AMD Zen ES leaked benchmarks: performance similar to a Core i5 4670K?!?!

[Dabombinable]

Just now, FPS-Russia said:

Why is there so much moaning?I have not seen anything new from AMD's CPU gates in such a long time, i just want to see a new CPU from them.

 

4670K performance is still quite a huge leap from FX 8000 series.

Its still too far behind though.

[themctipers]

make this 8 cores and under $200 cad

and i got a fucking amd cpu in my house. never thought that would happen

family is intel only, dads a fanboy. although if this shit is better than a 3770k ill somehow convince him to get a "shitty amd cpu" and he'll get back his shiny way fucking better i7 amirite? /s

[FPS-Russia]

Just now, Dabombinable said:

Its still too far behind though.

I have an i7 970.

[Dabombinable]

Just now, zMeul said:

from Haswell to Haswell-E is just a node shrink if I recall right

 

you have Haswel-E, Broadwell, Skylake, Broadwell-E and Kabi Lake later this year - so yeah ... AMD is kinda' fucked

Haswell to Broadwell was the shrink-Haswell E was 22nm still.

[zMeul]

Just now, Dabombinable said:

Haswell to Broadwell was the shrink-Haswell E was 22nm still.

Broadwell has some different arch if I recall; new memory controller, they also removed the VRM from the CPU die back to the mobo

[Coaxialgamer]

3 minutes ago, zMeul said:

 

 

people that still believe a architecture design that already has some years over it will magically compete with the Intel current and upcoming CPUs drank the Kool-Aid, just like they did with Polaris 10

What bulldozer ended up being ( and it's shortcomings)  is irrelevant to the potential performance of zen.  They  are starting  a completely new core design unlike piledriver and steamroller,  which were just desperate attempts to fix a broken architecture. 

If anything,  i would be more concerned about amd's R&D,  not their engineering team's capabilities or their previous mistakea.  After all AMD have been stuck with the bulldozer core for the past 6 years.  I would think of it like Intel's p4 back in the day,  when they were stuck with netburst.  After all,  k5 was decent,  k6 ( and subsequent k6-2 and k6-3)  was good,  so was k7. K8 was great,  and k10 was ok ( but was heavily based on k8) 

[Dabombinable]

Just now, FPS-Russia said:

I have an i7 970.

And that's what the FX 8*** truly competes with due to its IPC-multi threaded it scales far better of course being CMT not SMT

[FPS-Russia]

Just now, Dabombinable said:

And that's what the FX 8*** truly competes with due to its IPC-multi threaded it scales far better of course being CMT not SMT

Yeah that's what i mean so if zen is 4670k performance which none of us really know for fact yet, that's a good upgrade for me and many others. Pricing will be the kicker though.

[Dabombinable]

Just now, zMeul said:

Broadwell has some different arch if I recall; new memory controller, they also removed the VRM from the CPU die back to the mobo

Broadwell still has the FIVR-Skylake is the architecture where Intel removed it (I can see why though-Haswell and Broadwell throw out shitloads of heat due to it).

[zMeul]

On 8/10/2016 at 7:42 PM, Coaxialgamer said:

If anything,  i would be more concerned about amd's R&D,  not their engineering team's capabilities or their previous mistakea.  After all AMD have been stuck with the bulldozer core for the past 6 years.  I would think of it like Intel's p4 back in the day,  when they were stuck with netburst.  After all,  k5 was decent,  k6 ( and subsequent k6-2 and k6-3)  was good,  so was k7. K8 was great,  and k10 was ok ( but was heavily based on k8) 

back then Intel got caught of guard because they dragged their feet - then K8 hit them right in the ballz

but Intel isn't doing that anymore, they die shrinking, getting new arch to market - they aren't slouching around

 

On 8/10/2016 at 7:43 PM, Dabombinable said:

Broadwell still has the FIVR-Skylake is the architecture where Intel removed it

oh, sorry then .. I remembered it wrong 

Edited August 11, 2016 by zMeul

[Coaxialgamer]

Just now, zMeul said:

back then Intel got caught of guard because they dragged their feet - then K6 hit them right in the ballz

but Intel isn't doing that anymore, they die shrinking, getting new arch to market - they aren't slouching around

True.  But they have been stuck on 14nm for a while,  and will continue to be for at least a year... 

[Dabombinable]

1 minute ago, zMeul said:

oh, sorry then .. I remembered it wrong 

TBH, Broadwell is kind of easy to forget due to when it was released. Right before Skylake.

[zMeul]

1 minute ago, Coaxialgamer said:

True.  But they have been stuck on 14nm for a while,  and will continue to be for at least a year... 

node shrinking brings challenges that sometimes aren't feasible to just throw cash at it

[Coaxialgamer]

Just now, zMeul said:

node shrinking brings challenges that sometimes aren't feasible to just throw cash at it

It just takes time for the process to mature i guess.  

[Energycore]

I think I learnt my lesson not to trust leaks back when Polaris launched. Staying off this one. 

[patrickjp93]

1 hour ago, Dabombinable said:

There is a reason as to why Intel's 8 cores have such a high TDP or low base clocks. That many cores under the same IHS throws out shitloads of heat-regardless of the IHS being soldered on or using some other TIM.

Intel also gives the TDPs for standard workloads under the hottest instructions like vectorized work (SSE/AVX). If AMD is trying to get away with gaming TDP, it won't fly in the server world. The moment they're caught being less than truthful, everyone will pay to buy out of any AMD contracts and switch right back to Intel, because if AMD is willing to lie about TDP, what else is wrong under the hood?

 

1 hour ago, Coaxialgamer said:

True.  But the 5960x was a 22nm part.  Samsung's 14nm LPP should have better power characteristics... Anyway,  ill stay away from any performance for the time being,  especially from wccf... 

The original source is dresdenboy on overclock.net and overclockers.uk as well as computerbase.de. WCCF just reported on that.

 

1 hour ago, Coaxialgamer said:

True,  but every chip from the 6800k to the 6950x are all rated for 140w,  while they probably have quite different heat outputs.  Exact  TDP measures aren't always accurate,  as they are often extended to a whole cpu family ( I'm looking at you, 6600k and 6700k) 

This is much more a problem in consumer processors than server ones where you're really not allowed to lie.

[AresKrieger]

19 minutes ago, Energycore said:

I think I learnt my lesson not to trust leaks back when Polaris launched. Staying off this one. 

Yeah, didn't they say 980+ performance in the rumors on those lol, 970 performance is what we got, this is why "leaks" always need proper seasoning

[i_build_nanosuits]

4 hours ago, leelaa14 said:

Ashes of the singularity is AMD's marketing tool to con people into thinking there products are better than they are.

[Dabombinable]

48 minutes ago, patrickjp93 said:

Intel also gives the TDPs for standard workloads under the hottest instructions like vectorized work (SSE/AVX). If AMD is trying to get away with gaming TDP, it won't fly in the server world. The moment they're caught being less than truthful, everyone will pay to buy out of any AMD contracts and switch right back to Intel, because if AMD is willing to lie about TDP, what else is wrong under the hood?

 

The original source is dresdenboy on overclock.net and overclockers.uk as well as computerbase.de. WCCF just reported on that.

 

This is much more a problem in consumer processors than server ones where you're really not allowed to lie.

The funny thing is-if they get caught selling products in Australia based on a lie-they can get in deep shit.

[Okjoek]

Well even if it's not an intel-beater the 4670K is still a fantastic upgrade to what I'm using currently and I'd gladly take it as an upgrade if AMD can atleast make the price competitive.

[patrickjp93]

On 8/10/2016 at 9:16 PM, Dabombinable said:

The funny thing is-if they get caught selling products in Australia based on a lie-they can get in deep shit.

That's anywhere really.

 

On 8/10/2016 at 9:37 PM, Okjoek said:

Well even if it's not an intel-beater the 4670K is still a fantastic upgrade to what I'm using currently and I'd gladly take it as an upgrade if AMD can atleast make the price competitive.

When will you people learn?! AMD historically matches Intel's prices when it has competitive products. Athlon Extreme -$1000, Phenom X6 -$1000, FX 9590 -$1000! Even if for some reason it doesn't this one time, Intel will just respond by dropping prices, and you'll be able to get an even better upgrade for the same price!

[Dabombinable]

1 minute ago, patrickjp93 said:

When will you people learn?! AMD historically matches Intel's prices when it has competitive products. Athlon Extreme -$1000, Phenom X6 -$1000, FX 9590 -$1000! Even if for some reason it doesn't this one time, Intel will just respond by dropping prices, and you'll be able to get an even better upgrade for the same price!

The FX 9590 competes with nothing BTW........

[Energycore]

Just now, patrickjp93 said:

When will you people learn?! AMD historically matches Intel's prices when it has competitive products. Athlon Extreme -$1000, Phenom X6 -$1000, FX 9590 -$1000! Even if for some reason it doesn't this one time, Intel will just respond by dropping prices, and you'll be able to get an even better upgrade for the same price!

Indeed prices will dictate how the market unfolds. That said, it's nice to have an AMD that now has the option of offering a $200, 300, 500 CPU that perfoms up to par.

[handymanshandle]

Close enough to an i5 6600K, like many others said, but will it be cheaper or on par in price?

[patrickjp93]

3 minutes ago, Dabombinable said:

The FX 9590 competes with nothing BTW........

When it launched, there were some compelling cases for it, but they were integer-only workloads.