RTX 4090 & 53 Games: Ryzen 7 5800X vs Ryzen 7 5800X3D

[W1zzard]

How much performance can you gain with Ryzen 7 5800X3D over the non-3DV Cache Ryzen 7 5800X when using the mighty GeForce RTX 4090? Our review has the answer, we test 53 games at three resolutions.

Show full review

 

[Daven]

This means the small selection of games in the CPU review suite is not enough or the right type to show the performance differences. I believe the 5800X3D regressed with the choice of the 12 games in the latest reviews.

 

[seventy]

Dual rank sams b-die on 5800x3d vs "questionable" ram on 5800x.
Would really like to see the opposite, ie 5800x with good b-die vs 5800x3d with trash ram.

@W1zzard if possible, can you also add 0.1% lows to the graph?

 

[clopezi]

This means the small selection of games in the CPU review suite is not enough or the right type to show the performance differences. I believe the 5800X3D regressed with the choice of the 12 games in the latest reviews.

I think it's more than enough for 99% of the people. With those games, you can do a nice picture of the CPU performance.

If you need 100 games with pixel perfect versus to test if one or another CPU it's better for gaming, we can get two conclusions. First, both CPU's are great. Two, you are demanding a very specific scenario and it's impossible to media outlets to test 100 games in every single review...

 

[W1zzard]

and it's impossible to media outlets to test 100 games in every single review

I've been having wet dreams about a 100 game benchmark, but kinda difficult to find that many games (that are relevant and not impossible to test due to always-online, etc)

if possible, can you also add 0.1% lows to the graph?

I've been thinking about ways how to add some kind of min fps indication, but it's not easy. Definitely not gonna happen for this review or the upcoming 13900K

 

[Toss]

That's why I underclocked/volted my CPU to 4 GHZ while gaming at 4k60. It just can't be bottlenecked with modern GPU's so why bother?

 

[Daven]

I think it's more than enough for 99% of the people. With those games, you can do a nice picture of the CPU performance.

If you need 100 games with pixel perfect versus to test if one or another CPU it's better for gaming, we can get two conclusions. First, both CPU's are great. Two, you are demanding a very specific scenario and it's impossible to media outlets to test 100 games in every single review...

Its a pretty big difference. In the 53 games tested 12900k review, intel is 15.7% faster than the 5800x at 1080p. In this review, the 5800x3d is 18.5% faster. But in the latest 13700k review, the 12900k summary shows a 8.3% lead at 1080p versus the 5800x3d. That’s quite a lot when it should be a few percent deficit.

But you are right, all CPUs do gaming well from the last few years and it is hard to pick one over the other as tech reviewers cannot bench all scenarios. Either way it looks like the 13900k will have the same 1080p gaming performance as the 5800x3d. I’m guessing that’s the last 53 game test before TPU makes a decision on a new test bench.

 

[seventy]

I've been thinking about ways how to add some kind of min fps indication, but it's not easy.

Can I ask why? You should have the data already, but now you just throwing it away.
Even identical graph but with 0.1% fps lows instead of average fps would be better than nothing.
People who are interested in that kind of stuff can easily figure it out, just give a way to access it please.

 

[SL2]

@W1zzard if possible, can you also add 0.1% lows to the graph?

Given how many games there are, it really needs a separate graph.

 

[Xuper]

So far only this games are lower than 5% at 1080p :

AC Valhalla
Control
Devil May cry 5
GTA V
Rage 2
Strange Brigade
Witcher 3

about Witcher3 and GTA V, I'm quite surprised !

I checked this :

RTX 4090 & 53 Games: Ryzen 7 5800X vs Core i9-12900K Review

We test the NVIDIA GeForce RTX 4090 with 53 games at three resolutions, comparing the AMD Ryzen 7 5800X against the Intel Core i9-12900K. The idea here is to get a feel for how much graphics performance is lost by a weaker processor.

www.techpowerup.com

About Devil May Cry5 , What the hell ? 5800X is 59.7% Faster than 12900K?

 

[GreiverBlade]

mmmh, now that i can find a 5800X for 299.90chf (about same price of a 5700X ) and a 5800X3D is 409.90chf (hey! i said if it dropped sub 450 when it was 449 i would place an order asap )
given that i play in 1440p/1620p the 5800X3D is highly tempting and would still cost me way less than a full new Intel or AMD build
but dang ... 300chf for a 5800X, i could settle for that one and add 50chf for 16gb RAM more (same stick as i have ofc ) and go 32gb

 

[rv8000]

Dual rank sams b-die on 5800x3d vs "questionable" ram on 5800x.
Would really like to see the opposite, ie 5800x with good b-die vs 5800x3d with trash ram.

@W1zzard if possible, can you also add 0.1% lows to the graph?

This is sort of a good point, but neither should have a different kit in a good testing environment.

(Seems like an oversight in the review) Nevermind, looks like a poor decision as it’s pointed out in the test rig descriptions; from personal testing higher DDR frequencies and IF clocks have dramatic impacts on minimum frame rates (and thus average). Both CPUs should have been running at the lower maximum achievable IF clock between the two CPUs.

 

[LifeOnMars]

That's why I underclocked/volted my CPU to 4 GHZ while gaming at 4k60. It just can't be bottlenecked with modern GPU's so why bother?

I run my 5700x on eco mode @4k with 32Gb Slow DDR4 and my 3080 12Gb No issues and a nice cool and quiet gaming experience.

 

[cowcatwithahat]

It is scary how fast the technology moves now vs the time we were stuck with the 4 cores for so damn long.
As a 5800X user, it is hard to justify spending money again on a processor when 90% of it`s use is gaming & youtube;
I get it, it`s fast! But in most cases, at least for me, more then a steady 120 FPS at any resolution it is placebo.
I am amazed of the content you guys post. Thank you!

 

[Selaya]

this is kinda cool (story bro) ig but ... we all know 5800x3d is better at gaming. (ig it's telling us which games have cache scaling?)
5800x3d vs 13900k/13700k would've been far more interesting, tbh

EDIT: it's already on the menu i see (read first spam later ig)
however this still feels kinda ... superfluous? could be good for perspective later, ig

 

[rusTORK]

Didn't read it yet, but... FINALLY! W1zzard is true magician!

P.S. I got R7 5800X and thought about switch to R7 5800X3D.
P.P.S. Oh... and Dark Hero too... %)

 

[dirtyferret]

I've been having wet dreams about a 100 game benchmark, but kinda difficult to find that many games (that are relevant and not impossible to test due to always-online, etc)

how about a 101 game benchmark on a tuseday under a waning crescent moon while eating a ben & jerry half baked pint of ice cream? Otherwise, what's the point none of the data would be valid other any other scenario.

 

[BlueKnight]

10 days ago I sell my 5900X at 330 Euro to buy a 5800X3D for 405 Euro. I'm very satisfied for my choice!

 

[ARF]

It is scary how fast the technology moves now vs the time we were stuck with the 4 cores for so damn long.
As a 5800X user, it is hard to justify spending money again on a processor when 90% of it`s use is gaming & youtube;
I get it, it`s fast! But in most cases, at least for me, more then a steady 120 FPS at any resolution it is placebo.
I am amazed of the content you guys post. Thank you!

This is thanks to AMD achieving some type of a miracle with the design of a so potent Zen architecture. It's really fast and competitive, not like the previous AM3 and AM2 designs.

But speaking of stagnation - you can't imagine how bad it is to enter one of those big techno stores and look at large screens with 1080p - so ugly - bad colours, huge pixels and the screen-door effect everywhere.
Will we be stuck with 1080p forever?

 

[Gucky]

I bought a 5800X3D last week for 349€...my current CPU is slower as the 5800X in the review, but goes for 450€ on Ebay.
I already have an AM4 Board with 3200CL14 RAM here.

 

[Shazamy]

Thanks so much for another great comparison W1zzard! Always appreciated

 

[QuietBob]

Big thanks to @W1zzard for another comprehensive article! I can only imagine the amount of time that went into testing such a huge number of games

Since the previous article puts the same 5800X against the 12900K, we can draw some more comparisons by looking at the graph data for both:

In terms of average frame rate at 1080p, the 12900K is meaningfully faster than the 5800X3D (>10%) in only 2 out of the 53 games tested. In Civilization VI it posts 12% higher fps, and in Cyberpunk 2077 - 19% higher. Both titles are known to scale beyond 16 threads, so Intel's advantage here is most likely due to having 50% more threads than AMD.

On the other hand, the 5800X3D is outstripping the 12900K by a similar margin in 5 titles. Borderlands 3, Darksiders 3 and Dota 2 show 15%, 14%, and 16% higher scores, respectively. Prey is 20% faster, and Devil May Cry 5 -- while being a clear outlier -- 61% faster in terms of average fps.

It's worth noting that in 4K the tables are turned and the 5800X3D leads the 12900K by 11% in Civilization VI and equalizes in Cyberpunk 2077. AMD is still faster in 4K in Borderlands 3 by 14% and in Dota 2 by 15%. Intel catches up in Darksiders 3 and Devil May Cry 5, and ekes out a 5% advantage in Prey.

 

[zlobby]

Mmm, pants-wetting! Imagine the latest X3D Zen!

 

[ffolekram]

Due to the fact that 5800x3d is the only AM4 cpu with 3D V-cache, almost all AM4 mobos (B350, X370, B450, X470, B550, X570) can run 5800x3d with no problem, and very small performance difference because of the 3D V-cache between 3200cl16 and 3800cl14,
this would make the 5800x3D a very popular option for any existing AM4 users to consider upgrade to, when it's the time.
I currently have a 5600x on a x570 taichi with a 3080 12g at 1440p, personally I would definitely consider 5800x3D instead of AM5 cpus in the next few years when comes to upgrade (since I can reuse almost all of the existing parts in my system).

 

[blazeddd]

Any chance we can get FPS numbers instead of percentages?