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

[Abwx]

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.

Fritzchess should display the number of kilonods/second, either in ST or in MT, so what are exactly the numbers you posted..??.

 

[KTE]

Clock for clock, Zen should be on Haswell/Broadwell Level.

'should'... There really is no should.

We hope, is so far the best argument for ST.

Sent from HTC 10
(Opinions are own)

 

[PhonakV30]

hmm I see this numbers :

ST : 36.86
MT : 17693

 

[Arachnotronic]

The first number is "Relative Speed" compared to a Pentium 3 1GHz. Second number is Kilo Nodes per second.

 

[Nothingness]

Fritzchess should display the number of kilonods/second, either in ST or in MT, so what are exactly the numbers you posted..??.

The score sweepr used is the index (speed relative to P3). The picture has the kNPS: 17693.

EDIT: Am I slow or what?

 

[Arachnotronic]

It says there: P3 1.0 GHz "something" 480.

Yep, Pentium 3 1GHz (480 kilo nodes per second). The first number is the score of your CPU relative to that baseline, second score is the actual kilo nodes per second.

 

[Sweepr]

Fritzchess should display the number of kilonods/second, either in ST or in MT, so what are exactly the numbers you posted..??.

i7-6900K - 24.545
i7-6800K - 19.576
Ryzen 8C/16T - 17.693
i7-6700K - 16.239

i7-7700K would be in a dead heat with Ryzen here. Wow at i7-6900K 8C/16T - 38.6% faster.

 

[KTE]

i7-6900K - 24.545
i7-6800K - 19.576
Ryzen 8C/16T - 17.693
i7-6700K - 16.239

i7-7700K would be in a dead heat with Ryzen here. Wow at i7-6900K 8C/16T - 38.6% faster.

This is branchy int code?

Sent from HTC 10
(Opinions are own)

 

[Nothingness]

I expect Fritz Chess to be very hard on branch prediction. But most branch predictors would have a hard time because many of the branches will have a random direction.

 

[Abwx]

hmm I see this numbers :

ST : 36.86
MT : 17693

At 17693 Kn/s this would be slightly better than a 4790K, since Fritzchess is INT code based it couldnt beat a 6900K in whatever handbrake test if this result was genuine.

 

[Nothingness]

At 17693 Kn/s this would be slightly better than a 4790K, since Fritzchess in INT code it couldnt beat a 6900K in whatever handbrake test if this result was genuine.

Are you really trying to imply that all integer workloads have the same characteristics?

 

[rainy]

i7-6900K - 24.545
i7-6800K - 19.576
Ryzen 8C/16T - 17.693
i7-6700K - 16.239

i7-7700K would be in a dead heat with Ryzen here. Wow at i7-6900K 8C/16T - 38.6% faster.

You have no link, nor confirmation from other sources, however result is of course valid.
How your Wow isn't a trolling attempt?

 

[Sweepr]

You have no link, nor confirmation from other sources, however result is of course valid.
How your Wow isn't a trolling attempt?

The only trolling attempt here are your agressive replies when the leaked results don't match your expectations based purely on hype and controlled data provided by AMD. You have the option not to read this thread if you don't want to see leaked results or their respective comparisons. Nobody is forcing you to believe they are genuine.

Spoiler

 

[KTE]

At 17693 Kn/s this would be slightly better than a 4790K, since Fritzchess in INT code it couldnt beat a 6900K in whatever handbrake test if this result was genuine.

That link shows Zen losing to a i7-3930k.

Sent from HTC 10
(Opinions are own)

 

[Glo.]

The only trolling attempt here are your agressive replies when the leaked results don't match your expectations based purely on hype and controlled data provided by AMD. You have the option not to read this thread if you don't want to see leaked results or their respective comparisons. Nobody is forcing you to believe the are genuine.

So how you will comment on the presentation on New Horizon, and benchmarks for Blender and Handbrake? Does it contradict the scores, there or not?

 

[Nothingness]

I expect Fritz Chess to be very hard on branch prediction. But most branch predictors would have a hard time because many of the branches will have a random direction.

Answering to myself, sorry...

This is old but confirms my expectations: http://www.anandtech.com/show/2386/11

 

[mikk]

Isn't Cinebench biased against AMD? I remember there being a big stink about it, and Agner being rather miffed. http://www.agner.org/optimize/blog/read.php?i=49#49. Sure, Intel can't require benchmarks to have a biased cpuid check, but that doesn't stop 'em pushing for it.

We need open source benchmarks.

No. Cinebench is a good Benchmark for a general view which reflected an average, also it scales nicely over many cores, which favours CPUs with many cores, even AMD used this in the past. Cinebench gave a good overview in the past. Poor CPUs like Bulldozer did poorly because it was a poor CPU.

 

[rvborgh]

if Zen scales like it did with Blender (relative to the performance of K10 cores)... it should score 1592cb in CB R15 at 3.4 Ghz... that leaked score has to be on an ES with some feature not quite working and at lower clocks.

 

[Abwx]

Are you really trying to imply that all integer workloads have the same characteristics?

Granted there s no uarch that is favoured by compiler or special instructions the results should be in a same window, besides the FX8350 is at 13168 kn/s, given that Zen has SMT this would imply that the ST perf is about the same as the FX.

Just compare the ST perf of the FX to HW for instance, the difference is 41.5%, yet anothe hint that Zen has the same ST perf as the FX in this alleged review that show no CPUZ pic, or anything that would prove that it s a Zen used at stock settings.

Also there s no branching issue overall, if we look at more recent chess games the FX beat even HW, so Fritzchss is a case favourable to Intel but not to the point of producing such paradoxal numbers.

http://www.hardware.fr/articles/946-5/performances-applicatives.html

Houdini and Stockfish are currently the best chess games.

 

[rainy]

The only trolling attempt here are your agressive replies when the leaked results don't match your expectations based purely on hype and controlled results provided by AMD. You have the option not to read this thread if you don't want to see leaked results or their respective comparisons. Nobody is forcing you to believe the are genuine.

Maybe I have missed something but as far as I remember, you're posting only leaks which appear to be bad for AMD.
You've got definitely a "special" talent for that./s

 

[Sweepr]

Xeon E5-2670 - 8C/16T Sandy Bridge - 2.6 GHz base clock but manages to Turbo up to 3.0 GHz with all cores active:

At 3.4+ GHz it would easily match Ryzen's leaked score, but then again we don't know the clockspeeds or specifics for the AMD system.

 

[Abwx]

Maybe I have missed something but as far as I remember, you're posting only leaks which appear to be bad for AMD.
You've got definitely a "special" talent for that./s

Not counting "the controlled results provided by AMD", as if a "source" that give no CPUZ screenshot and no pic of the CPU could be more reliable, at least with AMD we are sure that it s Zen at the specified frequency, here it could be datas made up by some viral marketers..

 

[rainy]

Houdini and Stockfish are currently the best chess games.

Stockfish is a top freeware engine, Komodo and Houdini are best commercial ones.

 

[Arachnotronic]

Xeon E5-2670 - 8C/16T Sandy Bridge - 2.6 GHz base clock but manages to Turbo up to 3.0 GHz with all cores active:

At 3.4+ GHz it would easily match Ryzen's leaked score, but then again we don't know the clockspeeds or specifics for the AMD system.

Nice find!

 

[raghu78]

Maybe I have missed something but as far as I remember, you're posting only leaks which appear to be bad for AMD. You've got definitely a "special" talent for that./s

You are spot on. If anything shows AMD products in a poor light you can bet he will be the first to post it. It does not matter if its CPU or GPU or any other product. The next few weeks and months will see a lot more of such rubbish. I am just going to wait for the actual product reviews.