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

Page 219 - 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.

CentroX

Senior member
Apr 3, 2016
351
152
116
Yes. This will be the downfall I am afraid. "Neural net" is not as good as Intel's. Its not much better than a FX

Also KL does not have lower score when you consider it has half as much cores compared to 5960x
lol downfall? Seriously ryzen destroys in 8/10 tests and performs average in 2. And people focus on the negative only. :unamused:

It is performing way above expectations.
 

cytg111

Lifer
Mar 17, 2008
23,046
12,712
136
So a very different architecture than Intel's core indeed: at times better than Haswell (first two test), similar or worst in some cases. I wouldn't take take passmark as absolute benchmark value anyway given how it shows Haswell with better single thread performance (same clocks) than Kabylake in the slides above, we know that never happens in 99% of apps, so yeah it's a good indicator for Zen but that's all.
Anyone worried by prime numbers score? Branching etc, compared to brute calculus like integer test, where even old FX is shown to beat Kabylake already?
avx2 is first thought..
 

lobz

Platinum Member
Feb 10, 2017
2,057
2,856
136
The difference is at 144fps most games need to add artificial motion blur so you don't get a headache as your brain can't process or your eyes track that fast. The other difference is a synched 60 hz IPS panel's image is just going to look a shit load better than a washed out crappy viewing angle 144hz TN panel.

This is the single stupidest thing ever posted in this thread.
Also one small advice: when you game (why else would you get a 144 Hz screen), I suggest you sit in front of your monitor, not trying to watch it from the sides.
I play CS:GO w/o motion blur @ 144, and it is more than a <redacted profanity>better this way than it was on my much more nicely colored 60 Hz IPS screen. That, I use to lay on the couch and watch movies.
But your brain processing headache thingy.... you shouldn't use your personal shortcomings to make them up as universal facts.

Profanity is not allowed in the technical forums
Markfw
Anandtech Moderator
 
Last edited by a moderator:

Agent-47

Senior member
Jan 17, 2017
290
249
76
lol downfall? Seriously ryzen destroys in 8/10 tests and performs average in 2. And people focus on the negative only. :unamused:

It is performing way above expectations.
Downfall may have been a rather strong word, but coming from BD anything will be above expectations.

I'm just trying to dampen the hype here. People will always look at the negative. And coming from research industry, I will want to know why its performing lower in PN before buying it for a workstation where computation time is more important than price/perf. It may have been memory, but then please run a bench with the ryzen timings for ram and confirm it.

Higher you fly, harder you will fall :)
 

imported_jjj

Senior member
Feb 14, 2009
660
430
136
Pretty weird that so many folks seem to factor in the base clocks for Ryzen against, in some cases, single core turbo for Intel.
The 3DMark scores, accurate or not, are suggesting 6-11% behind Kaby Lake, if you factor in the all cores turbo clocks for everybody.

So above the 200$ price point, AMD wins with more cores, even in gaming and anyone saying differently,deserves to buy an Intel. You are not gonna be bottlenecked with either one in a realistic scenario. If some person has a 4k display or a GTX1080 and games at 1080p, that person needs medication not a new PC. At any display res that fits the GPU, there won't be a downside from more cores but there can be an upside in games that scale and a larger long term upside as more games start to scale. Aside from the upside in apps, not sure why some pretend that PCs are just for gaming.
Bellow 200$, AMD would likely match or beat Intel in ST (vs i5 7500 and 7400), if we include ST turbo and ofc offer more cache and threads.

Next year, If Zen+ gains a bit, they can catch up. Intel's core is very mature by now and their 15+% claimed gain is not in ST perf, it's overall with some impact from GPU and video encoder. We are gonna see again a minimal boost in ST, mostly from clocks. AMD with a new core might get 10-15% IPC boost with Zen+
 

itsmydamnation

Platinum Member
Feb 6, 2011
2,743
3,071
136
So interesting stuff, after about 25 fails to post and about 15 hard locks i found something that worked higher then DDR3-2000

i had to go to 2133 13-14-13 CR T3
Code:
##2133 mhz 13-14-13 CR T3 ....lol ( 12ns)
CPU Mark This Computer    10749
Integer Math This Computer    19761
Floating Point Math This Computer    8384
Prime Numbers This Computer    31.1
Extended Instructions (SSE) This Computer    226.9
Compression This Computer    15186
Encryption This Computer    2134
Physics This Computer    538
Sorting This Computer    8811
CPU Single Threaded This Computer    2377

SO it appears that all benchmarks outside of Prime numbers and Physics test get a higher score from more bandwidth at the expense of latency but those two tests which Zen was "poor" in massively prefer low latency. I was also able to get a 2000mhz 9-11-10 run in which i find quite interesting seeing i only just get 2133 to work at any timings and no go @ 2400....

Code:
###2000mhz 9-11-10 (9ns!)
CPU Mark This Computer    11094
Integer Math This Computer    19519
Floating Point Math This Computer    8436
Prime Numbers This Computer    34.6
Extended Instructions (SSE) This Computer    225.8
Compression This Computer    15175
Encryption This Computer    2124
Physics This Computer    634
Sorting This Computer    8839
CPU Single Threaded This Computer    2379

So here we can see that in most tests lower latency is still preferred over more throughput only tests bucking the trend are encryption, compression, int math?? But i think based off all of these results going from 14ns 2400 memory to say 10ns 3200 memory should net Zen a good 1000 points in this benchmark. I dont think 3200mhz 16-16-16 memory is anything super high end either.......


Code:
##1066 10-11-10 (18.7ns)
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




##1066 7-7-7 (13ns)
CPU Mark This Computer    9932
Integer Math This Computer    19862
Floating Point Math This Computer    8305
Prime Numbers This Computer    25.5
Extended Instructions (SSE) This Computer    227.5
Compression This Computer    15206
Encryption This Computer    2078
Physics This Computer    413.9
Sorting This Computer    8589
CPU Single Threaded This Computer    2327

###1333 8-8-8 (12ns)
CPU Mark This Computer    10140
Integer Math This Computer    17076
Floating Point Math This Computer    8182
Prime Numbers This Computer    29.1
Extended Instructions (SSE) This Computer    220.0
Compression This Computer    14959
Encryption This Computer    2040
Physics This Computer    476.2
Sorting This Computer    8557
CPU Single Threaded This Computer    2367

###1333 7-7-7 (10.5ns)
CPU Mark This Computer    10516
Integer Math This Computer    19445
Floating Point Math This Computer    8457
Prime Numbers This Computer    29.9
Extended Instructions (SSE) This Computer    228.5
Compression This Computer    15119
Encryption This Computer    2074
Physics This Computer    503
Sorting This Computer    8761
CPU Single Threaded This Computer    2375

##2000 10-11-10 (10ns)
CPU Mark 10673
Integer Math 19504
Floating Point Math 8336
Prime Numbers 31.5
Extended Instructions (SSE) 219.1
Compression 14806
Encryption 1944
Physics 597
Sorting 8577
CPU Single Threaded 2358

##2133 mhz 13-14-13 CR T3 ....lol ( 12ns)
CPU Mark This Computer    10749
Integer Math This Computer    19761
Floating Point Math This Computer    8384
Prime Numbers This Computer    31.1
Extended Instructions (SSE) This Computer    226.9
Compression This Computer    15186
Encryption This Computer    2134
Physics This Computer    538
Sorting This Computer    8811
CPU Single Threaded This Computer    2377

###2000mhz 9-11-10 (9ns!)
CPU Mark This Computer    11094
Integer Math This Computer    19519
Floating Point Math This Computer    8436
Prime Numbers This Computer    34.6
Extended Instructions (SSE) This Computer    225.8
Compression This Computer    15175
Encryption This Computer    2124
Physics This Computer    634
Sorting This Computer    8839
CPU Single Threaded This Computer    2379
 

Gideon

Golden Member
Nov 27, 2007
1,608
3,570
136
Downfall may have been a rather strong word, but coming from BD anything will be above expectations.
Well i'm not convinced ... You explicitly said that Zen was bad in prime numbers (compared to KL) because of it's branch predictor.

Just a couple pages ago, in this very thread, plenty of evidence was given how that very same prime number test test is 90% memory latency bound.

Looks more like deliberate trolling to me.
 
  • Like
Reactions: Drazick

imported_jjj

Senior member
Feb 14, 2009
660
430
136
You didn't understand the slide.

Really? lol
The slide mentions performance in SYSmark 2014 v1.5 and claims that Kaby Lake offered 15% gain in perf with the next gen offering over15%.
As for documentation on that benchmark, you can find it here https://bapco.com/products/sysmark-2014/
After you are done reading,come back and tell me again how i did not understand the slide.
Intel isn't even trying to match the configs of the platforms they show results for, they just compare their own reference platform.
Have a good day!
 
Last edited:

Greyguy1948

Member
Nov 29, 2008
156
16
91
Single thread:
Both int math and fp math are small needing only 240kB data. Now both Intel and Ryzen have included L2 so unique data in L2 is 256-32 kB for Intel and 512-32 kB for Ryzen.
This looks OK for Intel if not the icache will need some of the L2.
The old AMDs use a big but slow L2...
 

Agent-47

Senior member
Jan 17, 2017
290
249
76
Well i'm not convinced ... You explicitly said that Zen was bad in prime numbers (compared to KL) because of it's branch predictor.

Just a couple pages ago, in this very thread, plenty of evidence was given how that very same prime number test test is 90% memory latency bound.

Looks more like deliberate trolling to me.

it could be that memory or branch prediction, or any number of thing. if its inferior branch prediction, or L3 performance then that means there is nothing you can do to overcome these drawbacks.

and i have clarified my stance with the following post:

Downfall may have been a rather strong word, but coming from BD anything will be above expectations.

I'm just trying to dampen the hype here. People will always look at the negative. And coming from research industry, I will want to know why its performing lower in PN before buying it for a workstation where computation time is more important than price/perf. It may have been memory, but then please run a bench with the ryzen timings for ram and confirm it.

Higher you fly, harder you will fall :)

Happy reading :)

Hype train is breaking light speed!!!!! :cool::cool:
 
Last edited:
  • Like
Reactions: cytg111 and Drazick

JDG1980

Golden Member
Jul 18, 2013
1,663
570
136
From PassMark's website:

The Prime Number Test aims to test how fast the CPU can search for Prime numbers, reported as operations per second. A prime number is a number that can only be divided by itself and 1. For example, 1, 2, 3, 5, 7, 11 etc. This algorithm uses loops and CPU operations that are common in computer software, the most intensive being multiplication and modulo operations. All operations are performed using 64-bit integers. This test uses about 4MB of memory per core. The specific formula used for this test is the Sieve of Atkin with a limit of 32 million.​

Emphasis mine. If these modulo operations are done using the DIV/IDIV opcodes, then it wouldn't be surprising to see them poorly optimized. These are very slow instructions even on modern Intel CPUs. I doubt that this would have been a high priority for AMD's development team, and it would not be likely to have a substantial impact on a wide variety of real-world software. Most applications that need high-speed division and modulus operations use various tricks to speed this up.
 
  • Like
Reactions: Dresdenboy

zinfamous

No Lifer
Jul 12, 2006
110,509
29,090
146

Cliffs? That host is getting hammered

ah, it's retail listing with specs.

1700x = 3.4/3.8
1800x = 3.6/4.0

wow, 3.6 base as rumored earlier.

I'm tempted to get an 1800x (something that I would probably never need), but it looks like there is so much value to be had with the 1700x now, where you can hold out for the + or ++ refreshes in a year or two. A single component swap
 
  • Like
Reactions: Dresdenboy

cytg111

Lifer
Mar 17, 2008
23,046
12,712
136
I would at least want to wait and see OC potential of the different chips before pulling that trigger
 

zinfamous

No Lifer
Jul 12, 2006
110,509
29,090
146
I would at least want to wait and see OC potential of the different chips before pulling that trigger

well, yeah. I'm also waiting for later releases and a solid lineup of Mobos that are already several BIOS revisions deep before I make a decision. Zen is a mid-late summer purchase for me, along with whatever Vega brings.

I need things to work really well when I purchase them. I have maybe 4% of the skills the rest of you have and 0% of the patience dealing with early adopter issues. :D
 
  • Like
Reactions: richierich1212

richierich1212

Platinum Member
Jul 5, 2002
2,741
360
126
Yea I still remember the SATA motherboard issue that plagued Intel and their H67/P67 motherboards when Sandy Bridge was launched.

I'd like to hold off too but I know I would only be able to hold off for a couple of days at most before pulling the trigger. :D
 

Magic Hate Ball

Senior member
Feb 2, 2017
290
250
96
I would at least want to wait and see OC potential of the different chips before pulling that trigger

Hopefully some review sites will have a few days of testing before "launch" so their reviews all get posted at 9am EST at the exact same time for me to browse before pulling the trigger 5 minutes later.

I'm either going to beef up my HTPC enough that I can game smoothly on it, or I'll be upgrading my main rig. (I use it as my portable rig when I go hang at the lake in the summertime on the weekends)
 
Status
Not open for further replies.