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

[Veradun]

Yes, I can actually. They showed gpu limited gaming benchmarks showing BD as fast as the top of the line Hedt intel cpu.

It must be some bug in the forum software, but I can't see the link, can you send that via pm?

 

[SpaceBeer]

Good enough?
https://www.computerbase.de/2016-12...iagramm-the-division-mit-directx-12-1920-1080

If we have gaming, rendering, file compression... benchamrks showing Ryzen as fast as intel's HEDT CPUs, and some synthetic tests showing it's a little bit slower, will it still be good purchase?

 

[richaron]

*Snip* ...here we have Ryzen barely faster then an i7 7700k overclocked on cinebench...

If by "barely faster then i7 7700k overclocked" you also mean "performing about the same as a similar clocked Ivy/Sandy Bridge 8C(16T)" - then I'm not sure what else you were expecting...

Isn't that about exactly what everyone's been saying for years now? Including AMD? Given the right price I'm sure plenty of people would be happy with the equivalent of 2 x 2600K's glued together (@ half the power). And any higher performance would just widen the potential consumer base, we've already seen some cases where it matches or beats Broadwell for IPC and efficiency*.

I don't know much far as Fritz chess goes, and you may have reason to be disappointed here. But any benchmarks done on Zen now make me wonder about the "neural nets" and associated performance gains (if any). Is their going to be a different benchmarking method like running the software X number of times? Or can one "flash" the CPU to a mode pre-optimized for Y software? Or is all AMD's talk just marketing designed to allure and confuse gullible people like myself?

Edit:*assumptions from New Horizons demos

 

[krumme]

Yes, I can actually. They showed gpu limited gaming benchmarks showing BD as fast as the top of the line Hedt intel cpu.

As you can see you get abit of flak for this as its factual wrong but imo its perfectly fine for such an error as amd used gfx bound bm to put bd in a good ligh.

Where it fails is the comparison to this situation where we have a hedt comparison and for blender and handbrake. To compare the situation to what happen back then is idiotic at best and its just sidetracking the discussion. Its a stupid thing to discuss imo.

Besides - and thats not a remark to you - a ceo just cant go on stage lying. Because most people dont want to go to prison. Saying thats what going on now is naive. And to be frank the amd stock price have been on a rise for months in the middle in diluting the value. And i am pretty sure someone gots more info than we get. And some of the answer is zen. It all fits. Not that this is a 246mm2 bwe killer because its not but the numbers proves this is nowhere near bd situation.

 

[krumme]

We need a betting about the branch predictor in zen now ! And perhaps also some final freq base and turbo. Someone set it up ! dont waste time. The fun stops when full reviews get here.

 

[Nothingness]

We need a betting about the branch predictor in zen now !

What do you want to bet on? Do you really think we'll see serious public review that will do branch prediction analysis on properly compiled code? I'd love to see that for sure

As far as my bet goes, I'd be surprised if AMD was better than Intel at branch prediction. Intel has been very serious about that subject for very long, going as far as organizing public competitions of branch predictors. I'm not saying AMD isn't serious on that matter, but I've yet to see anything but claims from them

The fun stops when full reviews get here.

Yeah, the fun is in the speculation. The problem is that haters and fanboys speculations are spoiling the fun by being so utterly dumb

 

[Arachnotronic]

You don't think last minute microcode and or metal layer changes can take place before mass production? Do you think AMD is mass producing Zen right now with final silicon?

A metal layer change would pretty much require a respin at this point, pushing out the production schedule by several weeks.

The silicon has to be done if you believe AMD's statements. The firmware/microcode that's loaded onto it had better be finished or be finished VERY soon (like, pretty much today) if they want to hard launch this chip in Q1 2017.

 

[Arachnotronic]

If it looks bad for a certain company it's good enough to post and spread it as much as one can, even if the source is a website many blast as not reliable source when two other certain companies benches don't look so good either

The source that showed these Ryzen benchmarks also leaked 7700K info/benchmarks in advance. It's not a random, no-credibility source.

 

[Abwx]

The source that showed these Ryzen benchmarks also leaked 7700K info/benchmarks in advance. It's not a random, no-credibility source.

7700K was already in stores when he "leaked" the results, for Zen only OEMs have samples, you think that an e retailer has Zen samples at this point..?..

 

[coercitiv]

I don't understand why there's still any debate over this: this presumed leak is of no help as long as it does not come with CPU clocks attached. The clock range between 2.8 - 3.4 Ghz is enough to tilt IPC towards SNB and HW respectively, covering the entire spectrum of IPC prognosis in this thread.

Why bother to validate a leak when info isn't enough to tell you anything new?

 

[rainy]

The source that showed these Ryzen benchmarks also leaked 7700K info/benchmarks in advance. It's not a random, no-credibility source.

Looks like it was a fake:
https://www.reddit.com/r/Amd/commen...nebench_r15_against_core_i77700k_and/dbf3673/
http://www.overclock.net/t/1617227/...eview-on-12-13-at-3-pm-cst/1580#post_25723175

 

[raghu78]

Looks like it was a fake:
https://www.reddit.com/r/Amd/commen...nebench_r15_against_core_i77700k_and/dbf3673/
http://www.overclock.net/t/1617227/...eview-on-12-13-at-3-pm-cst/1580#post_25723175

So if those results are fake would the OP please remove them from his postings in this thread.

 

[Glo.]

So if those results are fake would the OP please remove them from his postings in this thread.

Quote from reddit:

Note: This rumor is coming from chinese Baidu forum. The post has been deleted over there and Baidu MODERATOR confirmed that it is fake. The result is actually run by E5 2660.

This: https://ark.intel.com/pl/products/6...E5-2660-20M-Cache-2_20-GHz-8_00-GTs-Intel-QPI

 

[Sweepr]

Looks like it was a fake:
https://www.reddit.com/r/Amd/commen...nebench_r15_against_core_i77700k_and/dbf3673/
http://www.overclock.net/t/1617227/...eview-on-12-13-at-3-pm-cst/1580#post_25723175

First of all, where is the actual Baidu moderator post? All you provided is a random Reddit post from Mr. 'AMD Ryzen Vega 4K 120hz OLED dream'.

''The result is actually run by E5 2660.''

Definitely not a regular Xeon E5-2660, considering Xeon E5-2670 (a faster model) scores 1035-1038 pts in CB R15. Also the processor name looks a lot like previous Ryzen ES leaks:

So if those results are fake would the OP please remove them from his postings in this thread.

Prove it. Until then, they remain there as 'rumor'.

 

[Abwx]

Prove it. Until then, they remain there as 'rumor'.

I would rather say remaining as FUD..

Besides burden of proof is on you, i can as well make such pics and be "right" if no one has the poof that i faked them...

So far the guy that faked those pics doesnt stand by his claims since he disappeared, besides he does not work in the industry and is a simple online shop holder..

He didnt even provide a screen of the system or a CPUZ shot, all this is prove that it s fake..

 

[inf64]

Prove it. Until then, they remain there as 'rumor'.

You know what I have just noticed in the OP? You neatly listed all the rumored/unconfirmed leaks that show poor Zen performance but you failed to put images of the only confirmed (live) runs we have had of Ryzen. You know, the Blender and Hanbrake results showing Ryzen rendering faster than 6900K in Blender and finishing the encode job in Handbrake 5% faster?

That is very telling

 

[Atari2600]

If you click on a username, in the popup, you have access to ignore functionality.

Might be worth doing if you think someone is being blatantly biased and you want to filter out their posts.

 

[Glo.]

You know what I have just noticed in the OP? You neatly listed all the rumored/unconfirmed leaks that show poor Zen performance but you failed to put images of the only confirmed (live) runs we have had of Ryzen. You know, the Blender and Hanbrake results showing Ryzen rendering faster than 6900K in Blender and finishing the encode job in Handbrake 5% faster?

That is very telling

It doesn't matter. Everybody is working for themselves. You are attacking the messenger rather than the message.

 

[krumme]

What do you want to bet on?

As far as my bet goes, I'd be surprised if AMD was better than Intel at branch prediction.

The problem is that haters and fanboys speculations are spoiling the fun by being so utterly dumb

Good we have clever people here that can predict it even without knowing what the question was. Thats ehh a really fancy neural predictor we have there

 

[inf64]

Aaaand he ninja edits the OP after my post

Sweepr, Aug 10, 2016
Last edited: 16 minutes ago

I still do not see Blender result tho, but hey we are making progress here

 

[misuspita]

If you click on a username, in the popup, you have access to ignore functionality.

Might be worth doing if you think someone is being blatantly biased and you want to filter out their posts.

That is true, but still, that kind of behaviour is telling. A user known to be antiAMD in his views does create frequent threads about his favorite punching bag so he can change the title and OP as he desires. It might not be trolling, but it is not really helping the anandtech forums.

This kind of sensitive topics, as a new architecture launch in 5 years, or a new gpu wuold be much better served if a mod opened it and maintained some verticality in the op and thread name. It would be flame retardant on a greater scale than to allow a known poster to do so at almost every oportunity.

And I say this as mostly a lurker, I don't post much, but read frequently. His posting behaviour is evident from skyhigh, even if it does not break the forum rules, it is definitely bending them, and not in a positive way. Allowing such behaviour is not doing a service to the community.

Even if I will be "fined" by a mod, this is something thwt needs to be said.

0.02$

 

[Sweepr]

If you click on a username, in the popup, you have access to ignore functionality.

Might be worth doing if you think someone is being blatantly biased and you want to filter out their posts.

Agreed.

You know, it's very telling that some people prefer attacking the messenger instead of the message here. Especially in light of their previous posting history regarding K8L/K10 and Bulldozer - but that's another topic.

Aaaand he ninja edits the OP after my post

Feel free to ninja edit your post now.

 

[Nothingness]

Good we have clever people here that can predict it even without knowing what the question was. Thats ehh a really fancy neural predictor we have there

And my neural predictor being better than yours and having been trained for longer, I'm sure to win. Want to bet?

 

[krumme]

And my neural predictor being better than yours and having been trained for longer, I'm sure to win. Want to bet?

Are you sure its not just because its truly decoupled? And i mean -truly- decoupled?

Edit. Btw not sure your predictor have been trained for longer than mine. At least i hope for your case its not so but hey perhaps its just difficult to assess the quality of the predictor on a 4 stage inorder non superscalar core.

 

[AtenRa]

Let me put it on perspective: can you link a pre-release benchmark posted by AMD in which Bulldozer was faster than the top of the line Intel processor?

Yes, I can actually. They showed gpu limited gaming benchmarks showing BD as fast as the top of the line Hedt intel cpu.

That was a perf/$ comparison for 1080p Gaming for a total Gaming System in page 26 of the official Bulldozer release PDF and not pre-release benchmark. Also on release Bulldozer was compared vs Core i5 2500K and Core i7 2600K in CPU benchmarks in the official release PDF.