AMD Ryzen 9 9950X3D Gaming Performance CPU Review

Brent_Justice

Administrator
Staff member
Joined
Apr 23, 2019
Messages
992
Points
93
Introduction AMD is launching its new Ryzen 9 9950X3D and Ryzen 9 9900X3D CPUs, based on the Zen 5 architecture on the AM5 platform, featuring 2nd Gen AMD 3D V-Cache for Gaming and Content Creation. In this Ryzen 9 9950X3D review, we are going to focus purely on the gaming performance of the CPU. This […]

See full article...
 
Good review, thanks!

Looks exactly like I suspected.

The multi CCD penalty in gaming is still real, so it consistently performs lower than the 9800x3d in games, but on average the penalty seems less than in the past, maybe because of slightly higher clocks.

It would be interesting to see how it performs in games with one of the CCD's disabled.

I wonder how much better a 9950x3d with one CCD disabled would perform compared to the 9800x3d...

Based on the clock speeds (5.7Ghz vs. 5.2Ghz) the upper bound should be 9.6% at stock clocks, but in reality it will likely be lower, as RAM bandwidth/latency will likely hold it back.

Lets not forget that overclocking is now possible with x3D CPU's. I'm going to presume that the CCD's in the 9950x3d are higher bins to hit the 5.7Ghz clocks, but the question to me is this:

If you are a gaming extremist and overclocker, and want the best gaming performance bar none....

Is it better to buy a 9950x3d for $699 (compared to the $479 9800x3d) and disable one CCD (to get rid of the CCD penalty)?

Or, if you are going to overclock both anyway, are you likely to be limited at about the same clocks/performance.

In other words, can you get higher by overclocking a one CCD 9950x3d than by overclocking a 9800x3d.


I've already seen many 9800x3d overclocks it 5.7Ghz, so we know it can tie the stock 9950x3d, but the question is, is that 9950x3d right at its limit, or will that higher bin allow it to go even higher?

...and if it does go even higher, (6+Ghz?) will it even matter, or will AMD's fabric/memory controller design hold it back anyway?

This would be a really cool test, especially if a few different memory configurations are thrown in for good measure. Which would be faster for games on an AMD system when really pushing the clocks like this? DDR5-6400 in a 1:1 MCLK to UCLK configuration, or DDR5-8000 in a 2:1 MCLK to UCLK configuration.

The DDR5-8000 will get slightly less top level memory bandwidth due to a slightly lower UCLK, but on the flip side, with the RAM running at 4000Mhz (8000MT/s) penalties from things like refreshes are going to be much smaller, as those don't use memory controller bandwidth, and instead operate at the speed of the RAM. Of course, timings and submitting will likely make a difference too.

I have yet to see a really good review that combines all of these complexities of overclocking X3d chips, multiple CCD's and RAM settings when pushed to the limit.
 
Last edited:
It's interesting. I've seen some reviewers where the 9950x3d is steadily trading blows with the 9800x3d. Yet the reviews here are showing it largely a step below for gaming. (Well more a 1/2 step really.) Not discounting it mind you.
 
Grimlakin said:
It's interesting. I've seen some reviewers where the 9950x3d is steadily trading blows with the 9800x3d. Yet the reviews here are showing it largely a step below for gaming. (Well more a 1/2 step really.) Not discounting it mind you.
Click to expand...
What resolutions are being looked at in those reviews? Our numbers are at 1080p where CPU bound is more anything. 4k testing is probably going to be within margin of error...
 
Grimlakin said:
It's interesting. I've seen some reviewers where the 9950x3d is steadily trading blows with the 9800x3d. Yet the reviews here are showing it largely a step below for gaming. (Well more a 1/2 step really.) Not discounting it mind you.
Click to expand...

Game Selection, Manual In-Game Run-Through vs Benchmark, Area used within game for manual run-through, length of run-through, Resolution, Game Settings, System Configuration, BIOS version used, BIPO settings used, Memory used and type and speed, Windows Updates applied, Game Bar Updated, Chipset Drivers used, GPU used, System resource minimization, VBS on or off, FPS capture method, Number of run-throughs performed, Fresh data or old data, Clean install of Windows or not, verification of results, etc... name your variable.
 
Well, the 9900x3d is 679€ and the 9950x3d is 789€ around here, not as bad as the GPU's, still not good though.
 
As of 1:15pm EST, my local Micro Center has 25 in stock. Newegg and Amazon have none. A friend of mine grabbed one from Newegg and said they sold out in 9 minutes. He also got one from Micro Center. He's gonna return the one he got from Newegg.
 
DrezKill said:
As of 1:15pm EST, my local Micro Center has 25 in stock. Newegg and Amazon have none. A friend of mine grabbed one from Newegg and said they sold out in 9 minutes. He also got one from Micro Center. He's gonna return the one he got from Newegg.
Click to expand...
This morning the Dallas Microcenter had 380 9950x3d's. Currently.. maybe out I don't know.
 
I'm having a hard time deciding which one I want now. 9800x3d or 9950x3d. Considering I'll have a separate workstation, and my AM5 system will pretty much only be used for games, I'm leaning towards the 9800x3d, but the draw is strong.
 
David_Schroth said:
What resolutions are being looked at in those reviews? Our numbers are at 1080p where CPU bound is more anything. 4k testing is probably going to be within margin of error...
Click to expand...

I have to admit, your results seem a bit different than most others I am seeing now.

Did you guys confirm that you had proper core parking working during the game tests?
 
Zarathustra said:
I have to admit, your results seem a bit different than most others I am seeing now.

Did you guys confirm that you had proper core parking working during the game tests?
Click to expand...

Our results make sense compared to the other CPUs used in our comparison, specifically the 9800X3D and 7950X3D, and the advantages over the 9950X that we saw. The results make sense when you compare our own data to itself. The 9800X3D has the overall advantage, and that makes sense, if core-parking wasn't working, then the results would be far far worse for the 9950X3D, and it isn't. BTW, the 9950X3D was the first CPU we installed with a fresh install of Windows and drivers, and updates, so there is literally nothing getting in the way of the PPM drivers, and other chipset driver optimizations.

Cross-comparing versus other review sites can be misleading due to the myriad of variables I described above, that I will copy below. Game selection and where you benchmark in the game makes a big difference, including if it's a game benchmark or manual run-through. we use manual run-throughs that emphasize CPU-dependency; I cannot speak for other reviews.

Game Selection, Manual In-Game Run-Through vs Benchmark, Area used within game for manual run-through, length of run-through, Resolution, Game Settings, System Configuration, BIOS version used, BIPO settings used, Memory used and type and speed, Windows Updates applied, Game Bar Updated, Chipset Drivers used, GPU used, System resource minimization, VBS on or off, FPS capture method, Number of run-throughs performed, Fresh data or old data, Clean install of Windows or not, verification of results, etc... name your variable.
Click to expand...

Please specify which games you think are questionable.
 
Last edited:
Compared to Gamers Nexus while graphical settings are different you went with Ultra and they went with Medium. Theirs is showing the 9950x3d = to the 9800x3d. Where yours is reflecting a nearly 5fps difference. That could easily be a margin of error.

F1 24, Gamers nexus running 1080p/High as opposed to Ultra again puts them at equal AS DOES yours. So that lines up.

Hardware unboxed CP 2077 shows a 3fps lead for the 9800x3d over the 9950x3d. Not bad.

(I'm not sighting these to prove or disprove, you asked I'm just doing a bit of research because I'm a nerd.)

Hardware unboxed results for The last of us Part 1 had a 2fps lead over the 9800xed, same effectively as yours.

Hardware unboxed Starfield results are effectively = with a 1 FPS lead for the stock clock 9950x3d. 1080p Ultra.

So your tests are largely the same or margin of error differences in results. It seems other review sites are leaning slightly toward the 9950x3d being the better for gaming depending largely on the game itself. :)
 
Brent_Justice said:
Our results make sense compared to the other CPUs used in our comparison, specifically the 9800X3D and 7950X3D, and the advantages over the 9950X that we saw. The results make sense when you compare our own data to itself. The 9800X3D has the overall advantage, and that makes sense, if core-parking wasn't working, then the results would be far far worse for the 9950X3D, and it isn't. BTW, the 9950X3D was the first CPU we installed with a fresh install of Windows and drivers, and updates, so there is literally nothing getting in the way of the PPM drivers, and other chipset driver optimizations.

Cross-comparing versus other review sites can be misleading due to the myriad of variables I described above, that I will copy below. Game selection and where you benchmark in the game makes a big difference, including if it's a game benchmark or manual run-through. we use manual run-throughs that emphasize CPU-dependency; I cannot speak for other reviews.



Please specify which games you think are questionable.
Click to expand...

Sorry, didn't mean to sound like I was making accusations or anything. I don't have enough information to say one way or another.

Just that most review sets out there I have seen have the two trading blows, and maybe even having the 9950x ahead by one or two percent on average. Most conclusions out there are along the lines of "wow, isn't it amazing. AMD has solved the split CCD performance penalty issue". Your review leads to a different conclusion, so I was just curious what might be different.

Overall game selection in the test suite can certainly be a possible explanation.

Yours seems to be a little bit of an outlier, and outliers are always interesting :) You can learn from them. Understanding the "why" the general conclusion is different can lead to a better understanding of the two CPU's comparative performance.
 
Perhaps it's more that they have solved the X3D frequency issue than the split ccd performance penalty....
 
Looking on videos about the CCD performance... it seems pretty solved.... probably... but... I have an issue.

Why simple park all of those other cores for everything happening on the system? Wouldn't it make more sense to have the non game processes on the other cores to free up the gaming cores for the game stuff?
 
Grimlakin said:
Why simple park all of those other cores for everything happening on the system? Wouldn't it make more sense to have the non game processes on the other cores to free up the gaming cores for the game stuff?
Click to expand...
I'd assume that they're trying to make things better while also not making them worse.

The hardware problem is two-fold - that putting processes on the non-X3D CCD results in a lack of benefit, and that putting threads on both CCDs slows down processes considerably.

And then we have the software problem where every piece of software can approach this problem differently - or not at all, right?

I assume that this is why AMD hasn't done something relatively simple like downclocking the non-X3D CCD when a game is 'detected', which would theoretically cause processes to avoid that CCD in favor of the X3D CCD, since this would complicate things when it doesn't work. And 'not working' is about how we'd describe the release experience of the 7900X3D and 7950X3D in general.

But now it seems like AMD has gotten things within a margin of error of their single CCD parts and that's pretty cool.
 
You must log in or register to reply here.
Become a Patron!
Back
Top