AMD Ryzen (Summit Ridge) Benchmarks Thread (use new thread)

[bjt2]

What doom and gloom? All I hear about Zen is solid clocks for an 8c part (3.8Ghz+ for MT loads) plus accurate and flexible power management. Except for power consumption at high clocks (4Ghz+) which many prefer to approach in a more cautious manner, Ryzen looks like a great CPU. Not ok, not good enough, but great. In marketing speech that would translate to freaking spectacular.

Even if the top 3.6/4Ghz part ends up being a ~120W TDP part like CPC Hardware suggests, the only fault Zen appears to have is obeying the laws of physics. At this point all I'm interested to find out about Zen is pricing and performance consistency. Frequency and power are no longer a curiosity, except maybe in relation to max oc numbers.

I don't know if AMD was hiding very well at new horizon event or the ES shown was the best that AMD can produce. Worst case is 3.4GHz@95W. We don't have true proof of higher frequencies. Only speculations on solid data as I have done here. But something could go wrong. Yield can be low. Process variation can be high. I am not sure that Zen can reach those theorethical frequencies in the first batches.

 

[Greyguy1948]

We all dream about the performance of the new Ryzen.
This is a real cold shower....
Today I went to SPEC.ORG and compared some 2006fp_base results.
Best Opteron with 16 cores was on 44
Best Xeon with 16 cores and 1 thread/core was on 127
Best Xeon with 8C/16T was on 112
40% improvement----> 62
55% improvement----> 68

 

[Abwx]

Even if the top 3.6/4Ghz part ends up being a ~120W TDP part like CPC Hardware suggests

Can you post a link where he suggest that it s 120W..?.

Of course you wont find one, so you are just putting your own numbers in his mouth...

What he suggest is just that it should be more than 95W, but he based his assumptions on the measurements of the sample he got and wich did consume about 82W at the CPU level, wich make me think that it was quite a bold, if not erroneous, statement of his, because from 3.2GHz to 3.6GHz there should be only 25% more power and we know that the ESs are overvolted.

 

[coercitiv]

Can you post a link where he suggest that it s 120W..?.

Of course you wont find one, so you are just putting your own numbers in his mouth...

You really are incapable of decent conversation. Here's your link, translate your way out of it like you enjoy doing.

Bon ceci dit, arriver à 3.6/4.0G avec 8C/16T dans 120W, ce serait déjà une belle prouesse. Il faut remettre les choses dans leur contexte.

 

[Abwx]

You really are incapable of decent conversation. Here's your link, translate your way out of it like you enjoy doing.

Talk for yourself since i fully understand what he said, in substance he state that it would be a prowess if they managed to be within 120W at 3.6GHz, and that things should be put in context.

Now looking at his own measurements and conclusions it s clear that he doesnt have a good knowledge of semiconductors laws, not that he s incompetent, he designed both the Memtest 86+ version and CPUZ and is a very capable software guy, but on the matter of TDP i can easily challenge his saying using his very datas, it s not because he has large software knowledge that he has better understanding of electronics than either BJT2 or myself...

 

[coercitiv]

Now looking at his own measurements and conclusions it s clear that he doesnt have a good knowledge of semiconductors laws

Unbelievable. Enjoy your Zen launch.

 

[Mechanical Man]

You really are incapable of decent conversation. Here's your link, translate your way out of it like you enjoy doing.

My google translate says, if 3,6/4,0 is under 120 its a good achievement. For me it implies, a) he donta have that part, b) he extrapolated from his ES

We cant know what was on or foo what comes to ES. We know XFR was off, that can also have big part in final power usage.

 

[BenjiBenjamin123]

Thanks for useful information and links, guys! I really appreciate that!

 

[CatMerc]

But CPC is sure that there is no 65W 8C chip...

Either it's a typo, or CPCHardware didn't catch wind of the sample. It might be something new, or it might be AMD managing to hide what they're doing from prying eyes.

The models exist, that's for near certainty right now. The only question is if it's an 8 core, or if the TDP is 65W. Could be a 6 core or a 95W part.

Edit:

 

[LTC8K6]

He has not seen it. But if that reseller have actual informations, then it should exist.

It wouldn't be unusual for a chip to

Either it's a typo, or CPCHardware didn't catch wind of the sample. It might be something new, or it might be AMD managing to hide what they're doing from prying eyes.

The models exist, that's for near certainty right now. The only question is if it's an 8 core, or if the TDP is 65W. Could be a 6 core or a 95W part.

Edit:

"AMD is not currently validating any 8C part @ 65W (nor any 8C part without SMT)"

That is what he said. He is changing his tune, imo.

 

[dacostafilipe]

Either it's a typo, or CPCHardware didn't catch wind of the sample. It might be something new, or it might be AMD managing to hide what they're doing from prying eyes.

The models exist, that's for near certainty right now. The only question is if it's an 8 core, or if the TDP is 65W. Could be a 6 core or a 95W part.

I think CPCH is just confusion TDP with peak consumption.

 

[CatMerc]

It wouldn't be unusual for a chip to


"AMD is not currently validating any 8C part @ 65W (nor any 8C part without SMT)"

That is what he said. He is changing his tune, imo.

He's clarifying his exact meaning. He's not aware of any 8C @65W parts going through validation, but it might not be necessary for them to do so.

They can release slightly modded parts w/o full tech valid

 

[Doom2pro]

He's clarifying his exact meaning. He's not aware of any 8C @65W parts going through validation, but it might not be necessary for them to do so.

Especially if they are products of binning, it's already validated just better silicon.

 

[LTC8K6]

He's clarifying his exact meaning. He's not aware of any 8C @65W parts going through validation, but it might not be necessary for them to do so.

I can't wait for his next clarification...

 

[CentroX]

CPC is backpedaling?

 

[Asterox]

 

[bjt2]

Is there any link?

 

[Doom2pro]

Is there any link?

To?

 

[LTC8K6]

CPC is backpedaling?

Instead of just saying he was wrong, or he doesn't actually know, he keeps "clarifying".

Pedaling in circles, I'd say.

 

[inf64]

I doubt there is any link but **supposedly** 5Ghz 16T Ryzen scores a tad over 2000pts in R15 x64. If top Ryzen somehow miraculously can run all 16T at 4Ghz (max boost) in R15 MT benchmark, it should **supposedly** score around 1600pts. All not verified of course

 

[guskline]

inf64, those Ryzen numbers for R15 MT look impressive. I just ran R15MT on my 5960X clocked at 4.4 and scored 1759. I dropped the clock speed to 4 Ghz on all 8c/16t and reran R15 MT with a score of 1618. If Ryzen comes in at @1600 THAT is impressive.

 

[Magic Hate Ball]

You'll have to change this, the COO just quit.

 

[lolfail9001]

I doubt there is any link but **supposedly** 5Ghz 16T Ryzen scores a tad over 2000pts in R15 x64.

Looks inline with Haswell-E http://hwbot.org/submission/2674251_mydog_cinebench___r15_core_i7_5960x_2045_cb
So, nothing that we did not know already about it's performance. Btw, makes Asterox macro even more stupid.

 

[DrMrLordX]

We all dream about the performance of the new Ryzen.
This is a real cold shower....
Today I went to SPEC.ORG and compared some 2006fp_base results.
Best Opteron with 16 cores was on 44
40% improvement----> 62
55% improvement----> 68

Seriously? That isn't even XV, and you haven't listed clockspeeds either. Yeah I ain't worried.

You'll have to change this, the COO just quit.

Death before dishonor!

 

[CentroX]

is there any reason for AMD to not reveal their launch lineup right now when retail listings are out?