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

[KTE]

OK, let's assume it's correct, on par with i7-6850K in Cinebench R15.

As we see, I'm afraid the 'on par with 6850k' statement is not enough to justify anything, because there's very limited difference between 5960X(8C) and 6850k(6C), not to mention we don't know the exact frequency of Zen.

BTW, it's about twice as fast as FX8350 which clocked at 4Ghz if true.

edit: I dislike Cinebench, anyway.

I would rather they be 20% behind 6850K than 20% behind 2600k, as Bulldozer was upon launch in that bench

20% behind 6850K in ST is what I'm expecting... No I don't think that's IPC because they woan't reach BD-E clocks.

Sent from HTC 10
(Opinions are own)

 

[Arachnotronic]

That rumor sais it can OC to 4.2ghz fairly easily with conventional cooling. That doesn't explain much, but at least it looks like it won't get bogged down below 4ghz like many thought it would. Even if the per core performance at 4.2 matched my 3930k@4.6, to be honest that's still pretty fast actually, especially with 8 cores. Well, its fast for a 5 year old CPU anyway, but its still a good deal at $300. That price means everything.

Bogg, wanna try something interesting? Download Geekbench 4, downclock your CPU to 4.2GHz, run the test, and post a link to the results.

I want to see how SNB @ 4.2GHz does in this test and how it compares to the known XV @ 4.2GHz results that are out there.

 

[Timmah!]

Well if the top model at 500 USD is performance wise comparable to i7 6850K, at least i wont hate myself for not waiting and getting exactly i7 6850K recently

Silver linings.

 

[Stormflux]

At those prices, I could be in a market for a dual socket desktop workstation. Have AMD historically locked their multi-socket capabilities behind a professional (expensive) line like the Xeons?

 

[DrMrLordX]

[rumor] January 17th release date for ZEN ??

That would be consistent with AMD press releases that walked back the Q4 2016 release date to a CES showing followed by commercial availability. They'll launch at CES, then start selling limited quantities of the SR7 with some x370 boards.

The only people that got screwed were the folks looking forward to Bristol Ridge + B350. *sniff*

 

[raghu78]

The rumours are so contradictory. How can 8C/16T Zen with 3.2-3.3 Ghz base and 3.5 Ghz boost be competing with 6C/12T 6850k Broadwell in Cinebench when AMD already demonstrated a 8C/16T edging out a 8C/16T Broadwell clock for clock at 3 Ghz. 6850k has a 3.8 Ghz turbo. 10% higher turbo cannot makeup for 25% lesser cores.

http://ark.intel.com/products/94188/Intel-Core-i7-6850K-Processor-15M-Cache-up-to-3_80-GHz

I think the earlier chiphell slide gives us an idea of pricing and performance. 4C/8T Zen at USD 200 going up against 4C/4T core i5 , 6C/12T Zen at USD 300 going up against 4C/8T and 8C/16T priced at USD 400-500 going up against 6900k. This looks more likely as long as AMD can deliver on OC headroom of 4.2 Ghz consistently on all 8C/16T chips and IPC in the Broadwell range.

 

[Arachnotronic]

NewThe rumours are so contradictory. How can 8C/16T Zen with 3.2-3.3 Ghz base and 3.5 Ghz boost be competing with 6C/12T 6850k Broadwell in Cinebench when AMD already demonstrated a 8C/16T edging out a 8C/16T Broadwell clock for clock at 3 Ghz. 6850k has a 3.8 Ghz turbo. 10% higher turbo cannot makeup for 25% lesser cores.

Because the Blender benchmark isn't Cinebench and was probably a best-case scenario for AMD's CPU architecture.

Seriously, +40% IPC over XV is not Broadwell territory, it is Sandy Bridge. I don't think it will be on average Broadwell class, just in cherry picked scenarios.

If it were, AMD would have probably seeded one of these chips to the press to test and benchmark to build excitement. Instead, they are remaining mum. I wonder why.

 

[raghu78]

Because the Blender benchmark isn't Cinebench and was probably a best-case scenario for AMD's CPU architecture.

Seriously, +40% IPC over XV is not Broadwell territory, it is Sandy Bridge. I don't think it will be on average Broadwell class, just in cherry picked scenarios.

If it were, AMD would have probably seeded one of these chips to the press to test and benchmark to build excitement. Instead, they are remaining mum. I wonder why.

oh i got confused between Blender and Cinebench. If IPC is sandy-ivy range the product stack pricing for SR will be USD 150 - USD 300 . It would be disappointing because 8C/16T Zen will not be competing at the high end against 8C/16T Broadwell. If IPC is Broadwell range the product stack pricing will be USD 200 - USD 500.

 

[AtenRa]

Because the Blender benchmark isn't Cinebench and was probably a best-case scenario for AMD's CPU architecture.

Seriously, +40% IPC over XV is not Broadwell territory, it is Sandy Bridge. I don't think it will be on average Broadwell class, just in cherry picked scenarios.

If it were, AMD would have probably seeded one of these chips to the press to test and benchmark to build excitement. Instead, they are remaining mum. I wonder why.

And Cinebench is a best-case scenario for Intel.

I would wait for more results than those two benchmarks to see ZENs real performance.

 

[mikk]

Cinebench is all but not a best case for Intel. Sandy Bridge IPC with clock frequencies lower than Skylake/Kabylake is a disaster if true. Because just like Bulldozer AMD has to rely on a higher core count, per core it would result in a huge disadvantage for them. All deficits remain the same in this case. For mobile there isn't even a core advantage for AMDs counterpart, Intel starts shipments of 4 core SKUs from 15W next year. And for S1151 desktop there is a core increase in sight.

 

[Arachnotronic]

And Cinebench is a best-case scenario for Intel.

I would wait for more results than those two benchmarks to see ZENs real performance.

Just take a GB4 XV result and multiply by 1.4x. You will see results about in-line with Sandy Bridge.

 

[KTE]

oh i got confused between Blender and Cinebench. If IPC is sandy-ivy range the product stack pricing for SR will be USD 150 - USD 300 . It would be disappointing because 8C/16T Zen will not be competing at the high end against 8C/16T Broadwell. If IPC is Broadwell range the product stack pricing will be USD 200 - USD 500.

If IPC and MT performance is Broadwell territory, opening stack will be >300-700. Minimum.

Basic economics in corps.

If any lower, that's performance dependent.

Phenom Agena 9600 was $283 at launch, higher than Q6600... Which beat it handedly. Overpriced by $83.
But Intel ended up throwing out the $1000 QX9770 just to show how little AMD is competing.

<$300 opening stack would mean a poor, noncompetitive product. Same as BD was.

Sent from HTC 10
(Opinions are own)

 

[lifeblood]

The rumor of Zen being labeled SR 3,5 & 7 with the bottom chip costing 1500 RMB is actually good news for Zen.

People buy based on bang for their buck. What constitutes “Bang” varies depending on user and scenario. For a laptop performance per watt is critical while it’s not as important on a desktop. But, simply put, AMD cannot charge more bucks for an “equal bang” CPU than Intel does (and expect to sell any). At the $200 price point we have the i5-6500 which is no slouch CPU. So assuming the rumor is correct, AMD is claiming their low end SR3 offers equivalent bang for the buck that an i5-6500 offers. How they do it, moar cores, faster clocks, better IGP, etc, is another question, but the fact they even think SR3 is worth that does give me reason to hope.

 

[master_shake_]

If IPC and MT performance is Broadwell territory, opening stack will be >300-700. Minimum.

Basic economics in corps.

If any lower, that's performance dependent.

Phenom Agena 9600 was $283 at launch, higher than Q6600... Which beat it handedly. Overpriced by $83.
But Intel ended up throwing out the $1000 QX9770 just to show how little AMD is competing.

<$300 opening stack would mean a poor, noncompetitive product. Same as BD was.

Sent from HTC 10
(Opinions are own)

um no.

The mainstream Core 2 Quad Q6600, clocked at 2.4 GHz, was launched on January 8, 2007 at US$851 (reduced to US$530 on April 7, 2007)

https://en.wikipedia.org/wiki/Kentsfield_(microprocessor)

it dropped price later on but at rellease it was a way more than a 9600

 

[Arachnotronic]

um no.


https://en.wikipedia.org/wiki/Kentsfield_(microprocessor)

it dropped price later on but at rellease it was a way more than a 9600

Bro...

The problem is, and I hate to ruin the surprise here, Phenom isn't faster than Intel's Core 2 Quad clock for clock. In other words, a 2.3GHz Phenom 9600 will set you back at least $283 and it's slower than a 2.4Ghz Core 2 Quad Q6600, which will only cost you $269. And you were wondering why this review wasn't called The Return of the Jedi.

http://www.anandtech.com/show/2378

 

[rainy]

I really wish that some of you Intel fans could openly admit that you don't want to see more competitive AMD in CPU space at all - hypocrisy is one of those things which annoy me like hell.



Inflammatory post. If you don't like certain members who post, then put them on ignore.


esquared
Anandtech Forum Director

 

[AtenRa]

Cinebench is all but not a best case for Intel. Sandy Bridge IPC with clock frequencies lower than Skylake/Kabylake is a disaster if true. Because just like Bulldozer AMD has to rely on a higher core count, per core it would result in a huge disadvantage for them. All deficits remain the same in this case. For mobile there isn't even a core advantage for AMDs counterpart, Intel starts shipments of 4 core SKUs from 15W next year. And for S1151 desktop there is a core increase in sight.

Firs of all CB IS best case for Intel, CB is highly Intel optimized, it cant be any better for Intel.

Secondly, we already know that ZEN IPC will not reach SKL/KBL, so we already know AMD will have to give more for less. What we are waiting for to see is how much closer they have managed to be this time and at what price.

Thirdly, 1151 desktop core increase its coming from 2018. So not really relevant for Q1 2017 ZEN release.

 

[Phynaz]

I really wish that some of you Intel fans could openly admit that you don't want to see more competitive AMD in CPU space at all - hypocrisy is one of those things which annoy me like hell.

Put them on ignore if they bother you.

 

[moonbogg]

Bogg, wanna try something interesting? Download Geekbench 4, downclock your CPU to 4.2GHz, run the test, and post a link to the results.

I want to see how SNB @ 4.2GHz does in this test and how it compares to the known XV @ 4.2GHz results that are out there.

Today is Friday and I got NOTHING TO DO TONIGHT besides whore myself out for science. It will be done.

 

[Artorius]

Just take a GB4 XV result and multiply by 1.4x. You will see results about in-line with Sandy Bridge.

Didn't they say some months ago that the IPC % increase wasn't counting the performance increase due to SMT? If so multiplying XV results by 1.68~1,82 would be more realistic. But that's wishful thinking.

 

[TheELF]

Didn't they say some months ago that the IPC % increase wasn't counting the performance increase due to SMT?

No they didn't,you have to be very very carefull whenever you are listening to tecnical stuff,they said that this 40% + will be fully available to one single thread running on one single core,that's completly different,if the core runs at 100% usage givving you the full 40% benefit then there is zero extra(idle) commands that SMT could be using.
40% + is the max IPC increase in the best case scenario.

 

[inf64]

No they didn't,you have to be very very carefull whenever you are listening to tecnical stuff,they said that this 40% + will be fully available to one single thread running on one single core,that's completly different,if the core runs at 100% usage givving you the full 40% benefit then there is zero extra(idle) commands that SMT could be using.
40% + is the max IPC increase in the best case scenario.

Are you following this thread closely or not?
What you said is simply not true. AMD did address this very question on the Hot Chips Q&A session:
https://mobile.twitter.com/Daniel_Bowers/status/768270633125806081

Question: Did 40% uplift on Zen IPC include boost from SMT? AMD: No, that was just 1-thread improvement. #HotChips

So SMT comes ON TOP of the 40% uplift (in MT workloads naturally).

 

[raghu78]

If IPC and MT performance is Broadwell territory, opening stack will be >300-700. Minimum.

Basic economics in corps.

If any lower, that's performance dependent.

Phenom Agena 9600 was $283 at launch, higher than Q6600... Which beat it handedly. Overpriced by $83.
But Intel ended up throwing out the $1000 QX9770 just to show how little AMD is competing.

<$300 opening stack would mean a poor, noncompetitive product. Same as BD was.

Sent from HTC 10
(Opinions are own)

Zen SR will be competing against Kabylake 4C/4T , 4C/8T and Broadwell 8C/16T when it launches in Q1 2017. In H2 2017 Zen will SR will have to contend with the monster Skylake HEDT. Even in best case scenarios with Broadwell IPC the 4C/8T Zen will have to contend with 10-15% higher IPC of Kabylake core i5 combined with roughly 20% higher OC headroom even if Zen SR can overclock to 4-4.2 Ghz as Kabylake looks to be able to easily hit 5 Ghz . So there is no way that AMD can price 4C/8T Zen even on par with core i5 kabylake unlocked as the kaby core i5 will pretty much dominate in the majority of desktop workloads which use upto 4 threads. The few apps which use 8 threads will still see core i5 kabylake come out pretty much ahead given the 20% higher max OC and 10-15% higher IPC. Zen's SMT is unlikely to give above 20% perf increase. So yeah there is no way 4C/8T Zen can launch for any price close to USD 300.

 

[AtenRa]

Zen's SMT is unlikely to give above 20% perf increase.

What makes you say this ??

 

[TheELF]

Are you following this thread closely or not?
What you said is simply not true. AMD did address this very question on the Hot Chips Q&A session:
https://mobile.twitter.com/Daniel_Bowers/status/768270633125806081



So SMT comes ON TOP of the 40% uplift (in MT workloads naturally).

Did you even read what I wrote? How can SMT come on top if the top is already reached?
SMT is about throughput anyway (MT as you stated) and not about speed so it doesn't even matter for finding out how fast they will be.