AMD Ryzen 8C/16T Cinebench R15 & Fritz Chess Scores

[-Sweeper_]

Summit Ridge 8C/16T Cinebench R15 MT & Fritz Chess Scores Leaked

Comparison #1:

Core i7-7700K (Stock) 971* cb
Core i7-7700K (OC) 1108* cb
Core i7-6900K (Stock) 1565 cb
Core i7-6950X (Stock) 1863 cb
AMD RYZEN (8/16) 1188 cb

*http://www.bit-tech.net/hardware/cpus/2016/12/05/core-i7-7700k-performance-overclock-preview/3

Comparison #2:

Comparison #1:

Core i7-7700K (Stock)
35.52
Core i7-7700K (5GHz)
41.44
Core i7-6900K (Stock)
47.80
Core i7-6950X (Stock)
51.50
AMD RYZEN (8/16)
36.86

Core i7-6900K 30% faster @ FritzChess, 31% faster @ Cinebench R15.

Via: http://wccftech.com/amd-ryzen-zen-cpu-benchmarks-leak/

 

[Shintai]

It seems all scores leaked keep showing the same pattern. Ryzen seems to be SB performance.

Actually performance may be below SB.
https://us.rebusfarm.net/en/tempbench?view=benchmark

 

[aztekk]

This is exactly why I said after that AMD demo that we need to wait for retail availability / independent non cherry picked benchmarks. A lot of people here naively bought into AMD's marketing about it being comparable to a $1000 CPU. Things are never as rosy as they look, especially with a for-profit corporation who is always looking to showcase their product in the best light possible. So I am not surprised it could be significantly slower than the i7-6900k.

However as I've said before, I will not be buying this processor if it's slower than Ivy Bridge in single-threaded workloads. That's just unacceptable for 2016 in my opinion. And yes I want AMD to do well, who doesn't, but we need to hold these companies to the highest standard and only give credit when they release a superior product overall. If you feel like doing charity, donate to the Red Cross or something.

I love AMD's past products when they made something truly competitive even in the highest end of the market, like with Athlon XP and A64 (I bought multiple of both back then). With performance shown here, Ryzen will clearly not be one of those legendary chips, and while it will have its place in the marketplace (barring AMD doing something stupid with pricing), enthusiasts looking for a truly premium processor will continue to have Intel as the only current provider for their needs (excluding certain highly multi-threaded use cases). Gamers will also be better off with a solid single-threaded performing quad core CPU instead.

That's what I take from this leak of benchmarks assuming this is the true performance we should expect. Again this could very well be fake, considering the source for this is just a post on some Chinese forum... Therefore no one should be making final conclusions until these chips hit the shelves next year.

 

[ecmaster76]

1. WCCFTECH article
2. No source link in article to "Chinese forums"
3. Camera picture of monitor cropped to only show score
4. No info on sys config/clock speed

Yep, lets start making our purchase plans now

 

[auntjemima]

I enjoyed the cropped photo with just a score. That was neat.

 

[SighTurtle]

Why would they crop out system info?

 

[aztekk]

Why would they crop out system info?

That's definitely the sketchiest part of this leak. There can be legitimate reasons to keep that hidden, but it really hurts the credibility of this. However I like to err on the side of expecting the worst, hoping for the best when it comes to CPU launches (or any high-profile tech launch). I think it is at least fair to expect Ryzen to be slower in most real world scenarios than what AMD showed us in their demo. The question is how much slower, could be very close to what AMD shows but could also be completely off with Blender being the best case scenario for the Zen architechture.

 

[ManofGod]

So I am not surprised it could be significantly slower than the i7-6900k.

Right, so if we are to believe this score, it will be slower than a current FX series of processors. Cropped image says nothing at all.

Edit: Remove the 8 additional threads.

 

[aztekk]

Right, so if we are to believe this score, it will be slower than a current FX series of processors. Cropped image says nothing at all.

Which FX series CPU scores 1188 on Cinebench R15?

 

[Shintai]

Right, so if we are to believe this score, it will be slower than a current FX series of processors. Cropped image says nothing at all.

Edit: Remove the 8 additional threads.

How did you make that conclusion? All the scores are higher than any FX.

 

[ManofGod]

Which FX series CPU scores 1188 on Cinebench R15?

Ryzen has 8 cores / 16 threads, FX only has 8 cores and some shared resources. Eliminate those 8 extra threads.

 

[Shintai]

Ryzen has 8 cores / 16 threads, FX only has 8 cores and some shared resources. Eliminate those 8 extra threads.

Perhaps you should look at the SMT benefit. Seems to be a stressful day for you.

4670K 3.4/3.8Ghz 4/4 531
4770K 3.5/3.9Ghz 4/8 756

And Ryzen scores almost twice of a FX 8350. And not only that, it scores about the same as an 8/16 SB Xeon at around 3-3.2Ghz.

 

[aztekk]

Perhaps you should look at the SMT benefit. Seems to be a stressful day for you.

4670K 3.4/3.8Ghz 4/4 531
4770K 3.5/3.9Ghz 4/8 756

And Ryzen scores almost twice of a FX 8350.

An 8370E which is a Piledriver core CPU @ 3.3ghz (closest to Zen's 3.4) scores 529 in this benchmark. So... assuming 30% performance increase (this is a big assumption that AMD has got SMT/HT right the first time, something Intel has been perfecting for a decade) it would still only increase the score to 687. Which is much slower than 1188... his logic really doesn't add up in this case at all.

 

[Shintai]

An 8370E which is a Piledriver core CPU @ 3.3ghz (closest to Zen's 3.4) scores 529 in this benchmark. So... assuming 30% performance increase (this is a big assumption that AMD has got SMT/HT right the first time, something Intel has been perfecting for a decade) it would still only increase the score to 687. Which is much slower than 1188... his logic really doesn't add up in this case at all.

Exactly. And he can just look at 8 core 16 thread Sandy Bridge Xeons if in doubt

 

[Dullard]

1. WCCFTECH article
2. No source link in article to "Chinese forums"
3. Camera picture of monitor cropped to only show score
4. No info on sys config/clock speed

Yep, lets start making our purchase plans now

You forgot

5. Cinebench is not a good bench for AMD processors

 

[lolfail9001]

5. Cinebench is not a good bench for AMD processors

In last 5 to whatever years, every bench was not a good bench for AMD processors.

 

[Shintai]

Imagine if AMD hadn't used CB15 to show Carrizo...oh wait

 

[aztekk]

I laugh any time someone starts the "X is not a good benchmark for AMD" nonsense. Unless you can prove there was some unfair business practices like with Intel's C++ compiler some years ago, you have no standing. When you can't make a competitive CPU accuse the benchmarks of being biased... sounds a bit like how Donald Trump conducts himself.

 

[ManofGod]

I laugh any time someone starts the "X is not a good benchmark for AMD" nonsense. Unless you can prove there was some unfair business practices like with Intel's C++ compiler some years ago, you have no standing. When you can't make a competitive CPU accuse the benchmarks of being biased... sounds a bit like how Donald Trump conducts himself.

Oh goody, interject politics into this thread, train crashing derailing in progress as we speak. Personally, I wish they would just release it already so I can buy it already. However, the release and my income tax refund should occur around the same time, goody for me.

 

[cageymaru]

An 8370E which is a Piledriver core CPU @ 3.3ghz (closest to Zen's 3.4) scores 529 in this benchmark. So... assuming 30% performance increase (this is a big assumption that AMD has got SMT/HT right the first time, something Intel has been perfecting for a decade) it would still only increase the score to 687. Which is much slower than 1188... his logic really doesn't add up in this case at all.

We think of FX processor in terms of FX-9370 / FX-9590 or a FX chip OC'd to at least 4.5GHz. But I understand your point.

 

[ecmaster76]

You forgot

5. Cinebench is not a good bench for AMD processors

Ryzen is a whole new architecture so no way to know for sure what its good and bad at until NDA is lifted.

 

[Palladium@SG]

I laugh any time someone starts the "X is not a good benchmark for AMD" nonsense. Unless you can prove there was some unfair business practices like with Intel's C++ compiler some years ago, you have no standing. When you can't make a competitive CPU accuse the benchmarks of being biased... sounds a bit like how Donald Trump conducts himself.

Yup, Broadwell-E IPC and $1100 6900K performance this close to release. Somehow AMD got this amazing chip and nothing to lose by this point yet they still feel the need to play hype mind games.

 

[mysticjbyrd]

Guys, I have the real results right here. The pic here is a fraud!

Check it out,

 

[rgMekanic]

Yup, Broadwell-E IPC and $1100 6900K performance this close to release. Somehow AMD got this amazing chip and nothing to lose by this point yet they still feel the need to play hype mind games.

I havn't seen AMD doing much hype at all

 

[OrangeKhrush]

A month odd ago I was in a conversation with a person that leaked Zen results that ended up going viral, the final top outcome reached around 1300 which I still think Zen is able to achieve Haswell level and I am boyed by that.

If the final Zen maxes out around 1300 that should be inside a good margin to a Haswell based octo core, probably hae the 4c/8t around 700 odd which should put it at 4770 levels.

 

[cageymaru]

Guys, I have the real results right here. The pic here is a fraud!

Check it out,

Need a 1819 version.

 

[cocdod]

It's been obvious for months that Zen would underperform. AMD has an odd habit of only starting the hype machine when they know they won't put out something good.

 

[Palladium@SG]

It's been obvious for months that Zen would underperform. AMD has an odd habit of only starting the hype machine when they know they won't put out something good.

If AMD is actually confident on a product that will decide their fate they will be doing this by now: http://www.anandtech.com/show/3871/the-sandy-bridge-preview-three-wins-in-a-row/11

Only 3.1GHz and no turbo, yet Intel has no qualms letting Anand comparing SB to their own $1000 980X and cannibalize their entire lineup a full 4 months before release.

 

[Prav]

Ponitless leak that doesnt even have CPU-Z. Certainly not deserving of its own thread. Cant tell if bored or obsessed for garbage like this to be posted.

 

[Gideon]

This is meaningless without clock speeds, most likely on a es sample chip with low clocks, thus why they didnt show the speed or mention it. Shintai lives for anything that puts AMD in a bad light, thus why he is here commenting on it.

 

[Peppercorn]

This is exactly why I said after that AMD demo that we need to wait for retail availability / independent non cherry picked benchmarks. A lot of people here naively bought into AMD's marketing about it being comparable to a $1000 CPU. Things are never as rosy as they look, especially with a for-profit corporation who is always looking to showcase their product in the best light possible. So I am not surprised it could be significantly slower than the i7-6900k.....

So instead of, as you say, 'naively buying into AMD's marketing', it seems to me you are suggesting people should naively buy into these cherry picked benchmarks instead, with no proof as to what CPU it is, no indication if it is an ES sample, no indication of clocks, no screenshot of CPU-Z, etc etc. Yup, that makes sense....

 

[Shintai]

It's been obvious for months that Zen would underperform. AMD has an odd habit of only starting the hype machine when they know they won't put out something good.

Last success they had was Bobcat, and they let everyone test it way before release. Whenever AMD doesn't have something that can live up to expectations, we get this.

Intel knows everything about Zen, just as AMD knows everything about Skylake-EP, Cannon Lake etc. So its not them they try to keep a secret for.

 

[aztekk]

So instead of, as you say, 'naively buying into AMD's marketing', it seems to me you are suggesting people should naively buy into these cherry picked benchmarks instead, with no proof as to what CPU it is, no indication if it is an ES sample, no indication of clocks, no screenshot of CPU-Z, etc etc. Yup, that makes sense....

Think you better read the post you quoted. I am saying you should not make any conclusions or premature buying decisions based on Ryzen's performance until you can actually walk into a store, buy one and test it yourself. I've been very consistent on that point as well.

 

[Gideon]

Last success they had was Bobcat, and they let everyone test it way before release. Whenever AMD doesn't have something that can live up to expectations, we get this.

Intel knows everything about Zen, just as AMD knows everything about Skylake-EP, Cannon Lake etc. So its not them they try to keep a secret for.

You realize that is a stupid argument right? They know everything thus they should show us performance now. No company releases info until they are ready to, even in the car industry where I work they dont release info almost till right before launch.. people try to guess but no official info. You only leak what you want to show and that is par for the course bud. In fact more often then not, when a company tries to make a lot of noise about a product, it is not performing as expected.

 

[Shintai]

You realize that is a stupid argument right? They know everything thus they should show us performance now. No company releases info until they are ready to, even in the car industry where I work they dont release info almost till right before launch.. people try to guess but no official info. You only leak what you want to show and that is par for the course bud. In fact more often then not, when a company tries to make a lot of noise about a product, it is not performing as expected.

Is it? Or do you just hope for some kind of excuse?

AMD even made a big blunder and talked about Skylake-EP SKUs at SC16.

Just look at history. Every time AMD got an underperforming product they do rigged/cherry picked closed benches/demos. When they got something good, they let everyone test it with no limits.

 

[Gideon]

No they dont, Athlon was not tested early by independent reviewers. They did the same thing they did with Zen, limited benches chosen by them and compare to Intel. That is a standard launch of a product for AMD. The FX was not a standard launch and they tried hard to steer you away from its weakness of IPC. Show me a full review of the Athlon processor 1 month before launch and I will stand corrected.

 

[Peppercorn]

Think you better read the post you quoted. I am saying you should not make any conclusions or premature buying decisions based on Ryzen's performance until you can actually walk into a store, buy one and test it yourself. I've been very consistent on that point as well.

I did read the post, and that may be the way you intended it, but it is not the way it reads... "a lot of people here naively bought into AMD's marketing...." "things are never as rosy as they look" "So I am not surprised it could be significantly slower than the i7-6900k". All these quips suggest that you are dismissing everything that AMD has shown, and are promoting these obscure results in their stead. The conclusion you are promoting is that what AMD has revealed is cherrypicked, and these results are more believable.

 

[SighTurtle]

Is it? Or do you just hope for some kind of excuse?

AMD even made a big blunder and talked about Skylake-EP SKUs at SC16.

Just look at history. Every time AMD got an underperforming product they do rigged/cherry picked closed benches/demos. When they got something good, they let everyone test it with no limits.

Bobcat was 2011, its been 5 years with multiple management changes (not to mention Bulldozer), you can't really expect AMD to do the same thing.

Irregardless, unless Sweeper's source can offer more evidence this is a actual up to date chip (or even Zen), these benchmarks are as much trash as what you feel AMD's rigged benchmarks are.

 

[FrgMstr]

Jeez guys, cut out the in-thread bickering. Take it to PM.

 

[OrangeKhrush]

maybe it was actually the ES sample