SFF.Network AMD Ryzen announced, detailed and available for pre-order TODAY !

Ever since AMD first announced the work that would later be branded as Ryzen, the company has been strategically and masterfully orchestrating a narrative of dramatic change and disruption to the staid status quo of consumer and enthusiast-grade processors. Today, however, AMD has built up this performance into a crescendo, by revealing their top-performing Ryzen AM4 CPUs today.

Read more here.
 

CC Ricers

Shrink Ray Wielder
Bronze Supporter
Nov 1, 2015
2,234
2,557
The better metric to look at is not framerate, but frame delivery time variance. So far I think only PCPerspective have tested using this method.

Tom's Hardware is also usually very good at testing frame variance over time, but usually with GPU reviews. They need to do the same now.

Suffice to say, Fine Wine technology will catch up to AMD CPUs as well as the kinks are smoothed out :p
 

Kmpkt

Innovation through Miniaturization
KMPKT
Feb 1, 2016
3,382
5,936
I really hope the partners do a bang up job on the ITX boards. I'm really tempted to go Team Red if for no other reason than to support them for a job well done.

The other thing I was wondering is don't these chips use deep learning processes to optimize over time? Shouldn't that yield better results over a month or two or something than immediately on release day?
 

3lfk1ng

King of Cable Management
SFFn Staff
Bronze Supporter
Jun 3, 2016
918
1,717
www.reihengaming.com
Unfortunately, no. From what I have read, the 'learning process' is effectively expunged after each restart like volatile ram.

While the following graphs are deliberately skewed to advertise a larger advantage that what is actually being gained (just a 1.2% boost in performance), it's still clear that when the same benchmark is run over and over again, the AMD processor's predictions will provide a slight increase in overall performance each and every time.
 

ChainedHope

Airflow Optimizer
Jun 5, 2016
306
459
Check this video. Steve was a test engineer before a reviewer so he has a bit more insight compared to other tech tubers. He includes an audio snippet from a phone call he had with someone at AMD.

 
  • Like
Reactions: Soul_Est and Phuncz

MarcParis

Spatial Philosopher
Apr 1, 2016
3,669
2,784
yes he already streamed yesterday with Joker production, to discuss already about all kind of gaps between all these review...:)
I like Gamers Nexus content, very professional
 
  • Like
Reactions: Soul_Est

Phuncz

Lord of the Boards
Original poster
SFFn Staff
May 9, 2015
5,943
4,952
I agree, being the top performer in gaming is not it's goal, with 8 cores at a severely lower base/boost clock this isn't going to impress in benchmarks.
Though if the future turns out to favor more cores than single-thread performance, this might change. I like to think AMD's presence in the console market will be important for this possible future.

Some aspect many people still forget is that AMD didn't design this CPU for gaming, it's an 8-core CPU that rivals Intel's 8-core CPUs, which don't target gaming either, but content creation and workstation applications. Maybe people are confused with the G4M3RZ motherboards or blinded by the LED lights, but I'm sure I can find games that'll run better on my dual-core G3258 running at 4.8GHz too.

If it was meant to compete with the i7-7700K, it wouldn't launch at a higher price (excluding R7 1700 here) and still underperform. That's where the quad-cores need to go, possible the hex-cores if they prove to clock well.

Bottom line, I'd still be happier to invest in AMD at this point because they have released a competing product (for the i7-6900K) for a lot less money and because a monopoly in tech isn't good for anyone in the long run.
 
Last edited:

MarcParis

Spatial Philosopher
Apr 1, 2016
3,669
2,784
Only issue on Ryzen is that 4-6 cores version won't clock higher than Ryzen 7 (ie around 4Ghz max).

Even Broadwell-e is keeping advantage on Ryzen 7 because of its far better OC potential and better IPC (R7 1800X single core @4.1Ghz is matching coe i7 6900K @3.7Ghz performance)

However all these reviews is making clear that more and more game love more cores vs more core clock, especially @ higher resolution (BF1, watchdogs 2 are the best examples)

In my personal case (for my upcoming cerberus-x build), I'm balancing between core i7 6800K/6850K vs R7-1700/1700X. 6850K is best-placed as it can support dual m.2 pcie 3 4x + sli/Cf pcie 3 16x, with better OC than 6800K..but it's the most expensive...:D
 

Soul_Est

SFF Guru
SFFn Staff
Feb 12, 2016
1,536
1,928
Only issue on Ryzen is that 4-6 cores version won't clock higher than Ryzen 7 (ie around 4Ghz max).

Even Broadwell-e is keeping advantage on Ryzen 7 because of its far better OC potential and better IPC (R7 1800X single core @4.1Ghz is matching coe i7 6900K @3.7Ghz performance)

However all these reviews is making clear that more and more game love more cores vs more core clock, especially @ higher resolution (BF1, watchdogs 2 are the best examples)

In my personal case (for my upcoming cerberus-x build), I'm balancing between core i7 6800K/6850K vs R7-1700/1700X. 6850K is best-placed as it can support dual m.2 pcie 3 4x + sli/Cf pcie 3 16x, with better OC than 6800K..but it's the most expensive...:D
Here's hoping that things improve with AMD's partners so that Ryzen can really shine and run higher clocks on the 4-core and 6-core parts.
 
  • Like
Reactions: cmyk78 and TheHig

MarcParis

Spatial Philosopher
Apr 1, 2016
3,669
2,784
Well, good news, I guess I found the best deal for my Cerberus-X setup..:D (even found a new name)..:)

I'll keep it confidential as long as it's not achieved..:D
 
  • Like
Reactions: Soul_Est

TheHig

King of Cable Management
Oct 13, 2016
951
1,171
Legit Reviews has a nice R7 1700 review. It and others are saying if you don't mind a little overclocking the 1700 is all you need. 4.0 at 1.38v. Of course this was on the MSI platinum which is spendy. If you could get a similar result on a cheaper board the price perf would be tremendous. I'm really looking at R5 personally.
 

ChainedHope

Airflow Optimizer
Jun 5, 2016
306
459
This is probably my last benchmark post, but its very applicable to what Ryzen 7 is supposed to be marketed at. Check the link, they benchmark Linux with a variety of server, render, and compiling benchmarks and show how it really shines as a productivity processor.

http://www.phoronix.com/scan.php?page=article&item=ryzen-1800x-linux&num=1

Im really loving the Linux support so far from what I have seen in the linked article and others. Seems that the microcode works as far back as kernel 4.4 (havent found any testing from earlier releases) with some fault errors and very few errors in the latest revisions. Seems GCC and other compilers do need to be updated however as they arent compiling correctly for multithreaded workloads, so I'd hold off for a few weeks if you are building a Ryzen Linux box.
 

MarcParis

Spatial Philosopher
Apr 1, 2016
3,669
2,784
This is probably my last benchmark post, but its very applicable to what Ryzen 7 is supposed to be marketed at. Check the link, they benchmark Linux with a variety of server, render, and compiling benchmarks and show how it really shines as a productivity processor.

http://www.phoronix.com/scan.php?page=article&item=ryzen-1800x-linux&num=1

Im really loving the Linux support so far from what I have seen in the linked article and others. Seems that the microcode works as far back as kernel 4.4 (havent found any testing from earlier releases) with some fault errors and very few errors in the latest revisions. Seems GCC and other compilers do need to be updated however as they arent compiling correctly for multithreaded workloads, so I'd hold off for a few weeks if you are building a Ryzen Linux box.
really nice!, thanks for sharing!
 

Phuncz

Lord of the Boards
Original poster
SFFn Staff
May 9, 2015
5,943
4,952
There seem to be issues with SMT on Ryzen in Windows 10 according to a few users at Anandtech forum:

https://forums.anandtech.com/threads/ryzen-strictly-technical.2500572/page-8#post-38775732

Windows 10 - 1080 Ultra DX11:

8C/16T - 49.39fps (Min), 72.36fps (Avg)
8C/8T - 57.16fps (Min), 72.46fps (Avg)

Windows 7 - 1080 Ultra DX11:

8C/16T - 62.33fps (Min), 78.18fps (Avg)
8C/8T - 62.00fps (Min), 73.22fps (Avg)

https://forums.anandtech.com/thread...and-discussion.2499879/page-128#post-38774366

each zen thread is being registered as an individual core with its own L2 and L3 cache. i.e. totaling 136 MB cache!!. this is using Windows Sysinternals. This explains the SMT troubles in the event that a thread bounced to a HT thinking its the real deal.

To me what stands out is how this happened to be present in Windows 10 in the first place. Is this Microsoft not having the needed update out yet ? Or did AMD fumble in sending the required information to Microsoft ? Either way, this will end up costing them in sales.
 
  • Like
Reactions: Soul_Est

BirdofPrey

Standards Guru
Sep 3, 2015
797
493
Who knows, but Bulldozer suffered exactly the same problem when it launched: Windows was treating each core equally rather than detecting cores with shared resources.
 
  • Like
Reactions: Soul_Est