do that optimizing, but it's not clear that AMD was optimizing the test to their hardware the same way Intel was.

I also don't see either of them saying how much power they used. Intel is getting a reputation for brute-forcing benchmark wins at disproportionate power. (Edit: Specifically in consumer desktop, at least. Less sure about Xeon)

Also Zen5 "Turin" looks like it supports more PCIe lanes and more memory channels per socket than 5th-gen "Emerald Rapids". If you can run your project on accellerators, that would be an advantage to AMD. If you have an absolute behemoth of a project that can't fit on accellerators and needs all the memory and bandwidth you can throw at it, that might also give AMD an advantage.

Just because Intel comes out ahead at their tested conditions and is transparent about what those are doesn't necessarilly mean they've got the better AI chip all around.

I'm sure next gen Epyc significantly beats current gen Intel in many if not most applications. They probably currently just have a disadvantage in use cases where Intel's accelerators are effectively used.
AMD should have chosen one of their many actual advantages to tout. They probably have more than time to show off in a single presentation.

Is AMD trying to discredit their reputation with the Computex keynote? This was far from the only misleading statement they made during it. Mind you it doesn't change the performance of their stuff either way.

 

[cyrusfox]

Yeah, they have a terrible corporate culture there. They tried copying AMD's tactics by hiring Jim Keller after the massive success of Ryzen. I don't think that Jim stayed there for even three months which means that he hated it there because they had to have been paying him a fortune.

Jim joined Intel in 4/2018 and left the company 6/2020. The reasons he left I believe have only been hinted at in a couple longstream videos he has done with Ian Cutress (

View: https://www.youtube.com/watch?v=fOUB73dZXEo
&

View: https://www.youtube.com/watch?v=AFVDZeg4RVY
), what I recall is he had great CEO support but huge organizational friction to make the necessary changes.

The chips coming out now, especially those outsourced to TSMC [Lunar lake] are believed to be influenced by his time at Intel. Intel is starting to get interesting, especially now that chiplet designs are coming to market.

 

[jkflipflop98]

Jim Keller jumps from job to job. He only stays in one place for 2-4 years then he bounces out. I found him to be highly over rated.

 

[NeoMorpheus]

1- Lets blindly believe a word from Intel.

2- Brave yourself for the Intel white knights.

 

[Kondamin]

Bah bogging down general use chips with “ai” optimisations.
Waste of silicon

 

[TheHerald]

Considering the marketing diarrhea that Intel has produced lately, I wouldn't believe a word they say:

I mean, come on... They might appear to be more transparent but they've been quite unhinged lately because over the past 30 years, they have only been the "also-ran" CPU producer for a year or two when the first Athlon 64 came out and they were never threatened in the data centre space by Opteron the way that EPYC has just been so dominant.

They don't know how to handle not being on top so beware what words you choose to believe from them.

I look at it this way... AMD has been spanking Intel all over the place in every sector of the industry and EPYC has been spanking Xeon since it was introduced, relegating Xeon to almost not existing.

I find the idea that a Xeon could out-perform a 128-core/256-thread Turin-based EPYC CPU in anything to be extremely far-fetched. If it turns out that it's true, then great, tech advances and competition continues.

I just don't see it being true based on everything else we've been seeing.

And since we know fully well how unbiased of an individual you are, I'd take your word over Intel's. Lol

 

[rluker5]

1- Lets blindly believe a word from Intel.

2- Brave yourself for the Intel white knights.

Intel left the instructions to replicate the results they got, if you have the hardware to test so it isn't blind trust, it is a verifiable statement. And it is in conflict with the AMD statement that you do have to blindly trust because they don't give details on how to replicate the results.

 

[NinoPino]

ARM doesn't make or sell processors, they only design them, as well as other chips, and develop and sell software and programming tools. They aren't really comparable to AMD or Intel: ARM just designs and licenses out IP, which their licensees then use as a base to manufacture processors and SoCs. There is plenty of competition worldwide in the field of ARM processors and SoCs.

ARM also sell ready to bake designs, so the difference from ARM and AMD/Intel in the CPU space is small from a technical pov.

Unlike with Apple, the ARM architecture will fracture the Windows software ecosystem if successful, with likely significant negative repercussions to the overall perceived stability and reliability of the Windows platform, as a PC can no longer be relied upon to "just work" with software, due to the split architecture.

Windows PCs never "just works", and in the last years the perception of "easy of use" and reliability is becoming more and more negative.
The problem is not ARM and will not be ARM.
ARM can be a chance for Microsoft to refresh the Windows image and leave the mistakes of the past relegated to x86 legacies.
But I bet they will miss it.

Emulation offers a band-aid solution for the time being, but for the average consumer, and especially the large SMB market, who don't fully understand the difference and simply need their software to work, which for many companies can be positively ancient and barely able to run on modern x86 systems, this is likely to result in significant frustration which, if not handled just right, may collapse majority opinion on Windows ARM machines, resulting in a death spiral they cannot recover from.

With modern emulators the problem is not relevant in my opinion. Games are limited by GPUs, a loooot of softwares are limited by user capacity and not by the CPU, another batch can leverage multiprocessing, so increasing cores can solve the problem, and the remaining, in a couple of years will be converted or outpaced from hardware advancements.

Which, honestly, is likely to be the best result for both end users and developers - the power efficiency benefits ARM offers can't really outweigh the chaos adding a second system architecture to the Windows ecosystem will result in.

Chaos is more relevant for devs and IT guys that end users. Users will benefit from light and efficient machines without any issue (if Microsoft works well).

Obviously, Microsoft, Qualcomm, and everybody else who has invested significant money and resources in the idea is likely to disagree. Just one of the reasons I've sold my stock in them before the launch.

I also disagree but I haven't got any stock.😀

 

[NeoMorpheus]

Intel left the instructions to replicate the results they got, if you have the hardware to test so it isn't blind trust, it is a verifiable statement. And it is in conflict with the AMD statement that you do have to blindly trust because they don't give details on how to replicate the results.

Thanks for validating both points i made.

 

[CmdrShepard]

Considering the marketing diarrhea that Intel has produced lately, I wouldn't believe a word they say:

But you expect us to believe some emo hippie on YouTube? Give me a break.

I look at it this way... AMD has been spanking Intel all over the place in every sector of the industry and EPYC has been spanking Xeon since it was introduced, relegating Xeon to almost not existing.

Last time I checked, Intel CPUs were still the top results in all relevant charts from performance to sales. Also, Intel launched into video card business and is doing very good considering how long both AMD and NVIDIA have been entrenched there as the only contenders. In other words, your spanking is wishful thinking.

I find the idea that a Xeon could out-perform a 128-core/256-thread Turin-based EPYC CPU in anything to be extremely far-fetched. If it turns out that it's true, then great, tech advances and competition continues.

4th Gen Intel Xeon Scalable processors running Intel AMX can perform 2,048 INT8 operations per cycle and 1,024 BF16 operations per cycle.

For an overview of the tech look here:
https://www.intel.com/content/www/u...ines/advanced-matrix-extensions/overview.html

It is already supported as an extension for PyTorch so it's already possible to use it and verify the results.