New Zen microarchitecture details

[cytg111]

How would a quad with a even with a 40% IPC increase at 4ghz be a game changer? We already have that in Skylake, probably even Haswell. If you cant tell any difference between SB and skylake, you could not tell any difference with this mythical quad with a 40% IPC increase either. And if you believe AMD will bring a competitive hex core at 200.00, well, I have some land you might wish to look at. Even AMD's marketing is not that inept.

I guess it would be a game changer because now there would be TWO viable brands of quads in existence instead of just one. I guess that by different amounts of "gamechanger" your mileage may vary.

 

[Arachnotronic]

I thought she said she was done being the discount brand.

MrTeal is being oh-so-deliciously sarcastic. Love it!

 

[frozentundra123456]

Yea, all you have to do is look at the prices of Fury X and what they tried to get away with charging for the 9590 and even 7850K at launch to see that they are going to charge as much as the market will bear, just like intel. As well they should, BTW.

 

[ShintaiDK]

MrTeal is being oh-so-deliciously sarcastic. Love it!

Ohh, it was so sophisticated I didn't notice.

 

[ShintaiDK]

Yea, all you have to do is look at the prices of Fury X and what they tried to get away with charging for the 9590 and even 7850K at launch to see that they are going to charge as much as the market will bear, just like intel. As well they should, BTW.

And they originally wanted 850$ for Fury X if we are to believe rumours. But the 800$ 220W spaceheater was enough evidence.

But I like to see the product first. It may be a 200$ product because it performs like a 200$ product.

Performance/watt should hopefully improve drastically.

 

[Shehriazad]

Looks very nice...now if the frequency is anywhere NEAR 4 Ghz...this would be a nice product.

BUT my gut tells me that the 8core/16thread CPU won't sit near that(95W limit)


I mean just based off of way uneducated guesses.


4core/8thread will prolly be 4 ghz "boost"
8core/16 would likely be 3.2-3.5 ghz "boost".


I just don't see a 16 threaded monster chip to have 4 ghz @ 95W...this just does not seem likely at all.
Anyone who thinks that must be mad...there's just no way AMD has that much efficiency...because not even Intel is anywhere near that...and they got the budget to throw at such research.

 

[IntelUser2000]

I love how when Intel provides a certain level of performance/functionality, they are accused of screwing customers, but if AMD is able to provide it, it's a "game changer" and "most exciting architecture in the history of ever."

AMD = Underdog with history of lower prices
Intel = At the moment, using quad cores with minimal IPC increase for the past 4 years, and probably for another couple

Performance at a value will bring comments such as above you know? Considering the screw-up with Intel's 14nm and seeing how well the foundries are doing, I would like to see a formidable Zen chip come to the market.

Perhaps they'll be the one to do what I think Intel should have done - Core i3 with top end Iris Pro class iGPU at $200.

 

[dark zero]

Waiting how Zen performs... If they sell their Quad with HT at Core i3 prices and their non HT Quads at Celeron prices it would change everything.

 

[IntelUser2000]

Waiting how Zen performs... If they sell their Quad with HT at Core i3 prices and their non HT Quads at Celeron prices it would change everything.

The "40%" claim has a lot of leeway. It could be that 40% is from select applications that does not reflect real world workloads. Or that 40% might be a slight underestimation and we get 45% in average.

"IPC" can easily end up being worse than Sandy Bridge to threatening Kabylake parts depending on how realistic their claim of 40% is and how well they are executing.

If we get perf/clock equivalent to a Broadwell/Skylake/Kabylake then I don't believe they'll charge at such low prices(nor they should). But if they make parts at a key segment Intel does not care about entering(like the Core i3 + Iris Pro), and just outperform every part at the same price point by 5-10%, it'll be enough to make a big impact and earn good money for the company. But they do need a flagship part that will earn goodwill among people. It won't be good enough for the top part to be slightly worse than the competition.

 

[Bryf50]

It is a fact -- before BD launched, these forums were buzzing with claims that AMD was going to kick SNB into the stone age with 8 cores, more IPC than Phenom II, etc. Your join date is 2008, so I am sure you saw all of the internet buzz around Bulldozer, too.

That hype led to a lot of disappointment/angst, particularly among those who held off on buying SNB because they were told to "wait for Bulldozer."

Zen might be different, but to borrow the words of swilli89, "knowing AMD" it'll under-perform comparable Intel parts. Past performance is the best indicator of future performance and AMD has a lot to prove in CPUs.

Telling someone to "wait for Bulldozer" is only bad advice in hindsight. It made perfect sense for anyone to wait and see what Bulldozer brings if they were planning on buying a CPU at the eve of it's launch. Anyone looking to buy a CPU close to the end of this year should probably wait for Zen too.

 

[dark zero]

Considering that the BLCK OC from Skylake is locked out, some people stopped to buy it... They are waiting for Kabylake wondering if the glitch is back or going to Zen.

Now.... That 40% increase seems to be only applied on Single Thread. On the Multi Thread, I hardly see increasing more than 15%...

 

[NostaSeronx]

GMI is a HT-PCIe 4.0 Combo PHY which has a max of 32 GB/s per 16-bit/x16 link.

Primarily to elevate low yields AMD is separating the CPU and GPU dies.

Raven Ridge is literally;
One Summit Ridge die with an active GMI link 25-32 GB/s. (8c/16t)
One mainstream Polaris dGPU die with an active GMI link 25-32 GB/s. (2048 ALUs)

 

[AtenRa]

Personally i would be fine with a 6 Core 12 Threads Excavator IPC at 14nm power consumption for $200

With DX-12, I would really want to spend more on the GPU than the CPU. Hell, i even believe a 4.5GHz 8-Core Vishera will be more than fine for the next 1-2 years of DX-12 games.

 

[IntelUser2000]

Now.... That 40% increase seems to be only applied on Single Thread. On the Multi Thread, I hardly see increasing more than 15%...

That's what all of us want. Atom is close to low end Core chips in Multi-threading. It's in Single Thread that its hopelessly behind. We consider it slow based on the Single Thread factor.

And single thread performance helps multi-thread performance too. Not that it will lack multi-thread performance thanks to SMT if they keep their promise of 40%(hopefully more).

With DX-12, I would really want to spend more on the GPU than the CPU. Hell, i even believe a 4.5GHz 8-Core Vishera will be more than fine for the next 1-2 years of DX-12 games.

VR might mess that up. If it gets popular, its requirement of 90 fps or greater starts differentiating CPUs pretty easily.

 

[Burpo]

Thanks Dresdenboy.. Always appreciate what you bring us

 

[USER8000]

Thanks OP for the digging and giving us more information about the uarch as we have only had dribs and drabs so far.

Lol, this reminds me of what the AMD fans were saying on the eve of the Derpdozer launch.

Since you know its going to be a failure,I assume you won't be posting in any more Zen threads until launch?? Let all the fools get hyped up while you are snug in your realisation of how wrong they are in reality! :thumbsup:

 

[Arachnotronic]

Thanks OP for the digging and giving us more information about the uarch as we have only had dribs and drabs so far.



Since you know its going to be a failure,I assume you won't be posting in any more Zen threads until launch?? Let all the fools get hyped up while you are snug in your realisation of how wrong they are in reality! :thumbsup:

Didn't say it'll be a failure, just saying that expectations are unrealistic. Let's wait and see what SKUs AMD brings out & what kind of performance Zen brings before wetting ourselves, yeah?

 

[USER8000]

Didn't say it'll be a failure, just saying that expectations are unrealistic. Let's wait and see what SKUs AMD brings out & what kind of performance Zen brings before wetting ourselves, yeah?

Dude,nobody is wetting themselves though. Even the guy you quoted was thinking it would be closer to SB. I just think you are just being overly negative for the sake of it.

 

[Arachnotronic]

Dude,nobody is wetting themselves though. Even the guy you quoted was thinking it would be closer to SB. I just think you are just being overly negative for the sake of it.

I am generally a skeptical person.

 

[USER8000]

I am generally a skeptical person.

Well,considering that every release from Intel in the last few years has been also hyped to come within a year of the last chip being released(people thought IB would replace SB in a year and Haswell would replace IB within a year and so on),being game changing,etc - you might need to go into those threads and give them a dose of your scepticism too. It happens all round on tech forums at every tech release like going back for the last 20 years,whether it is from ATI,AMD,Nvidia or Intel.

I can still remember the hype for non-AMD/ATI disasters like Prescott,various versions of Atom,Nvidia FX,Fermi,etc.

Its what enthusiasts do. Look at things like cameras and cars - the same thing.

Edit to post.

Basically most tech released today - latest Samsung or Android phone,etc.

"Revolutionary","will change your life","ground braking" and all the marketing bumpf they use.

If someone is not saying it,it seems to indicate you have no faith in your product apparently.

 

[swilli89]

Lol, this reminds me of what the AMD fans were saying on the eve of the Derpdozer launch.

And reminds me of the hype before the 10ghz capable netburst Pentium 4. What did we used to call that back in the day? PressHot? or something like that. Ah.. good times.

 

[Arachnotronic]

And reminds me of the hype before the 10ghz capable netburst Pentium 4. What did we used to call that back in the day? PressHot? or something like that. Ah.. good times.

Prescott was a piece of crap. But it's interesting that you basically have to go back 12 years to find an instance of Intel truly falling flat on its face in the PC MPU market.

 

[swilli89]

How would a quad with a even with a 40% IPC increase at 4ghz be a game changer? We already have that in Skylake, probably even Haswell. If you cant tell any difference between SB and skylake, you could not tell any difference with this mythical quad with a 40% IPC increase either. And if you believe AMD will bring a competitive hex core at 200.00, well, I have some land you might wish to look at. Even AMD's marketing is not that inept.

I merely said a 6C @ 200 would be awesome. Are you a marketing analyst that understands AMD's long term plans? Yes they may very well sell one at $200 to regain marketshare quickly. Maybe its worth it to them to take a margin cut to increase unit volume sold, WHO KNOWS. And yes a Skylake 4c level CPU unlocked thats 30% cheaper than what's available is a game changer.

A lot of people throwing shade in this thread for some reason. Like they absolutely care so much about their fellow human beings getting their hopes up they are rushing in to save everyone from the hype. Why not let the hopefuls hope? Does it hurt you? Isn't this entire thing about enthusiasts? Maybe everyone here but me is some kind if IT investor and that's why they are so bullish on anything prospective out of AMD. Again.. who knows.

 

[Sweepr]

Didn't say it'll be a failure, just saying that expectations are unrealistic. Let's wait and see what SKUs AMD brings out & what kind of performance Zen brings before wetting ourselves, yeah?

The message here is, after many successive failures we are still not allowed to question the insane hype surrounding any AMD launch. Either you agree with it, or you're a fanboy.

 

[monstercameron]

nice thread derail guys, no AMD thread is complete without it.
@dresdenboy at a high level, how does zen -so far- compare with *dozer uarch