updatebitsandchipsAM4 Socket will be µOPGA and it will have 1331 pins

Page 4 - Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.

itsmydamnation

Platinum Member
Feb 6, 2011
2,731
3,063
136
And they did the same with Phenom. 40-50% faster than Core 2.

Any it was for anything greater the P1, yet you conveniently ignore that. You also ignore that the person who said it was a sales guy for the Server chips......

But you know ignore the parts of history that you don't like.

So why dont you give us a detailed run down on why you are so pessimistic on Zen based of the TECHNICAL details we know so far. Don't worry about keeping it simple go into the Gory details for us, we all might learn something.
 

MiddleOfTheRoad

Golden Member
Aug 6, 2014
1,123
5
0
I see going after the person is still easier than the ball when you lack arguments.

It's not even remotely a personal attack.

You are praising the position of a monopoly -- which is absurd. As I pointed out -- Ask Russia or East Germany how great it is to have a single supplier for a product -- their cars, electronics, everything..... Vastly inferior and decades behind what was produced in the West due to the lack of competition. It's a ridiculous argument so flawed that nobody is even willing to go there -- because anyone with common sense doesn't want to invest the time correcting such a bizarre opinion. Anyone with a minimal amount of business experience is laughing at the suggestion. ROFL

Die shrinks, performance gains, architectural enhancement, innovations..... They all slow or completely stop if Intel becomes the only producer. The history of previous monopolies has a very clear track record/history.

No thanks -- I don't want to start driving a Volga to work, either.
But it's the best car I can buy,
because it's the only car I can buy.
 
Last edited:

ShintaiDK

Lifer
Apr 22, 2012
20,378
145
106
Die shrinks, performance gains, architectural enhancement, innovations..... They all slow or completely stop if Intel becomes the only producer. The history of previous monopolies has a very clear track record/history.

Since you didn't understand the cash flow part. Lets try again.

If Intel doesn't innovate, increase performance metrics, doesn't give any new incentive to upgrade and so on.

Who on earth will ever upgrade their CPU? They wouldn't even last a year before entering chapter 11.

It seems you think they sell tap water with a static demand. Hence dont understand why competition doesn't matter.

Oh, and we had what now, 10 years without competition? It have went quite well so far.
 

ShintaiDK

Lifer
Apr 22, 2012
20,378
145
106
You also ignore that the person who said it was a sales guy for the Server chips......

I didn't know AMD staff works for Intel. That's a new one. Maybe you should refresh your memory. The numbers came from AMDs internal simulations and published by AMD. No less than from Randy Allen, Corporate VP for server and workstation products at the time.

But you know ignore the parts of history that you don't like.

http://forums.anandtech.com/showpost.php?p=26252206&postcount=1

Welcome to reality yet again.
 
Last edited:

The Stilt

Golden Member
Dec 5, 2015
1,709
3,057
106
So AM4 and FM2+ will have a carrizo CPU in common? Will AM4 support FM2+ CPUs?

FM2+ only got Carrizo based Athlon with FCH and GPU disabled. On FM2+ it was impossible to have these enabled, since all FM2+ boards have external FCH (Hudson or Bolton) and the socket infrastructure lacks the third main voltage plane. On FM2+ both the NB and the GPU are fed from the same voltage plane, while Carrizo has VDDNB and VDDGFX separately to improve the power efficiency.

Bristol Ridge on AM4 is essentially exactly the same thing as Carrizo, but this time installed on a desktop package.
 

JDG1980

Golden Member
Jul 18, 2013
1,663
570
136
We have to see how it pans out. But 40% IPC without some very heavy clock reductions is very unlikely. Unless 40% IPC to begin with is the problem.

Don't be absurd. Sandy Bridge had roughly 40% better IPC than the construction cores, and similar clock rates (4.0-4.5 GHz before power consumption starts shooting way up). And that was a 2011 design fabbed on 32nm. It's ridiculous to think AMD can't do at least that well on a clean-sheet design to debut in late 2016 on a 14nm FinFET process.
 

itsmydamnation

Platinum Member
Feb 6, 2011
2,731
3,063
136
I didn't know AMD staff works for Intel. That's a new one. Maybe you should refresh your memory. The numbers came from AMDs internal simulations and published by AMD. No less than from Randy Allen, Corporate VP for server and workstation products at the time.

O RLY.JPG

http://www.zdnet.com/article/amd-claims-quad-core-barcelona-will-outperform-intel-clovertown-by-40/

Still waiting for you to address the other points of my post........ :thumbsup:


Don't be absurd. Sandy Bridge had roughly 40% better IPC than the construction cores, and similar clock rates (4.0-4.5 GHz before power consumption starts shooting way up). And that was a 2011 design fabbed on 32nm. It's ridiculous to think AMD can't do at least that well on a clean-sheet design to debut in late 2016 on a 14nm FinFET process.

So why dont you give us a detailed run down on why you are so pessimistic on Zen based of the TECHNICAL details we know so far. Don't worry about keeping it simple go into the Gory details for us, we all might learn something.

see now two of us are waiting :sneaky:
 

ShintaiDK

Lifer
Apr 22, 2012
20,378
145
106
Don't be absurd. Sandy Bridge had roughly 40% better IPC than the construction cores, and similar clock rates (4.0-4.5 GHz before power consumption starts shooting way up). And that was a 2011 design fabbed on 32nm. It's ridiculous to think AMD can't do at least that well on a clean-sheet design to debut in late 2016 on a 14nm FinFET process.

With what, 1/10th the budget on a 3rd party node?

So you expect 40% IPC increase and 4Ghz+? :)
 

ShintaiDK

Lifer
Apr 22, 2012
20,378
145
106

Doesn't change the history.

http://www.dailytech.com/Kanter+AMD+Will+Wipe+the+Floor+with+Intel+for+HPC/article7338.htm

You can also get it in a video format:
https://www.youtube.com/watch?v=G_n3wvsfq4Y

AMD have a track history for promising more than they can deliver. Yet anyone questioning the latest amazing numbers gets frowned upon. That's a 10 year old story too.
 
Last edited:

itsmydamnation

Platinum Member
Feb 6, 2011
2,731
3,063
136
You wouldn't accept any details as valid ;)

try me, i haven't made one prediction about Zen, what i have tried to do is understand what decisions they are making and why.

Like my current thought is why is the Uop cache 4k in size, that 3 times bigger then sandy bridge. Do they really need 3 times the entries just to get some front end power savings, or is it like the crazy man says and its a trace cache sitting around the retirement unit. That would really help miss latency and allow to still keep a longer pipeline length and thus clocks.

Or i could be completely wrong, So how about you give it a go.


edit: also if your going to use Thevenin posts like you have been when being negativity on Zen remember to explain all of the post like "40501415"
 
Last edited:

AtenRa

Lifer
Feb 2, 2009
13,996
3,356
136
Now that we know AM4 is 140W TDP capable, do we expect more than 95W TDP CPUs ??
 

BigDaveX

Senior member
Jun 12, 2014
440
216
116
If they're making 8-core chips for the desktop, it'll practically be a requirement. The fastest 8-core Xeon that Intel has under the 95W power envelope has a base clock of 2.6GHz, and unless Global Foundries just happen to have an industry-best 14nm process, I can't see them delivering 6+ core chips with decent clocks at sub-100W.
 

itsmydamnation

Platinum Member
Feb 6, 2011
2,731
3,063
136
If they're making 8-core chips for the desktop, it'll practically be a requirement. The fastest 8-core Xeon that Intel has under the 95W power envelope has a base clock of 2.6GHz, and unless Global Foundries just happen to have an industry-best 14nm process, I can't see them delivering 6+ core chips with decent clocks at sub-100W.

if 256bit FPU didn't exist those 8 core HEDT/Xeon cpu's could have a lower tdp.
 

NTMBK

Lifer
Nov 14, 2011
10,208
4,939
136
If they're making 8-core chips for the desktop, it'll practically be a requirement. The fastest 8-core Xeon that Intel has under the 95W power envelope has a base clock of 2.6GHz, and unless Global Foundries just happen to have an industry-best 14nm process, I can't see them delivering 6+ core chips with decent clocks at sub-100W.

Those 8-core Xeons also have big, power hungry 256-bit vector units. Rumour has it that Zen won't.
 

majord

Senior member
Jul 26, 2015
433
523
136
Doesn't change the history.

http://www.dailytech.com/Kanter+AMD+Will+Wipe+the+Floor+with+Intel+for+HPC/article7338.htm

You can also get it in a video format:
https://www.youtube.com/watch?v=G_n3wvsfq4Y

AMD have a track history for promising more than they can deliver. Yet anyone questioning the latest amazing numbers gets frowned upon. That's a 10 year old story too.

You're quite hung up on these 10 year old claims. (Can't believe it's been that long)

Firstly though, that was never a straight up IPC claim. It was vague, and as you're well aware focussed on FP and Int_Rate benchmarks which were not a good indication of IPC.

Secondly, Since then, actual projected IPC claims have been against their own Architectures, and have been pretty modest and on the mark. I'm talking Phenom II, Piledriver, Steamroller, and Excavator Here. (Notice I left out Bulldozer, since there was never an IPC claim on paper - quite obvious why in hindsight!)

So really, I have no reason do doubt the claims of Zen are in the ballpark, based on recent history / track record..

still waiting for you details on why not.

I have to say I agree Frequency will take a hit, Just not dramatically as Shintai is suggesting. There's several pieces of the puzzle here though IMHO:

1. It's difficult to know how FF14nm will play out with higher clockspeeds, but I really don't think there will be much to gain vs 28nm at the higher end. Look at Skylake for example.. The trend is definitely towards lower Freq and has been so since 22nm.

2 Will AMD be utilizing high density libraries in all variants of Zen? Doing so had profound (positive) implications for lower clocked Excavator, but has the opposite effect at the upper end of the freq curve.

3. Pipeline depth - I'm certain it will be shorter: If you think about the above two points, it's obvious the current architecture is basically at war with both 'itself' (think HD libraries) , and the process node when it comes to achieving high performance - both of which have collectively brought clockspeed down, negating the significant IPC / CMT gains that have been made.

I honestly think the CON core's concept would have actually been half decent back in the days of in-house SOI. Instead, we've basically stalled - arguably regressed when it comes to max frequency scaling.. Intel included.
 
Last edited:

deasd

Senior member
Dec 31, 2013
508
711
136
If they're making 8-core chips for the desktop, it'll practically be a requirement. The fastest 8-core Xeon that Intel has under the 95W power envelope has a base clock of 2.6GHz, and unless Global Foundries just happen to have an industry-best 14nm process, I can't see them delivering 6+ core chips with decent clocks at sub-100W.

Any confirmation that claim Zen will be fabbed at Glo Fo?
 
Aug 11, 2008
10,451
642
126
Don't be absurd. Sandy Bridge had roughly 40% better IPC than the construction cores, and similar clock rates (4.0-4.5 GHz before power consumption starts shooting way up). And that was a 2011 design fabbed on 32nm. It's ridiculous to think AMD can't do at least that well on a clean-sheet design to debut in late 2016 on a 14nm FinFET process.

Yes, and it was ridiculous to think they would release Bulldozer with worse IPC in a lot of workloads that Phenom, but they did it. And that was a "clean sheet" design. What I am saying is that it is really not that easy, and bad designs make it to market.
 

Unoid

Senior member
Dec 20, 2012
461
0
76
It's not even remotely a personal attack.

You are praising the position of a monopoly -- which is absurd. As I pointed out -- Ask Russia or East Germany how great it is to have a single supplier for a product -- their cars, electronics, everything..... Vastly inferior and decades behind what was produced in the West due to the lack of competition. It's a ridiculous argument so flawed that nobody is even willing to go there -- because anyone with common sense doesn't want to invest the time correcting such a bizarre opinion. Anyone with a minimal amount of business experience is laughing at the suggestion. ROFL

Die shrinks, performance gains, architectural enhancement, innovations..... They all slow or completely stop if Intel becomes the only producer. The history of previous monopolies has a very clear track record/history.

No thanks -- I don't want to start driving a Volga to work, either.
But it's the best car I can buy,
because it's the only car I can buy.


Well said and correct.
 

Dresdenboy

Golden Member
Jul 28, 2003
1,730
554
136
citavia.blog.de
Doesn't change the history.

http://www.dailytech.com/Kanter+AMD+Will+Wipe+the+Floor+with+Intel+for+HPC/article7338.htm

You can also get it in a video format:
https://www.youtube.com/watch?v=G_n3wvsfq4Y

AMD have a track history for promising more than they can deliver. Yet anyone questioning the latest amazing numbers gets frowned upon. That's a 10 year old story too.
You use a technically based opinion and a well known "PR" (explaining sth to public) guy's exaggerated claims as a base for predicting the future? You might be right with this method as long as today's situation is exactly the same as nearly 10 years before.

Any change to this would be like throwing dice again. But humans tend to base their assumptions on simplified models of reality. OK, since AMD didn't deliver CPUs beating Intel CPUs in the respective markets (excluding cat cores vs. pre-Silvermont cores here), this might simply happen again. Intel might - other than already known so far - bring a significant performance and power efficiency improvement with Kaby Lake (not only more or improved IP blocks), while AMD's 40% claim might just be an "up to" for Dhrystone only, while the 14LPP improvements will fall short (ex-AMD fabs are casting their shadows), only delivering an AMD special with 25% power reduction for Zen because AMD didn't learn from the bad B's Barcelona and Bulldozer and continued to apply highschool homework quality logic synthesis workflows. Maybe the probability of this is a non-zero one, so we should take it as 100% (rounding up).

Satisfied? :)
 
Last edited:
Aug 11, 2008
10,451
642
126
Well said and correct.

Quite an overstatement actually. I agree we need competition, but all the hysteria and hand-wringing about the lack of competition is crazy as well. Just look at the server market, Intel *is* essentially a monopoly there, yet advancements are made each generation, probably more than in the consumer market. Why? Because Intel needs to bring out an improved product to drive new sales and especially upgrades.
 

cbn

Lifer
Mar 27, 2009
12,968
221
106
Now that we know AM4 is 140W TDP capable, do we expect more than 95W TDP CPUs ??

It makes sense that would be true.

And maybe is still developing the cooler for that?

P.S. Wraith is 125W+ cooler being used on 95W CPUs, I wonder what kind of HSF specs AMD is developing to keep a 140W CPU quiet? Maybe a slightly taller Wraith? Or possibly a Wraith that is also wider using a 120mm blow down fan?
 
Last edited:

NTMBK

Lifer
Nov 14, 2011
10,208
4,939
136
P.S. Wraith is 125W+ cooler being used on 95W CPUs, I wonder what kind of HSF specs AMD is developing to keep a 140W CPU quiet? Maybe a slightly taller Wraith? Or possibly a Wraith that is also wider using a 120mm blow down fan?

AIO liquid cooler, judging by AMD's past experience.