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

[DXMember]

9 minutes ago, patrickjp93 said:

How is it ridiculously low? That's basically 5960X range.

i guess... but considering 5960X easily overclocks past 4 GHz...

I sometimes think Intel like nVidia intentionally clocks lower to give the illusion of overclockable parts

[patrickjp93]

2 minutes ago, Prysin said:

did the leaks you posted about ES run at 2.5GHz?

Yes, that's an older sample though, so it's not too surprising.

 

2 minutes ago, DXMember said:

i guess... but considering 5960X easily overclocks past 4 GHz...

I sometimes think Intel like nVidia intentionally clocks lower to give the illusion of overclockable parts

and the 5960X doesn't sit on a 95W TDP (we don't know yet if this is just gaming workloads or if this miraculously does include vectorized workloads).

 

No that has more to do with the expense of binning. I myself am sitting on a 2600K that can't overclock more than 100MHz without sitting on 1.55 volts.

[DildorTheDecent]

19 minutes ago, patrickjp93 said:

FX 9590 were all $1000 at launch),

After quick Google it was. 

 

Absolutely laughable

[Murasaki]

>LOW 1080P

[Rapsky]

if its still around a i5 4670k performance I wont be getting it. Next plan was getting a i5 or i7 6700k and a high to mid end gpu but waiting for new stuff to come this year.

[DXMember]

[Prysin]

1 hour ago, patrickjp93 said:

Yes, that's an older sample though, so it's not too surprising.

 

and the 5960X doesn't sit on a 95W TDP (we don't know yet if this is just gaming workloads or if this miraculously does include vectorized workloads).

 

No that has more to do with the expense of binning. I myself am sitting on a 2600K that can't overclock more than 100MHz without sitting on 1.55 volts.

well, you shoulda bought a FX 8350 then
nearly all of em hit 4.8GHz with ease. And imagine how much fun you could have with them, trying to make the float and the ALU work at the same time through a custom compiler

 

non related topic

is skype still borked for you?

[Djole123]

8 hours ago, TidaLWaveZ said:

I hope the initial Zen offering is at least the performance of a 5820k.  I'd really like to see a powerful AMD cpu that overclocks like the dickens!

I doubt it will get to 5820K performance. I'd say it'll compete with the 5675C or maaaaaybe with the 6500.

[patrickjp93]

1 hour ago, Prysin said:

well, you shoulda bought a FX 8350 then
nearly all of em hit 4.8GHz with ease. And imagine how much fun you could have with them, trying to make the float and the ALU work at the same time through a custom compiler

 

non related topic

is skype still borked for you?

I just haven't been on it. It's fine now.

[MeDownYou]

 

 

 

aaaaand with that i'm out (fury X used in tests with intel based cores) also beta benches

[Humbug]

On 8/11/2016 at 9:20 AM, SkilledRebuilds said:

Zen rev 2.0 should be quite the amazeballs, higher clocks and lower tdp and all that...

 

When Zen rev 1.0 releases.. great.. but people just totally crave it to surpass any and all... it may not and people will get their pitchforks.

 

It will be great either way..

You will have options either way.

zen doesn't really need to be the 2nd coming.

but it needs to be good enough so that AMD can start selling at decent margins and decent quantities and make money from CPUs again.

 

On 8/11/2016 at 1:02 PM, Rapsky said:

if its still around a i5 4670k performance I wont be getting it.

this was an engineering sample.

 

normally clockspeeds are ramped up as we approach launch. Hard to say where it will end up..

[patrickjp93]

2 minutes ago, Humbug said:

this was an engineering sample.

 

normally clockspeeds are ramped up as we approach launch. Hard to say where it will end up..

But under a promised 95W TDP for 8 cores, this is about as far as we can reasonably expect unless AMD intends to cheat and just report what TDPs are like under SISD workloads.

[Prysin]

3 hours ago, Humbug said:

this was an engineering sample.

 

normally clockspeeds are ramped up as we approach launch. Hard to say where it will end up..

probably around the 3.4-3.7GHz area. If intel could do it on 22nm, there is no way that AMD couldnt do it too at 14nm.

[ApolloFury]

 

On 11/08/2016 at 7:14 AM, leelaa14 said:

Nope, I'm not. But any time AMD advertise there products lately, they use ashes, solely. The performance that you see them advertising on this game doesn't translate well to other games. Just saying what I see.

 

No need to be a bell end about it.

Project Cars - You see Nvidia logos all over the place, on the cars and on the side of the tracks, in the early days Nvidia really loves to advertise this game a lot like AOTS. You get cards like GTX 770 or GTX 960 matching/outpeforming R9 290X's. Tom Hardware always uses this game for benchmarking which is just misleading, they skew the performance graphs significantly by benchmarking this game alone.

[DXMember]

Jim Keller, make AMD great again!

[Prysin]

4 hours ago, DXMember said:

Jim Keller, make AMD great again!

he already left.

 

appropriate slogan is "Jim Keller: I'll be back"

[zMeul]

Guru3D did something: https://www.guru3d.com/news-story/amd-zen-engineering-sample-aos-further-analysis.html

they took a  i7 5960X Haswell-E (because it matches the core count) and downclocked it to match the ZEN engi sample, put a RX480 and ran Ashes at the same settings - it does not look good for ZEN

 

example:

[Dabombinable]

10 hours ago, zMeul said:

TheGuru3D did something: https://www.guru3d.com/news-story/amd-zen-engineering-sample-aos-further-analysis.html

they took a  i7 5960X Haswell-E (because it matches the core count) and downclocked it to match the ZEN engi sample, put a RX480 and ran Ashes at the same settings:

it does not look good for ZEN

 

example:

Ok. AMD is screwed if that's the way Zen really is.

[zMeul]

1 minute ago, Dabombinable said:

Ok. AMD is screwed if that's the way Zen really is.

they could price it accordingly and maybe could save their bacon for 1-2y before (IF) they come up with some actual nice arch

[Dabombinable]

Just now, zMeul said:

they could price it accordingly and maybe could save their bacon for 1-2y before (IF) they come up with some actual nice arch

The problem is-what are the actual manufacturing costs? And would AMD be able to keep them in supply (and therefore keep the prices from increasing significantly) if they can be priced accordingly? Those are going to be the 2 factors that make or break AMD if Zen does shape up to be this much of a let down. Its supposed to have near Haswell performance FFS.

[Sintezza]

Interesting, but i highly doubt those Guru3D test methods and scoring numbers.

 

[Doobeedoo]

It would seriously be quite odd if Zen doesn't have massive IPC increase over previous 8-core. I mean, it's on quite smaller node, 14nm, it's completely different architecture from ground up and also it's SMT and not CMT anymore which results in quite higher single-core performance. 

Can't wait to see final sample with actual benchmarks already. Curiosity is killing me.

[Prysin]

10 hours ago, zMeul said:

Guru3D did something: https://www.guru3d.com/news-story/amd-zen-engineering-sample-aos-further-analysis.html

they took a  i7 5960X Haswell-E (because it matches the core count) and downclocked it to match the ZEN engi sample, put a RX480 and ran Ashes at the same settings - it does not look good for ZEN

 

example:

well there is factors at play here. Such as turbo speeds. How aggressive AMDs power tuning is, remenber the ZEN 8c 16t is supposed to be 95w... the 5960x is a 140w TDP part... meaning that if AMD went for power efficiency rather then absolute performance, then their CPU would throttle. That being said, if it "out of the box" is power starved, then overclocking it should boost performance by ridiculous amounts.

 

As for the gap in "low" batches. That could be a cache thing, or a compiler thing. Hard to say atm, although the information is interesting.

[leelaa14]

On 10/08/2016 at 10:17 PM, leelaa14 said:

Well then, seems that I was wrong, still though, seems a bit weird how they are all over this game to show there performance. We will see when it releases then.

 

22 minutes ago, Nicholatian said:

https://en.wikipedia.org/wiki/Argumentum_ad_populum

 

A legion of idiots with an opinion doesn’t change anything about the facts of the matter. Come on now, you should be smarter than that!

 

1 minute ago, Nicholatian said:

Sadly, no. And I don’t think you do either.

I think I do.

[TidaLWaveZ]

19 hours ago, Djole123 said:

I doubt it will get to 5820K performance. I'd say it'll compete with the 5675C or maaaaaybe with the 6500.

 

You're probably right, we can always dream though.