[Techpowerup] AMD "Zen" CPU Prototypes Tested, "Meet all Expectations"

[Phynaz]

"I read on the Internet that this guy knows a guy who knows a guy who said..."

The ADF is desperate.

 

[cytg111]

If Zen was only 10% slower than Intel's best but cost half as much, I'd buy Zen. Ok fine, no I wouldn't. I'm dying to see benchmarks. This is going to be one hell of a release day. I can't wait to read all about it over a massive cup of coffee and see the nukes start to drop in the forums.

+1. I hope it hit targets for this reason alone.
Oh it would be a good day at the zoo.. only one thing would make it better, an apple a10x that also puts the smack down... released in the same week.

 

[Rickyyy369]

I read about this the other day and I thought it was one of the most vague, meaningless statements I've ever heard in my life. I have high hopes for Zen, but in all honesty I'm really only expecting Ivy-bridge level of ST performance with MT coming in a little lower because this is AMD's first attempt at SMT.

 

[myocardia]

I have high hopes for Zen, but in all honesty I'm really only expecting Ivy-bridge level of ST performance with MT coming in a little lower because this is AMD's first attempt at SMT.

And even that is a best-case scenario, that very likely won't come to fruition. If Zen gets anywhere near Ivy Bridge IPC, and it has anywhere near the same efficiency, AMD's engineers will indeed be dancing in the aisles.

 

[zir_blazer]

They may compete in performance, but with a two node disadvantage, performance-per-watt would be terrible, and that matters a lot these days. It would be a GeForce 980 vs Radeon Fury style fight. The miracle that they can pull out a competitive architecture could be possible, but closing the process node gap, not so much.

 

[Boze]

Its easy to meet expectations when you don't expect jack s**t.

That said, I predict Zen performance to be somewhere between i5 and i7 Skylake performance.

 

[Boze]

I'm sick of forum drama. I'd rather talk about the facts when we have them, not fight endlessly over nothing.

I just want to bottle up all the tears when the crying starts when performance isn't nearly as dramatic as the hype train has building it up to be.

There's gonna be a lot of QQing when the ole hype train finally pulls into the station, and I'm going to be here waiting with my ticket puncher.

 

[waffleironhead]

Google is your friend then.
http://www.theinquirer.net/inquirer/news/1016287/amd-to-offer-speed-boosts-on-barcelona

So one hit piece from 2007? Damn. You sure are getting some mileage out of that.

 

[Arachnotronic]

I dont expect quad core 8 thread desktop ZEN to be less than 4GHz.

On a low power optimized 14nm Samsung process? I disagree.

 

[nismotigerwvu]

Time will tell. Would be nice if it traded blows with Intel's best, but even if they got closeish it would dramatically help AMD's competitiveness.

Exactly. Even a Phenom II level of competitiveness would be great for us consumers. The price cuts after that launch were great

 

[Vesku]

Exactly. Even a Phenom II level of competitiveness would be great for us consumers. The price cuts after that launch were great

Yes, even Haswell or Broadwell type of chips should lead to some price competition. That's where Bulldozer design tripped up, they could have met a modest target of Phenom II IPC + 10-20% and 3.8-4GHz top end SKUs if they weren't chasing Intel's Pentium 4 philosophy for whatever silly reasons. Sure it may be a bit of "armchair general" but I can't see how it would not have been a better strategy and it was what I was posting as my hoped for Bulldozer results prior to its launch.

 

[dark zero]

On a low power optimized 14nm Samsung process? I disagree.

And who said Samsung? IBM-Glofo is behind this

 

[nismotigerwvu]

Yes, even Haswell or Broadwell type of chips should lead to some price competition. That's where Bulldozer design tripped up, they could have met a modest target of Phenom II IPC + 10-20% and 3.8-4GHz top end SKUs if they weren't chasing Intel's Pentium 4 philosophy for whatever silly reasons. Sure it may be a bit of "armchair general" but I can't see how it would not have been a better strategy and it was what I was posting as my hoped for Bulldozer results prior to its launch.

They had an even easier option already laid in front of them. LLano had brought something near a 10% increase in IPC over the Athlon II Stars based cores and the shrink to 32nm had a whole host of advantages over the older 45nm process. A Phenom III would have made perfect sense here alongside the initial APUs, just lop off the GPU and APU specific logic and drop in an L3 cache and you've got an attractive package. They likely would have had enough of a transistor budget to sell an 8 core sku with a die size comparable to Deneb (Thuban at the absolute worst). Sandy Bridge would have held a comfortable lead on the high end, but they at least could have competed with the i3 and i5 lines while they developed a proper successor or at least polished Bulldozer enough.

 

[CHADBOGA]

So one hit piece from 2007? Damn. You sure are getting some mileage out of that.

The author of that piece is a renown AMD booster and was himself dancing in the aisles constantly, until Barcelona came out.

 

[Hi-Fi Man]

It's been too long since we've had a really interesting x86 CPU launch but I think some people are getting a bit too excited. Come Zen launch the forums will be engulfed in flames from hell and I'll probably just watch from the sidelines.

 

[Elixer]

The only thing we know as a fact is AMD is betting the farm on Zen.
Therefore, it will be their fastest part they ever made.
Anything else is speculation.

Even if they do have a killer chip, I have high doubts on if GloFlo can deliver the quantity & cost needed to be competitive--even if they can handle sammy's tech (which is a big if).

 

[TurtleCrusher]

If Zen was only 10% slower than Intel's best but cost half as much, I'd buy Zen. Ok fine, no I wouldn't. I'm dying to see benchmarks. This is going to be one hell of a release day. I can't wait to read all about it over a massive cup of coffee and see the nukes start to drop in the forums.

It'll be a good day, regardless of outcome. Finally something interesting to talk about.

 

[JimmiG]

Dont think so,

40% would not be anywhere near enough though. They would have to more than double the IPC (100% more) to remain competitive with Intel in the mid-range, because currently their cores have about half the IPC of Intel.

 

[TechGod123]

I can't wait. I want to build an AMD build next year (Arctic Islands and Zen) So I'm very excited about this.

 

[Shehriazad]

40% would not be anywhere near enough though. They would have to more than double the IPC (100% more) to remain competitive with Intel in the mid-range, because currently their cores have about half the IPC of Intel.

So you're telling me that AMD needs 5 Ghz to be equal to a 2.5 Ghz Intel CPU? I'm not really sure about that.

Now compare Kaveri to that Haswell Pentium (G3258). Kaveri not having an L3 cache it seems to compare in per core performance @ 3.7/4 Ghz (Kaveri) vs 3.2 Ghz (Pentium). I'm looking at titles like Skyrim here btw...those only use 2 cores.
Or even put some old FX 8core chip vs an i7 and clock the i7 @ 2.5 Ghz vs 5 Ghz FX 8core...I'm pretty sure your argument won't really hold up.

In my book that's certainly not 100% IPC difference...or are you telling me that there is a 50% IPC boost from Haswell to Skylake?

Let's not forget that AMD is talking about +40% IPC compared to EXCAVATOR. Which indeed was a bit better than Kaveri in terms of IPC as well.


Take Excavator, shovel on +40% IPC and give back a fast L3 cache and I assure you that the product will be very competitive...even more so if the pricing is right.

I'm not saying AMD can manage that...but I'm also not saying they can't. Regardless if the news up there is fake or not...your post seems a bit...wishy washy. I certainly do not see a 100% IPC difference existing here. That would certainly mean that Intel would be getting twice the scores and double the fps in everything they do...wouldn't it? But I have yet to see that. 50-60% seems more like it.

 

[MisterLilBig]

40% would not be anywhere near enough though. They would have to more than double the IPC (100% more) to remain competitive with Intel in the mid-range, because currently their cores have about half the IPC of Intel.

If AMD was that far behind I would not had bought an AMD chip. Cause you know, then it would had made sense to pay more than double.

 

[stuff_me_good]

Not gonna hop on hype train just yet. Why?

1. Even though IPC would be on par with haswell or even skylake, clocks are mystery. If it clocks only 3GHz tops, it's not going to be on my (down)upgrade path.

2. Power consumption is total mystery as well.

3. GloFo is going to fuck all AMD:s plans anyway so, prepare to get in the waiting line if this is going to even remotely be success.


While I genuinely hope this would be success, I don't want to get as disappointed as with faildoser.

I would gladly give up some IPC compared to my hasswell, if I could get 8 real cores with reasonable price and clocks(4GHz).

 

[TheELF]

Now compare Kaveri to that Haswell Pentium (G3258). Kaveri not having an L3 cache it seems to compare in per core performance @ 3.7/4 Ghz (Kaveri) vs 3.2 Ghz (Pentium). I'm looking at titles like Skyrim here btw...those only use 2 cores.

No skyrim doesn't use only 2 cores,it uses 4 main threads like any other console port and it's locked at 60FPS anyway so even a celeron won't make a difference to an i7.
If you want, search for dolphin benchmark or check out mame performance,those two do use 2 cores, are compiled with 3rd party compilers (not ICC) and will show you just how far behind excavator is.
No,the difference is not 100% but it is huge.

 

[Tuna-Fish]

And who said Samsung? IBM-Glofo is behind this

GloFo doesn't have their own 14nm process, their own efforts completely failed. Instead, they licensed the Samsung 14nm. Zen will be manufactured on this process.

They are developing their own process for 10nm, but that is still years out.

 

[Gikaseixas]

GloFo is going to fuck all AMs plans anyway so, prepare to get in the waiting line if this is going to even remotely be success

GloFo doesn't have their own 14nm process, their own efforts completely failed

This sums it up perfectly. I'm much more careful these days when it comes to forecast cpus, specially AMD made because of how much GF dependent they're. AMD comes up with decent designs that are let down by poor fab processes.