AMD RyZen CPU Architecture for 2017

Some zen latency tests:

http://www.planet3dnow.de/cms/30520-ryzen-7-1800x-teil-2/subpage-tiefenanalyse-der-architektur/

You're wrong and you're right, but don't feel bad because it's a common mistake on the web. It actually IS up to Google to decide who is or isn't allowed to run videos on THEIR site. Now you're right that this is censorship, but it is very much legal and prudent of Google to do so.

You don't really have the right to free speech on the internet for the most part, you have the right to say whatever the site owners will tolerate. Google is a privately held company who gets to make up their own rules about their services, and people get to choose if they care to use those services based on those rules. I know it sounds dumb and hokey, but you would not BELIEVE how many times I'd have to explain to people that just because they're a member of a website it doesn't give them the right to say whatever they wanted.

Sorry, just a personal peeve of mine. Too many years as a mod/admin at too many sites having to explain that to people. My apologies for the rant, but just pointing it out. :oops:

Its not illegal being incompetent and get to not true-to-reality conclusion base on data and methodology that is not 100% correct. You don't deal with this by censoring people, you deal with this by creating a healthier ecosystem to and encourage more highly technical content and let the community judge itself which channels are better.
 
You're wrong and you're right, but don't feel bad because it's a common mistake on the web. It actually IS up to Google to decide who is or isn't allowed to run videos on THEIR site. Now you're right that this is censorship, but it is very much legal and prudent of Google to do so.

You don't really have the right to free speech on the internet for the most part, you have the right to say whatever the site owners will tolerate. Google is a privately held company who gets to make up their own rules about their services, and people get to choose if they care to use those services based on those rules. I know it sounds dumb and hokey, but you would not BELIEVE how many times I'd have to explain to people that just because they're a member of a website it doesn't give them the right to say whatever they wanted.

Sorry, just a personal peeve of mine. Too many years as a mod/admin at too many sites having to explain that to people. My apologies for the rant, but just pointing it out. :oops:
The problem isn't google deleting illegal videos or accounts from youtube. They rightfully do that all the time. The problem is that google is marking whole (legal) channels as "advertiser unfriendly" while leaving their videos online (since they are legal), and so denying them their revenue streams. That's an insidious form of censorship, but it's offtopic so I'll stop now.
 
Damn it, MicroCenter is HURTING ME!!!!

EZdKtMH.jpg


Can't wait to see their Ry5 deals! :D
 
Ended up changing my kit of Trident Z for the exact same RGB kit because of stock issues and because they didn't have anything else in stock that was 100% Samsung b-die. The G.Skill RGB kits are essentially the same exact memory kits as the non-rgb, correct? At least I'll be getting my parts a lot sooner.
 
Anandtech has a active core vs power test that might be interesting.
http://www.anandtech.com/show/11244/the-amd-ryzen-5-1600x-vs-core-i5-review-twelve-threads-vs-four/2

It sequentially loads cores, and shows some of the transitions points in AMD's architecture for XFR and turbo.
There are signs in the clocking and power behavior for why the assumption that parts with fewer cores could use the spare TDP to clock higher did not happen.
The single-core XFR state is probably approaching the ceiling for Zen's implementation in Summit Ridge, and at its highest a CCX seems to top out at ~50W for power.

There are signs of cores taking a bit more power individually in a CCX when there are idle ones. One item I cannot determine from the test is if the cores are activating in the order of physical adjacency. It's not entirely certain how much of this "slack" is the cores taking advantage of having cool silicon next to them, or if there's some restriction in how much power can be delivered to a specific region.
If there is a thermal adjacency factor, perhaps turning two cores separated by the L3 in a CCX or one core in each CCX wouldn't behave in exactly the same way. The interference of actively using both L3s in the latter case might show up.
Some sign of a core behaving a little differently based on not having any active neighbors is the transition to having the first core in the second CCX activated. At that point, the core gets a bit more power.
Perhaps changing cooling in either direction could tease out if core temp is managing this.
 
Just like NetBurst ... or wait.

Im not quite sure to understand what Nerburst ( P4 ) are doing with that.. Actually Intel quadcores have an advantage on gaming due to higher clock ( 7700), if AMD can push higher clock than the actual .. where is the problem ?
 
Im not quite sure to understand what Nerburst ( P4 ) are doing with that.. Actually Intel quadcores have an advantage on gaming due to higher clock ( 7700), if AMD can push higher clock than the actual .. where is the problem ?
It was a joke, Intel executives at that time claimed that Netburst was designed to scale to 10 Ghz...
 
:D Well in this case Ryzen efficiency is impressive first arq in a first process with tdp equal or better than Intels.

Absolutely, we all love a good underdog story. I'm just happy that we finally have competition again. Intel have been milking people with the Core i5 for too long. Finally there is an affordable 8 core / 16 thread processor on the market with more than competitive performance.
 
Back
Top