• Some users have recently had their accounts hijacked. It seems that the now defunct EVGA forums might have compromised your password there and seems many are using the same PW here. We would suggest you UPDATE YOUR PASSWORD and TURN ON 2FA for your account here to further secure it. None of the compromised accounts had 2FA turned on.
    Once you have enabled 2FA, your account will be updated soon to show a badge, letting other members know that you use 2FA to protect your account. This should be beneficial for everyone that uses FSFT.

AMD Ryzen 9 9950X3D Review - Great for Gaming and Productivity

erek

[H]F Junkie
2FA Enabled
Joined
Dec 19, 2005
Messages
12,824
Gaming Performance
Gaming performance of the 9950X3D is impressive—it beats every single processor on the market, the only exception is the 9800X3D. At 1080p Full HD, which is of course an unrealistic resolution for the RTX 5090, AMD's newest release is almost 10% faster on average than the 9950X without 3D V-Cache. If you look at individual games, which benefit from the larger cache, the differences are even bigger. The 9950X3D is also around 10% faster than the 14900K and 22% faster than the 285K. Last generation's Ryzen 7 7800X3D is still a strong option that's just 3% slower than the 9950X3D. Once you are playing at 4K though, the bottleneck shifts from the CPU to the GPU, which means the differences between all processors get compressed a lot. Here we see only around 10% between the entries in our test group, which isn't that much. If you don't have to own the best, and want to smartly manage your budget, it might make sense to opt for a more affordable CPU and spend the savings on a faster graphics card, where you are getting more FPS for your money.”

1741701996494.png

1741702113270.png

Source: https://www.techpowerup.com/forums/threads/amd-ryzen-9-9950x3d.333793/
 
Last edited:
Just watched HUB's review, and I think it's clear at this point that the 9950x3D is the new king, both for gaming and (most) productivity tasks. It's crazy how good it is.
 
I wish the reviews already covered the performance when disabling the non v-cache CCD, and a more in depth look at the clocks of the different CCDs. Out of the box seems overall decent, but I'm sure there's games that would perform better without crossing data between the two CCDs.The proof is the few games that run slower which you can already see in some reviews. It's not as bad as 7950X3D, but still there.

Surely the clocks on the v-cache CCD are not 5.7ghz peak? That'd be a lot faster than a 9800X3D (5.2ghz peak).

In that theoretical setup it should have a very definite lead over the 9800X3D.

The 7950X3D had a 200mhz lead over the 7800X3D, so disabling the non v-cache CCD got you a slightly better 7800X3D, with even better efficiency since better silicon binning.
 
It doesn’t seem the king of gaming?

It’s slightly slower
In some games its 1% slower in other games its 1% faster. Pretty much margin of error for the most part.
It is the co king of gaming.
The only reason to buy this as always is if you really need the extra 8 core chip for productivity stuff. If you do it looks like AMD has fixed their windows scheduling issues. So you get 9950x and 9800x3d parts all in one with no downsides. Even bests the non X3D 9950 in some productivity where the extra cache does mean something.
 
3d cache version faster than the 9950x in high frequency type non-gaming task is really impressive, could make some decision easier.

Looking at review now, with all the path in, the superbe zen5 efficacy talk of its launch feel like a fever dream
 
  • Like
Reactions: ChadD
like this
It doesn’t seem the king of gaming?

It’s slightly slower @ 4K
But anybody who wants pure gaming would just get the 9800x3d, the x950’s are for people who want something Threadripper like but can’t justify or afford the money for Threadripper.

If you need something you can work and game on then this seems to be the best of both worlds sort of product.
 
I really wish these reviews would have also done a fresh test of the 7950x, since it's only one generation old and the 9950x, on release, had regression in some applications compared to the 7950x. Supposedly that's been fixed and the 9950x gotten further uplifts that are also now benefitting the 9950x3d, but it would be nice to know what the actual present difference in both gaming and productivity between the 7950x and the 9950x3d.

There was not enough of an uplift to switch from the 7950x to 7950x3d and certainly not to switch to a 9950x. However, is the performance gap between 7950x a 9950x3d sufficient to warrant an upgrade? These reviews don't make that clear.
 

AMD Ryzen 9 9950X3D Delivers Excellent Performance For Linux Developers, Creators & Technical Computing
6 Hours Ago - Processors - 31 Comments
Ahead of tomorrow's availability of the Ryzen 9 9900X3D and Ryzen 9 9950X3D CPUs in retail channels, today the embargo lifts on being able to deliver Ryzen 9 9950X3D reviews and performance benchmarks. Simply put, for Linux creators, developers, enthusiasts, and others running technical computing workloads and other similar tasks on their desktop, the Ryzen 9 9950X3D with its 16 cores / 32 threads and 144MB total cache makes for an excellent desktop CPU. In this review are around 400 Linux benchmarks looking at the captivating performance and competitive power efficiency of the AMD Ryzen 9 9950X3D.”
 
I really wish these reviews would have also done a fresh test of the 7950x, since it's only one generation old and the 9950x, on release, had regression in some applications compared to the 7950x. Supposedly that's been fixed and the 9950x gotten further uplifts that are also now benefitting the 9950x3d, but it would be nice to know what the actual present difference in both gaming and productivity between the 7950x and the 9950x3d.

There was not enough of an uplift to switch from the 7950x to 7950x3d and certainly not to switch to a 9950x. However, is the performance gap between 7950x a 9950x3d sufficient to warrant an upgrade? These reviews don't make that clear.
Hardware unboxed includes 7950x in their review
 
Now I have to decide if its worth it to step up from a 7950X3D. I lean towards no or at least not now, but I don't want to see a stock and scalping issue either. I want AMD to get their scheduling fixed on Windows and otherwise optimize stuff without precarious software defined fixes for the asymmetry (ie Xbox game bar dependency etc) and both the 7000 and 9000 X3D series should benefit from that. of course, the Linux side of things has their own scheduler that handles things differently and has evaded some of those issues it seems; I imagine it will be even easier now. Still, good to see this is a solid CPU, though I am a bit disappointed that the extra vcache was not on both CCDs.
 
Now I have to decide if its worth it to step up from a 7950X3D. I lean towards no or at least not now, but I don't want to see a stock and scalping issue either. I want AMD to get their scheduling fixed on Windows and otherwise optimize stuff without precarious software defined fixes for the asymmetry (ie Xbox game bar dependency etc) and both the 7000 and 9000 X3D series should benefit from that. of course, the Linux side of things has their own scheduler that handles things differently and has evaded some of those issues it seems; I imagine it will be even easier now. Still, good to see this is a solid CPU, though I am a bit disappointed that the extra vcache was not on both CCDs.
Personally, I don't see why anyone would go from any Zen4X3D to the same core count Zen5 X3D. Wait till Zen6, at least.
 
It's pretty incredible how every generation of X3D cpu (5X3D, 7X3D, 9X3D) brings significant improvement. I would have thought there would be a wall hit at some point...
 
It's basically a 9800X3D for gaming and a 9950X in productivity, rolled into one chip. If you only care about one or the other, there's not much reason to upgrade to this.
 
  • Like
Reactions: erek
like this
Personally, I don't see why anyone would go from any Zen4X3D to the same core count Zen5 X3D. Wait till Zen6, at least.
I'll probably end up doing that but in the Phoronix average for instance the 9950X3D had a 25%+ uplift over the 7950X3D which is a bit more than I expected. If it seems like the CCD-related stuff, Vcache, drivers/firmware scheduler issues on Windows being fixed or improved (and Linux, but the latter was working from a better place), as well as the generational uplift and slight frequency bumps on the cache cores really add up. Normally I wouldnt rush out tomorrow or anything, but it seems a larger than standard improvement and the current environment makes me worry about tariff nonsense, scalping, and the like that jacks the price or makes it impossible to find at close to retail etc.. for awhile. Still not as likely to pull the trigger, but I'm far more tempted than I'd usually be in a situation where I knew that it would always be available on the shelf at the same or lower price.

This also makes me wonder what kind of benefit the Zen5 powered Threadripper will have...
 
Last edited:
  • Like
Reactions: erek
like this
Will wait to see what Zen6 brings.
Already have four maxed out AM5 systems with 7950Xs ($240 ea) running at -20 offset, and don't have to deal with the Game Bar nonsense.
 
  • Like
Reactions: erek
like this
At $699 on shelves, this is cheaper than the launch of the 5950X at $749 a few generations back.. nice, imo
 
I'll probably end up doing that but in the Phoronix average for instance the 9950X3D had a 25%+ uplift over the 7950X3D which is a bit more than I expected. If it seems like the CCD-related stuff, Vcache, drivers/firmware scheduler issues on Windows being fixed or improved (and Linux, but the latter was working from a better place), as well as the generational uplift and slight frequency bumps on the cache cores really add up. Normally I wouldnt rush out tomorrow or anything, but it seems a larger than standard improvement and the current environment makes me worry about tariff nonsense, scalping, and the like that jacks the price or makes it impossible to find at close to retail etc.. for awhile. Still not as likely to pull the trigger, but I'm far more tempted than I'd usually be in a situation where I knew that it would always be available on the shelf at the same or lower price.

This also makes me wonder what kind of benefit the Zen5 powered Threadripper will have...

I'm same boat, ordered one from Amazon. This is a larger uplift than I expected, in both gaming and productivity vs 7950x3d.
 
  • Like
Reactions: erek
like this
When you take the GPU out of the equation the 9800X3D still clearly comes out on top. The 300 MHz clock advantage after disabling the non-3D cache CCD can't make up for that.

1000000218.jpg
 
  • Like
Reactions: Niner
like this
Why not? Those reviews didn't disable CCD1 at the BIOS level and the scheduling is still imperfect (they are all using the stock setup and trusting AMD and Windows software), with processes randomly waking cores on the second CCD (JayZ shows that in one of his vids).

The CCD0 is a 9800X3D with higher max boost clock, better silicon, no reason why it shouldn't win.

It was the same story with the 7950X3D.

edit:
some Korean review here with non cache CCD disabled: https://quasarzone.com/bbs/qc_bench/views/94164
>> 9950X3D with non cache CCD disabled, ends up about 3% faster than 9800X3D at 1080p... lines up with the small clock bump
 
Last edited:
I'll keep my 14900KS at 6.2Ghz / DDR5 8200 Cas 38 no point in upgrading. I only use my PC to play games. At 1440p there isn't much of a difference in most games.

On another note I love how a lot of the reviewers are only showing 1080p benchmarks. This is bad. Newbies will see those benchmarks and think the 9950X3D is so much better than Intel. Then they go out and buy it and play at 4K when the difference between AMD and Intel at 4K is very very small. Intel overclocks so much better than AMD and if you are not overclocking you are not [H]ard. I am strictly ONLY talking about the people that only use their computers for gaming here.
 
Last edited:
That's not true there are plenty of 1440p and 4k benchmarks but they are just boring to look at since most games are GPU bound then. But there are plenty of games that manage to be CPU bound at 4k still, and this is more true than ever thanks to upscaling. Whether you play those games or not is another matter. Like:
1741959636243.png
1741959651228.png


And absolutely not everyone can run a 14900k@6.2ghz and DDR5@8200mhz (if that is even stable and not degrading).
 
I'll keep my 14900KS at 6.2Ghz / DDR5 8200 Cas 38 no point in upgrading. I only use my PC to play games. At 1440p there isn't much of a difference in most games.

On another note I love how a lot of the reviewers are only showing 1080p benchmarks. This is bad. Newbies will see those benchmarks and think the 9950X3D is so much better than Intel. Then they go out and buy it and play at 4K when the difference between AMD and Intel at 4K is very very small. Intel overclocks so much better than AMD and if you are not overclocking you are not [H]ard. I am strictly ONLY talking about the people that only use their computers for gaming here.
If games still allowed the resolution to drop to something like 640x480 we should be using that. GPU-bound scenarios tell us nothing about a CPU's performance in games. I criticize TPU for still using "high" settings in their 1280x720 test. If you want to pinch pennies instead of going for the best, then go ahead and look at the results for the resolution you game at. Clock speed isn't everything, and I thought we had moved beyond that mentality when first generation Core processors came out.
That's not true there are plenty of 1440p and 4k benchmarks but they are just boring to look at since most games are GPU bound then. But there are plenty of games that manage to be CPU bound at 4k still, and this is more true than ever thanks to upscaling. Whether you play those games or not is another matter. Like:
View attachment 716795View attachment 716796

And absolutely not everyone can run a 14900k@6.2ghz and DDR5@8200mhz (if that is even stable and not degrading).
While true, TPU tests scaling separately.
 
Now I have to decide if its worth it to step up from a 7950X3D. I lean towards no or at least not now, but I don't want to see a stock and scalping issue either. I want AMD to get their scheduling fixed on Windows and otherwise optimize stuff without precarious software defined fixes for the asymmetry (ie Xbox game bar dependency etc) and both the 7000 and 9000 X3D series should benefit from that. of course, the Linux side of things has their own scheduler that handles things differently and has evaded some of those issues it seems; I imagine it will be even easier now. Still, good to see this is a solid CPU, though I am a bit disappointed that the extra vcache was not on both CCDs.
It would have been an instabuy for me if it was dual 3dvcache CCDs so there wouldn't be all the scheduling issues, even if I had to pay a bit more. Maybe they will give us the option with zen6.
 
I’m tempted to pick one up, but most things I do that requires that many cores I do on my Mac now (lots of VSTs). I don’t really want to migrate everything back to windows.
 
If games still allowed the resolution to drop to something like 640x480 we should be using that. GPU-bound scenarios tell us nothing about a CPU's performance in games. I criticize TPU for still using "high" settings in their 1280x720 test. If you want to pinch pennies instead of going for the best, then go ahead and look at the results for the resolution you game at. Clock speed isn't everything, and I thought we had moved beyond that mentality when first generation Core processors came out.
video game setting can affect CPU demands level (LOD, crowds and world complexity can be affecter per settings), even some engine way to do raytracing can stress cpu performance more than without.

There is not really an easy way, single answer fit all to do it.

Newbies will see those benchmarks and think the 9950X3D is so much better than Intel. Then they go out and buy it and play at 4K when the difference between AMD and Intel at 4K is very very small
I think it is more an issue the other way around, we have still good amount of 4k benchmark around for new CPU on big publication like TPU, low and mid-range GPU tested on regular CPU not so much, making the upgrade gap for someone with an average CPU look more impressive that it will be.

Of course have people test with a 9800x3d even the 9600 vs 5060 vs 7600 vs 4060 vs 3060, but having someone test with a 12600k/3700x/5600x type of cpu, 7700x/14600k would be useful information for potential buyers, as if a buy a pre-built with that type of cpu how much more would I by willing to pay to go from the 4060 model to the 5060 or if I am upgrading my gpu in my reasonable desktop system.

Everyone is well aware of the 4k playing GPU limitation for some title (can even get exaggerated) and that rich enthusiast (either quite informed or quite rich), less so the low-mid range GPU benchmark all made with 9800x3d CPUs.
 
I'll keep my 14900KS at 6.2Ghz / DDR5 8200 Cas 38 no point in upgrading. I only use my PC to play games. At 1440p there isn't much of a difference in most games.

On another note I love how a lot of the reviewers are only showing 1080p benchmarks. This is bad. Newbies will see those benchmarks and think the 9950X3D is so much better than Intel. Then they go out and buy it and play at 4K when the difference between AMD and Intel at 4K is very very small. Intel overclocks so much better than AMD and if you are not overclocking you are not [H]ard. I am strictly ONLY talking about the people that only use their computers for gaming here.
With all due respect, what sort of copium is this? We’re glad you like your 14900KS, and we’re glad that most situations for you are not CPI limited. It is apparent the 9950x3D is not a good path for you. Everything else you said is literally opinion. Some people simply want the best, and in this case, the 9950x3D is it. It is extremely [H].
 
Back
Top