News AMD unveils Ryzen processors, boasts of Intel-beating performance

TDP's are often just brackets. 1700 might have fallen on one side of the 65W divider, while 1700X might have fallen on the other side. 60W => 65W bracket, 70W=95W bracket. Other option might been what Vault-Tec said, that is more aggressive clocks, staying at higher clocks for longer time, before thermal/TDP limitation come in the play.

 

More price leaks...

http://www.guru3d.com/news-story/us...cessors-surface-as-well-starts-at-316-59.html

Also some more details on Zen architecture, mostly in japanese but also some english bits in between:

http://pc.watch.impress.co.jp/docs/column/kaigai/1043349.html#1_s.png

 

Yeah I think the expensive ones will overclock quite aggressively. According to something said somewhere at some point (of which I can not remember) it was said that depending on your cooling etc Ryzen would clock itself as high as it could.

I think that is why they did not talk about, nor set in stone, the boost clocks when they ran Blender and Handbrake. I think they were still working on it.

 

Or the in between option:
On the non X its like with Intels Turbo Booost 3.0 where it doesn't work on all cores while the X version gets it on all.

 

I agree Gareth, I think the expensive ones will do all of the work for you so you don't have to get your hands dirty. So for enthusiasts the 1700 will probably be the one to have so long as you are not scared by the overclocking part.

 

Looks like AMD are only [officially] supporting RyZen on Windows 10... linky.

*idly wonders if they'll get the same flak intel did*

 

Silly decision, given that many won't upgrade to Kaby because they don't want to leave 7. However, it was obviously more trouble than it was worth.

 

Good decision by AMD, at this point windows 7 is just legacy support.

 

Just to follow up on this: AMD had said that all Ryzen chips would be unlocked, but actually, they specifically said "multiplier unlocked", per this slide (which VT originally posted a month ago):

 

Yeah, I even went with an i5 for my "budget isn't even a thing" machine, because why pay hundreds of pounds more for MHz you're never going to use.

Also, gareth, I'll take that bet just to make things interesting. PM me when you win it otherwise I'll forget.

Come to think of it, I'd better not win as I have absolutely no idea what charity to choose...

 

What is believed to be a 1700X 8c 16t vs 6950x 10c 20t.



Edit. 8C 16T Ivy @ 3.4

 

What looks like Asus's flagship board to cost $209.

https://www.overclock3d.net/news/gp..._x370_and_b350_motherboards_have_now_leaked/1

 

https://www.overclock3d.net/news/cpu_mainboard/amd_ryzen_3dmark_physic_score_leaks/1

Take with a large pinch of salt.

 

Also logic of that "per core" is flawed. Especially in the era when you have different Turbo clocks depending on core load count, new features like that new automatic AMD OC and so on.

When you do a 8-core load and divide the result by 8, you might get a different "per core" score than if you did a 4 or 1 core load.

 

I know, hence the pinch of salt.

Edit. Just to say. With all of these benchmarks and leaked info you usually have three scenarios.

1. Total BS.
2. Totally the truth, but usually on a ES or what not.
3. Somewhere in the middle, which turns out to be correct after the launch.

It's the same every single time.

 

Again more leaked stuff but it does have actual benchmarks but as ever massive pinch of salt:

http://wccftech.com/amd-ryzen-3dmark-benchmarks-leaked-faster-core-intels-i7-6950x/

Edit: posted already by vault via another website.

Sent from my SM-N915FY using Tapatalk

 

Already posted that

https://www.overclock3d.net/news/cp...ws_are_rumoured_to_go_live_on_february_28th/1

Reviews going live on the 28th apparently. One day before I get paid /facepalm