• Guest, The rules for the P & N subforum have been updated to prohibit "ad hominem" or personal attacks against other posters. See the full details in the post "Politics and News Rules & Guidelines."

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

Page 206 - Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.
Status
Not open for further replies.

PhonakV30

Senior member
Oct 26, 2009
977
347
136
Yes, that worked fully until Excavator.
Excavator was the first design to implement the Shadow PStates. To make it unnecessary to fully redesign the whole power management, the implementation is much more simple than in Zen.
In Zen they are truly invisible and only the power management processor can access and control them.
More advanced than Intel's CPU ? I mean not possible with Software?
 

Agent-47

Senior member
Jan 17, 2017
290
249
76
Lol got banned from Toms for posting the turbo off/on state and how it appears on Passmark as N/A or only shows base clocks. Aparently if you use the "you" term you are calling out. I guess they rather promote wrong information or assumptions over someone that actually wasted 10 minutes running the test. Good riddence.
Take it as a complement. They are a thick bunch of mods and the discussion is polarized by juan and his alternate facts. :p
 

Rngwn

Member
Dec 17, 2015
143
24
36
Man i am so right it just hurts :cool:

my IVB @4.3 DDR3 1066 10-11-10-30 T2

CPU Mark This Computer 9230
Integer Math This Computer 19408
Floating Point Math This Computer 8121
Prime Numbers This Computer 19.8
Extended Instructions (SSE) This Computer 225.8
Compression This Computer 14193
Encryption This Computer 2024
Physics This Computer 359.7
Sorting This Computer 8723
CPU Single Threaded This Computer 2370

Memory Mark This Computer 2059
Database Operations This Computer 82.6
Memory Read Cached This Computer 27786
Memory Read Uncached This Computer 11720
Memory Write This Computer 6455
Available RAM This Computer 13141
Memory Latency This Computer 33.6
Memory Threaded This Computer 14229


Be gone you trolling, shill!

Would you look at prime number with your benches, 31 vs ~20 with slower ram, plus some "minor" decreases in other categories. Should start thinking about fast RAM sticks for Ryzen now.
 

itsmydamnation

Platinum Member
Feb 6, 2011
2,180
1,754
136
Would you look at prime number with your benches, 31 vs ~20 with slower ram, plus some "minor" decreases in other categories. Should start thinking about fast RAM sticks for Ryzen now.
prime numbers:
Prime Numbers This Computer 31.5 (2000 10-11-10)
Prime Numbers This Computer 25 (1066 7-7-7)
Prime Numbers This Computer 19.8 (1066mhz 10-11-10)
 
  • Like
Reactions: ultima_trev

bjt2

Senior member
Sep 11, 2016
784
180
86
Yes, that worked fully until Excavator.
Excavator was the first design to implement the Shadow PStates. To make it unnecessary to fully redesign the whole power management, the implementation is much more simple than in Zen.
In Zen they are truly invisible and only the power management processor can access and control them.
I suppose that those shadow p-states must be programmed at boot by BIOS? It seems strange to me that the CPU comes with predefined and pre-programmed p-states. I have already read of shadow p-states on the BKDG (i read the last version and talks also of BR), but i could remember bad on shadow p-states...
 

lolfail9001

Golden Member
Sep 9, 2016
1,056
353
96
prime numbers:
Prime Numbers This Computer 31.5 (2000 10-11-10)
Prime Numbers This Computer 25 (1066 7-7-7)
Prime Numbers This Computer 19.8 (1066mhz 10-11-10)
Okay, so AVX is probably irrelevant, the root is in latency (and from what i remember of prime number sieves, it makes perfect sense, the memory jumps there are ridiculous at times).
 

Crumpet

Senior member
Jan 15, 2017
745
538
96
Okay, so AVX is probably irrelevant, the root is in latency (and from what i remember of prime number sieves, it makes perfect sense, the memory jumps there are ridiculous at times).
Would you look at prime number with your benches, 31 vs ~20 with slower ram, plus some "minor" decreases in other categories. Should start thinking about fast RAM sticks for Ryzen now.
my memory is 4x4gb @3200 cas16.. is that going to be considered fast RAM or not?

Sorry if that is a dumb question, i'm a relative noob when it comes to memory.
 

itsmydamnation

Platinum Member
Feb 6, 2011
2,180
1,754
136
my memory is 4x4gb @3200 cas16.. is that going to be considered fast RAM or not?

Sorry if that is a dumb question, i'm a relative noob when it comes to memory.
3200 @ 16 cas is 10ns which i would call it better then average. High end super expensive your looking @ high 8ns. DDR4 @ 2400 with 17 cas (the Zen sample) is looking at 14ns which is slower then DDR3 1066 7-7-7 which is 13ns.

edit lol there is a user 2400, adding a space between @ and 2400
 

.vodka

Golden Member
Dec 5, 2014
1,173
1,468
136
Would you look at prime number with your benches, 31 vs ~20 with slower ram, plus some "minor" decreases in other categories. Should start thinking about fast RAM sticks for Ryzen now.
Okay, so AVX is probably irrelevant, the root is in latency (and from what i remember of prime number sieves, it makes perfect sense, the memory jumps there are ridiculous at times).

Yes, at least going by these synthetic benchmarks it seems Zen would really benefit from fast RAM. We know Skylake does scale well into the DDR4 >4000 range in real world workloads.

I don't think Ryzen's IMC will be able to go that far, but it sure can do much better than DDR4 2400 17-17-17-39. Yuck at that RAM. After all, there's eight cores feeding off dual channel class bandwidth. It would be expected to see faster RAM having an effect in real world workloads outside synthetic benchmarks.
 

inf64

Diamond Member
Mar 11, 2011
3,130
2,089
136
I think damage control is in full swing. Ryzen is clearly NOT a Skylake killer but is very close to high IPC intel cores. This is striking a nerve ;). Next year, with a base lifted this high, one can only go higher ;). 10%? 15%? 256b load/store units? 6 core CCX? AMD has managed to catch up, that is the point. With intel not having new uarch until 2018 or 2019, AMD is set to grab significant market share in all segments.
Competition is good because end user always wins ;).
 

lolfail9001

Golden Member
Sep 9, 2016
1,056
353
96
I think damage control is in full swing. Ryzen is clearly NOT a Skylake killer but is very close to high IPC intel cores. This is striking a nerve ;). Next year, with a base lifted this high, one can only go higher ;). 10%? 15%? 256b load/store units? 6 core CCX? AMD has managed to catch up, that is the point. With intel not having new uarch until 2018 or 2019, AMD is set to grab significant market share in all segments.
Competition is good because end user always wins ;).
Implying you can lose as end user in CPU business.
 

DrMrLordX

Lifer
Apr 27, 2000
16,888
5,845
136
Intel should ship Cannonlake this year. It doesn't seem like it'll show up in anything but notebooks/2-in-1s though. Things may be getting a little weird in the Intel camp with Cannonlake in Y/U format but Kabylake/Coffeelake on mainstream desktop and Skylake/Kabylake on HEDT.

That would put HEDT two generations behind the current one.

I am sure they would pull another Xeon-D (Cannonlake replacing Broadwell) if necessary.

Implying you can lose as end user in CPU business.
Which is possible, and happens often.
 

Doom2pro

Senior member
Apr 2, 2016
587
619
106
Intel should ship Cannonlake this year. It doesn't seem like it'll show up in anything but notebooks/2-in-1s though. Things may be getting a little weird in the Intel camp with Cannonlake in Y/U format but Kabylake/Coffeelake on mainstream desktop and Skylake/Kabylake on HEDT.

That would put HEDT two generations behind the current one.

I am sure they would pull another Xeon-D (Cannonlake replacing Broadwell) if necessary.



Which is possible, and happens often.
Whatever they have in the pipe is several years in the making, and the question should be asked, a couple of years ago did Intel think AMD was going to be competing at this point? Unlikely from the rumors so far... So why should the products coming to fruition now be any different than the ones of the past? Wasn't Kaby Lake supposed to be 15% better performance? Seen this movie before.
 

JDG1980

Golden Member
Jul 18, 2013
1,663
569
136
Intel should ship Cannonlake this year. It doesn't seem like it'll show up in anything but notebooks/2-in-1s though. Things may be getting a little weird in the Intel camp with Cannonlake in Y/U format but Kabylake/Coffeelake on mainstream desktop and Skylake/Kabylake on HEDT.
Expediting desktop Cannonlake probably isn't feasible, because from what we know so far, Intel's 10nm process is a disaster. It might be more reasonable to expedite Coffee Lake, which (leaks indicate) is basically just Kaby Lake with 6 cores instead of 4. The problem is that Intel is not at all an agile company; something that would take Apple 6 months would take Intel 2 years, or more.
 
  • Like
Reactions: CHADBOGA and inf64

IllogicalGlory

Senior member
Mar 8, 2013
932
337
136
Elaborate. In the worst case you just end up stuck on what you've had.
Getting the same core counts and 5-10% performance increases year over year for the same price sounds like "losing" to me. The desktop CPU market with an Intel monopoly is a boring place with a 5 year upgrade cycle and subpar improvements even then. When was the last time anyone was excited about a new CPU arch other than Zen? Sandy Bridge? Yeah, they've made a significant improvements since then, but really, Skylake was the first arch that was worth upgrading to since then - 4.5 years later.
 
Last edited:
  • Like
Reactions: looncraz

OrangeKhrush

Senior member
Feb 11, 2017
220
343
96
From a person you know that tests motherboards?
Yes, he is also a reviewer so he knows enough about Ryzen but will never tell me as he is on official warning from AMD for a leak of information that went viral on youtube. Lets just say there are the same issues as CPC is having with turbo stability and boards being awful at the moment. He says Asus and Asrock have great products ready for release though.
 
  • Like
Reactions: Dresdenboy

ultima_trev

Member
Nov 4, 2015
148
66
66
prime numbers:
Prime Numbers This Computer 31.5 (2000 10-11-10)
Prime Numbers This Computer 25 (1066 7-7-7)
Prime Numbers This Computer 19.8 (1066mhz 10-11-10)
Thanks for posting your physics and prime number scores. Thanks to that we can be almost certain that memory bandwidth is the reason Zen is behind Intel HEDT in Physics and Prime Numbers.

Faster RAM (lower latency and higher frequency) seems it would help, but most likely to still be a bit lower than Intel HEDT due to quad channel versus dual channel memory.
 

itsmydamnation

Platinum Member
Feb 6, 2011
2,180
1,754
136
Thanks for posting your physics and prime number scores. Thanks to that we can be almost certain that memory bandwidth is the reason Zen is behind Intel HEDT in Physics and Prime Numbers.

Faster RAM (lower latency and higher frequency) seems it would help, but most likely to still be a bit lower than Intel HEDT due to quad channel versus dual channel memory.
It's latency not throughput, you can see based off the two ddr3 1066 tests. I'll try get to 10ns cas at 1066 tonight and I bet the CPU score will be the same as 2000 10-11-10
 

looncraz

Senior member
Sep 12, 2011
718
1,642
136
Just to be sure i tried to run my memory @ 2000mhz 20-22-20, i cant get post, or anything with such high numbers.
So i did 1066@ 7-7-7, which 100% confirms a strong memory latency to performance of the two "poor" performing sub tests:

Code:
Ivb@4.3                       2000mhz 10-11-10    1066mhz 10-11-10    1066 mhz 7-7-7
Database Operations           106                   82.6                  89
Physics                       597                   359.7               435
ns access time time aprox   10ns               20ns               13ns
edit: typo in the data

I took my 4.5GHz i7 2600k and ran some numbers after seeing your post on SemiAccurate - totally confirmed what you said.

DDR3-1333 9-11-10-28 timings:


DDR3-1600 8-10-9-21 timings: (wouldn't work with loose timings... no idea why)


DDR3-2133 9-11-10-28 timings:


Big difference!
 

tential

Diamond Member
May 13, 2008
7,363
641
121
As I switch my PC usage more to a server based usage, Zen really is the only option. At ~$300 for 65 W, that's a dream come true. I may get this + 2 RX 480s (or equivalent GPU) to complete my gaming server. I'm excited to see how this thing performs, and how my server does in its new enclosure.
 

wembley

Junior Member
Jan 16, 2017
11
6
41
The cpu is a great option, not a fan of Intel or amd. But my concern is the platform, incompatibility, issues that can be a real pain, for example : not be able to boot from ssd. I m OK if the cpu is not the most powerful, I m not ok with stability and incompatibility that May occur. I will wait for reviews

Sent from my HTC One_M8 using Tapatalk
 
Status
Not open for further replies.

ASK THE COMMUNITY