AMD Ryzen 9 5900X and Ryzen 9 5950X Review

Ooo that looks posh !

If I didn't have my road bike I would get a downhill. Last thing I need is another big bike that's hard to hide tho. The 20s all go up through the attic opening very easy. 29" rims on the other hand don't.
 
How the could Intel not do anything to counter this? (and the same to Nvidia)
Maybe the FBI should start snooping around in the back of Intels development department, looking for newly dug holes about 9 feet deep :D

6 years on 14nm is just plain not good enough :)
 
How the could Intel not do anything to counter this? (and the same to Nvidia)
Maybe the FBI should start snooping around in the back of Intels development department, looking for newly dug holes about 9 feet deep :D

6 years on 14nm is just plain not good enough :)
Depending on IPC gains of 11th gen, they might become the marginally better gaming processor - if you look at reviews which supply worst case frame times / 1% lows, they're still rather competitive with Ryzen, though average FPS is a different story, and a much more marketable statistic.

Now, if they manage that, seeing that they're only shooting for maximum of 8/16 cores, they might aim to become the budget choice of CPUs. Which I doubt they have the guts to do, since they'd probably stubbornly market Intel as a premium product. But until they actually manage a node shrink, they're not looking good, since I also reckon that whatever Intel releases next is going to be undercut by non-X Ryzens.

Nvidia? Well we don't even have third party reviews of Big Navi yet. Early to call that a slam dunk.
 
Depending on IPC gains of 11th gen, they might become the marginally better gaming processor - if you look at reviews which supply worst case frame times / 1% lows, they're still rather competitive with Ryzen, though average FPS is a different story, and a much more marketable statistic..


Sorry, but I just have to ask, what are all those 1% lows etc that all reviewers put in their reviews?...
 
Sorry, but I just have to ask, what are all those 1% lows etc that all reviewers put in their reviews?...
Well it helps to look at performance not through FPS but by frame times, from which FPS is derived from. Stable 60fps means that every single frame lasts 16.66.. milliseconds, but reality is that frames aren't evenly paced.

So some reviewers post the 1% lows and .1% lows. 1% means that 1% of frames lasted a certain amount or longer, and .1% lows means that 0.1% of frames lasted a certain amount or longer.

Why this is meaningful is that with a decently performing gaming computer the average FPS does feel good already, but if there are frametime spikes, you do certainly feel that. Let's say a .1% low is 33ms, well that means you're dropping to 30fps every thousand frames. In a mouse controlled first person shooter that at least throws me off. And if one of every thousand frames seems like a rare occasion to you, keep in mind that you're ideally getting a hundred frames a second.
 
Sorry, but I just have to ask, what are all those 1% lows etc that all reviewers put in their reviews?...

https://www.gamersnexus.net/guides/...PS means,latency between each frame delivered.


This presents three key metrics: Average FPS, 1% low, and 0.1% low. Because minimum and maximum FPS are useless outliers, we rely more heavily on 1% and 0.1% lows as our indicators of frame dips and lag. 1% time low FPS means the average of the lowest framerates 1% of the time, a noticeable dip percentage when gaming.
 
Well it helps to look at performance not through FPS but by frame times, from which FPS is derived from. Stable 60fps means that every single frame lasts 16.66.. milliseconds, but reality is that frames aren't evenly paced.

So some reviewers post the 1% lows and .1% lows. 1% means that 1% of frames lasted a certain amount or longer, and .1% lows means that 0.1% of frames lasted a certain amount or longer.

Why this is meaningful is that with a decently performing gaming computer the average FPS does feel good already, but if there are frametime spikes, you do certainly feel that. Let's say a .1% low is 33ms, well that means you're dropping to 30fps every thousand frames. In a mouse controlled first person shooter that at least throws me off. And if one of every thousand frames seems like a rare occasion to you, keep in mind that you're ideally getting a hundred frames a second.



Yep, this is exactly as my thoughts on the matter... Chinese to me :huh:
 
Interesting, didn't realise GN reported the average the worst 1%, instead of the cutoff point.


Also, IMO if 1% low is noticable, you encounter that so often that the game isn't running well.


Edit: Well they demonstrate FRAPS in the article, which does report the cutoff point, not the average.
 
Yep, this is exactly as my thoughts on the matter... Chinese to me :huh:
Well if you run a game for 10 seconds, and for 9 seconds you get 100frames per second, and then it freezes for the last second entirely.


That means 901 frames in 10 seconds, or 90.1 FPS on average. Which on surface sounds pretty good! But clearly the game didn't run well since it froze for a whole second. However, the .1% low for that was 1 FPS or 1000ms. So the metric highlights if the game has stutters.
 
Feels like this is more of a GPU discussion now..

The lower frametimes can be addressed with AGESA and BIOS updates. Not that'll it'll entirely remove the problem but considering AMD has said 3800-4000mhz memory is not possible right now and they need more time with optimization of AGESA to maintain a 1:1 infinity fabric ratio at those speeds, it's likely they have more optimizations in general to address.
 
At high refresh rates, frametime percentiles are very much a CPU discussion. And catching up Intel in frame times is a big deal, especially since they achieved it with chiplet architecture.
 
Back
Top