Question Zen 6 Speculation Thread

[Win2012R2]

Ok, so it follows Zen 5 gets away with 1T 4 wide decoder and does pretty decent job at it, and Lion Cove got 8 wide and is hardly better - what gives?

 

[adroc_thurston]

Ok, so it follows Zen 5 gets away with 1T 4 wide decoder and does pretty decent job at it, and Lion Cove got 8 wide and is hardly better - what gives?

Because decode is hardly the limit for x86 cores. Any cores really.
Now, caches, reordering capacity, branch prediction, memory op capability, that's the stuff that matters.
Zen5 is still weirdly skinny by modern standards. In some places it's smaller than Skymont which is funny.

 

[Win2012R2]

Because decode is hardly the limit for x86 cores.

Ok, why did they make Lion Cove with 10 wide decoders, I thought going that wide was pretty expensive transistor wise?

 

[yuri69]

Ok, why did they make Lion Cove with 10 wide decoders, I thought going that wide was pretty expensive transistor wise?

AMD invests a bit more transistors in uOp cache. Intel invests in decoders.

 

[adroc_thurston]

Ok, why did they make Lion Cove with 10 wide decoders, I thought going that wide was pretty expensive transistor wise?

It's 8-wide but IDC is incompetent so whatever they do matters not.

 

[Win2012R2]

Great, so if decoding variable length x86 ops is not a problem, then ARM's main argument about x86 in-efficiency is gone?

 

[adroc_thurston]

Great, so if decoding variable length x86 ops is not a problem, then ARM's main argument about x86 in-efficiency is gone?

Aa64 is still a far cleaner ISA so not really?

 

[StefanR5R]

Great, so if decoding variable length x86 ops is not a problem, then ARM's main argument about x86 in-efficiency is gone?

It is not "not a problem", as it makes decoders more complex and has to be alleviated with big op caches. However, variable lengths of x86 ops isn't as big as a problem to make 2025 the year in which AARCH64 made x86-64 obsolete. Or x86-32 even.

 

[Meteor Late]

With Intel failing so badly why should AMD try hard with Zen 6?

Because there is ARM coming.
Yeah you might think it will never take off on Windows or in Servers whatever, but it's just a matter of performance per year. AMD cannot afford to, for example, only increase performance by 10% every 2 years when ARM increases performance by 20% in 2 years at minimum. This is one of the reasons why AMD will not sit still, x86 had a huge advantage against ARM 10 years ago, it's not the case anymore, sitting still is not an option.

 

[yuri69]

Because there is ARM coming.

ARM has been coming for a while. CSPs are pushing their in-house ARMs a lot. Nowadays, it doesn't make sense to run majority of AWS services on anything else than the 'g'-suffixed instances. This cuts the share from both AMD and Intel.

Both x86 vendors must offer some competitive advantage vs regular ARMs - performance and/or features.

Sure, it's not a problem to start an ARM project, like K12. But there are still no leaks about a high-performance ARM core being prepared by AMD nor Intel. This means there are no ARMs coming in 3-5 years time span.

Besides, Intel server chips are closer than they have been for a while with multiple upcoming chips coming. Client is still totally owned by Intel.

 

[JustViewing]

Great, so if decoding variable length x86 ops is not a problem, then ARM's main argument about x86 in-efficiency is gone?

It is not just decoding, the instruction set architecture also matter. Since ARM have more general purpose registers, there are more opportunities to execute code in parallel (Register rename is only a partial solution). Thus more wider designs are possible.
Another thing, many ARM processors also have specialized accelerators.

 

[adroc_thurston]

when ARM increases performance by 20% in 2 years at minimum.

See there's a problem in this statement.

CSPs are pushing their in-house ARMs a lot. Nowadays, it doesn't make sense to run majority of AWS services on anything else than the 'g'-suffixed instances. This cuts the share from both AMD and Intel.

It's not in-house. It's Neoverse. Aka ARM subsidizing demand by doing all the hard work and taking no money in return.
Which would end the moment ARM itself ships merchant server Si.

Since ARM have more general purpose registers, there are more opportunities to execute code in parallel (Register rename is only a partial solution).

Mega irrelevant meme.

Another thing, many ARM processors also have specialized accelerators.

Tf does that mean?

Thus more wider designs are possible.

Royal Core was wider than literally any design shipped evah by anyone and that was still good ol' smelly x86.

 

[Meteor Late]

ARM has generally been able to get that increase, if it stops now it's another thing, AMD will for sure not be banking on that, it would be irresponsible.

 

[JustViewing]

Mega irrelevant meme.

Care to explain?

 

[adroc_thurston]

Care to explain?

16GPRs haven't been the limit like ever.

 

[JustViewing]

16GPRs haven't been the limit like ever.

It is, if you do assembly programming you will know.

 

[adroc_thurston]

ARM has generally been able to get that increase, if it stops now it's another thing, AMD will for sure not be banking on that, it would be irresponsible.

There's a problem in this statement.

It is, if you do assembly programming you will know.

compiler explorer is a wonderful thing you should use instead of poasting this.

 

[MS_AT]

16GPRs haven't been the limit like ever.

Well, I am limited by them in my SIMD code Or maybe not strictly limited but if I had more GPRs I could use faster loads/store (without complex addressing). Another thing is that many people disable frame pointer making debugging harder to get one more usable GPR so while adding more of them will not make x64 beat M cores it will for sure make some code faster and lives easier as there no longer will be a need to omit frame pointers.

 

[Hulk]

Because there is ARM coming.
Yeah you might think it will never take off on Windows or in Servers whatever, but it's just a matter of performance per year. AMD cannot afford to, for example, only increase performance by 10% every 2 years when ARM increases performance by 20% in 2 years at minimum. This is one of the reasons why AMD will not sit still, x86 had a huge advantage against ARM 10 years ago, it's not the case anymore, sitting still is not an option.

Here's my take on that, which is a great point BTW.

Unless there is a massive paradigm shift, Windows will continue to be the ubiquitous operating system for businesses, gamers, and most others for the forseeable future. Windows is x86. Until ARM can outperform Windows in Windows I don't see much of a threat for Intel or AMD in that department.

The only reason ARM exists at all is because it was well-suited for the phone craze, which started in 2007. ARM has tried to "push up" ARM while Wintel has tried to "push down" x86. While there is some crossover in the tablet zone, for the most part ARM owns phones (outside of Apple of course) and x86 owns the desktop and server space.

By paradigm shift I mean something one hardware can do that the other simply cannot. For example, smart phones could simply not run on x86 when they were developed. It had to be something else. Boom, market opening, paradigm shift.

Now, with all that being said I do see what you are saying and AMD/Intel cannot stand still. One scenario could be ARM gets so good that it is competitive with x86 in Windows even in emulation mode. If that should occur and Windows ported over to ARM, and many software manufacturers followed suit that would be the end of x86. So yes that is a possibility, x86 does need to keep some distance but it has the advantage of running natively against ARM doing emulation.

Also, games are a huge part of the x86 stronghold. There is a massive amount of inertia to overcome in that space.

The real question is why Microsoft doesn't port Windows over to ARM now. Bring along their applications and right off the bat those systems would be good to go for many business users. That would be a SERIOUS threat to AMD and Intel. As long as MS staying in their (x86) lane AMD/Intel will remain relatively insulated. Snapdragon gave it a go with Windows/ARM laptops and while they "work" there are too many compatibility problems so most people are passing. I gave it a hard look and moved on.

Looking at my use case I'm actually seeing a need for more MT performance with my 9950X than ST performance. It's kind of unbelievable. The following applications slam all cores on my 9950X. Ultimate Vocal Remover, Handbrake, Vegas Pro, Topaz Photo and Video AI, and some others I think that don't come to mind right now. The point is I'm not wanting so much for ST as for more cores. 24 cores with Zen 6 would be most welcome for me.

 

[igor_kavinski]

AMD has the capability to actually cause massive damage to Intel's marketshare thanks to their in-house unreleased ARM design. What if they release their ARM SoC with even just a single Zen6c core for accelerating the emulation of legacy x86 code? Outside of outdated multithreaded software, that will cover almost all x86 applications that are single threaded or poorly threaded while the popular software will just get rewritten for the newer ARM cores.

If this happens and thanks to Intel's stupidity in not doing enough to popularize AVX-512, more than 50% laptop marketshare could get taken by AMD's hybrid ARM/Zen6c solution to the Wintel monopoly. And no one else would be able to challenge AMD without a proper x86 license.

The real question is why Microsoft doesn't port Windows over to ARM now. Bring along their applications and right off the bat those systems would be good to go for many business users. That would be a SERIOUS threat to AMD and Intel.

ACTUALLY, https://windowsonarm.org/?status=0

Almost 50% of the listed software is NATIVE already. And AMD has nothing to fear because they are working on their ARM design too. It's Intel that stands to lose.

 

[adroc_thurston]

AMD has the capability to actually cause massive damage to Intel's marketshare thanks to their in-house unreleased ARM design. What if they release their ARM SoC with even just a single Zen6c core for accelerating the emulation of legacy x86 code? Outside of outdated multithreaded software, that will cover almost all x86 applications that are single threaded or poorly threaded while the popular software will just get rewritten for the newer ARM cores.

No one wants ARM.

 

[igor_kavinski]

No one wants ARM.

No one knowledgeable. The sheep don't care which is what the majority are. Remember M$ Office having both 32 bit and 64-bit setup files but most people being content with the default 32-bits because they didn't know any better? ARM will sell once it gets to a reasonable price point and the software compatibility issues will become secondary once most popular software is recompiled. If this weren't a possibility, AMD wouldn't have expended resources preparing for it.

 

[adroc_thurston]

No one knowledgeable. The sheep don't care which is what the majority are. Remember M$ Office having both 32 bit and 64-bit setup files but most people being content with the default 32-bits because they didn't know any better? ARM will sell once it gets to a reasonable price point and the software compatibility issues will become secondary once most popular software is recompiled. If this weren't a possibility, AMD wouldn't have expended resources preparing for it.

Again, LNL is a proof that all you need is a good SOC.

 

[SiliconFly]

.

 

[igor_kavinski]

Again, LNL is a proof that all you need is a good SOC.

Which AMD may have with Sound Wave APU.