Ryzen 9 5900x vs Ryzen 7 5800x3d

[shastasm]

Hey all.

I've been considering getting a Ryzen 7 5800x3d for DCS & MSFS2020.  I am currently running a Ryzen 9 5900x and getting good performance, but would like to understand if there is a big performance gain for DCS if I switch to the R7 5800x3d.  

Appreciate any feedback from anyone that has made this switch.  

[tomcat_driver]

More performance, yes, but worth the price, not in my opinion. Is your PC only for gaming?

Edited October 21, 2022 by tomcat_driver

[Hiob]

What resolution and GPU are you running? Are you going to use VR?

In pancake, single-player and resolutions above 1080p the answer is a definite "no!"

I benched those exact cpus against each other and the only meaningful advantage of the 3D above the 5900X was in VR.

If you don't feel any pain regarding the performance now, I would save the money for a more meaningful upgrade down the road.....

Edit: When you are playing in 4K - don't even think about it.

Edit2: When the GPU in your signature is still up-to-date, you should upgrade that first!

Edit3: The longer I look at it.... concluding from your signature you upgraded from a 3900X to a 5900X while running a 2070 as a GPU  - just from a gaming point of view that was already a sub optimal route, but ok-ish. Rinsing that money down the drain now with side/up-grading to a 5800X3D would be insane. Add a couple of bucks and get a (used) 3080. That will push your fps in a much more meaningful way. 

Edited October 21, 2022 by Hiob

[VirusAM]

5 hours ago, Hiob said:

What resolution and GPU are you running? Are you going to use VR?

In pancake, single-player and resolutions above 1080p the answer is a definite "no!"

I benched those exact cpus against each other and the only meaningful advantage of the 3D above the 5900X was in VR.

If you don't feel any pain regarding the performance now, I would save the money for a more meaningful upgrade down the road.....

Edit: When you are playing in 4K - don't even think about it.

Edit2: When the GPU in your signature is still up-to-date, you should upgrade that first!

Edit3: The longer I look at it.... concluding from your signature you upgraded from a 3900X to a 5900X while running a 2070 as a GPU  - just from a gaming point of view that was already a sub optimal route, but ok-ish. Rinsing that money down the drain now with side/up-grading to a 5800X3D would be insane. Add a couple of bucks and get a (used) 3080. That will push your fps in a much more meaningful way. 

 

I have the same CPU (5900x), actually I am not using VR, but I am considering going back to it (will try to get an used reverb g2).

Right now i have a 2080ti, but this evening I will switch it to an Evga 3080ti FTW3 (used from amazon warehouse, so I can send it back). 

Undecided if the swap is needed (5900x -> 5800x3d and 2080ti --> 3080ti), or if I should save money and wait for next year.

On the Flight Sim official forums I have seen reports from users which used it at resolutions higher then 1080p and still having benefits (the one I recall switched from a 5800x and not a 5900x anyway).

Edited October 21, 2022 by VirusAM

[Hiob]

14 minutes ago, VirusAM said:

I have the same CPU (5900x), actually I am not using VR, but I am considering going back to it (will try to get an used reverb g2).

Right now i have a 2080ti, but this evening I will switch it to an Evga 3080ti FTW3 (used from amazon warehouse, so I can send it back). 

Undecided if the swap is needed (5900x -> 5800x3d and 2080ti --> 3080ti), or if I should save money and wait for next year.

On the Flight Sim official forums I have seen reports from users which used it at resolutions higher then 1080p and still having benefits (the one I recall switched from a 5800x and not a 5900x anyway).

 

I did the test only in DCS (with a 3080). A Benchmark in DCS (pancake, single player, high settings) in 1440p and 4K showed no fps gains whatsoever.
However, CPU frametimes in VR (steam VR, Index) decreased by 20-25%.

[VirusAM]

I did the test only in DCS (with a 3080). A Benchmark in DCS (pancake, single player, high settings) in 1440p and 4K showed no fps gains whatsoever.
However, CPU frametimes in VR (steam VR, Index) decreased by 20-25%.

Thanks, for the informations...
Very useful, and I think it is the only upgrade path, as it doesn't involve changing anything else.

[shastasm]

Great comments and feedback everyone.  Thank you and appreciate it.  Currently running a rtx6900xt at 1440p with my 5900x so I'm good with existing performance. I was exploring if the x3d was a big changer or not.  My rig is mostly for general use and flight sims; exploring video creation with some of my liberation campaigns too.   No VR either.  I tried it once with my 2070  and hp reverb 2, but didn't care for the quality.  TrackIR for me   

Liberation Campaign can be a resource hog too when playing the full maps and drops FPS tremendously if not culled and ran on a dedicated server.  This is where I do most of my flying until EG comes out with their dynamic campaign.  Hence this is where I thought maybe the L3 cache would help.  

[Beirut]

Not to be a stick in the mud, but the comparison videos I watch show just incremental increases when upping the CPU. For instance, in the civvy sim, my 3700X has a 1% low of 72 and the 5800X3D is at 78 at 1080p. That's basically nothing for the outlay of a good chunk of change. Even against a 5950X, the 3700X looks way to close to justify the expense of the upgrade. 

 

So where is the actual CPU sweet spot in DCS? Isn't 90% of the beef in the GPU?

 

 

[Mars Exulte]

The benefit is in specific use cases, and 1080p ain't one of em.

VR (ie ultra high resolutions), with a 3080ti+ that can make use of the extra bandwidth. That's where you see a difference. Not in 1080p pancake. If you're playing 1080p you don't need a 5x00 anything, or anything later than a 1000 series GPU. So, it doesn't apply.

 

The sweetspot, btw, is when your CPU and GPU frametimes are roughly equal, and obviously as low as possible.

Edited October 22, 2022 by Mars Exulte

[some1]

5 hours ago, Beirut said:

the comparison videos I watch

Comparison videos from small channels that look like this one you've posted, are usually fake. 

[Beirut]

4 hours ago, some1 said:

Comparison videos from small channels that look like this one you've posted, are usually fake. 

 

Hmmmm, that's a new one on me. I admit to being surprised. 

 

Just watched this. 

 

 

[some1]

Unfortunately that's the way it is. For example, the guy you've linked previously uploaded an RTX4090 "test" on 11th Oct. That's the day when press embargo has been lifted, but the cards were not yet available to buy, and more importantly, the drivers which worked with 4090 were not yet available for download. Only the selected reviewers received them earlier.

What are the chances nvidia would send a kit to this random guy? Not very high. Yes, they give cards to smaller reviewers and various influencers as part of marketing campaign, but those people are usually expected to show the card on a video, talk about it, etc. You know, do some marketing work or at least make an actual review and not just drop a bunch of numbers on the screen in total silence. 

[Beirut]

No doubt you're right. I suppose I never saw the reasoning to faking results, but if it gets you views and views mean money, there it is. 

 

Good to know. 

[Mars Exulte]

6 hours ago, Beirut said:

if it gets you views and views mean money

  Unfortunately, this drives 99% of online discourse these days. Ever wonder why people like/karma farm on websites with bots? So they can turn around and sell that bot with established ''credibility'' and use it to push marketing.

Lovely dystopia we have here.

[Beirut]

That's it. I'm quitting my job and starting a YouTube channel. 

[Hiob]

1 hour ago, Beirut said:

That's it. I'm quitting my job and starting a YouTube channel. 

Wait with the quitting until you have 100.000 subs…..

[Mars Exulte]

3 hours ago, Hiob said:

Wait with the quitting until you have 100.000 subs…..

Even better if you film from a hot tub and bottle the water afterward. That's where the real money is.

[BitMaster]

Am 23.10.2022 um 02:40 schrieb Mars Exulte:

Even better if you film from a hot tub and bottle the water afterward. That's where the real money is.

[xoxen]

On 10/21/2022 at 9:25 AM, Hiob said:

What resolution and GPU are you running? Are you going to use VR?

In pancake, single-player and resolutions above 1080p the answer is a definite "no!".......

 

Many thanks Hiob,

I had the idea of switching to the 3D from my 5800X. But I´m running 3 displays at a resolution of 3x 2560x1440. Based on your experience I will not go into that direction. At that resolution I already believed that I´m GPU bound. So many thanks on saving me a lot of money and disappointment. I owe you at least one beer. Whenever you are in my area (CLP), let me know 

[Eaglewings]

In the same vein, I want to throw in here my cpu upgrade plans here too and get your opinions if possible.

Upgraded my gpu from a gtx 1060 6g to rtx 3070.
I am still using my i5 7600k cpu.
Upgrading to the amd 7000 series cpu is high in terms cost.

Considering getting the Ryzen 5600 without the x as I have seen there is no huge difference between the x and non x cpu.
I really like the amd long platform upgrade feature so I plan to get into that when the cost of upgrading to that platform becomes reasonable.

Would the Ryzen 5600 be a big upgrade now?
5800x 3d with motherboard is still high for me now and beside that is the highest upgrade I can go on the am4 platform.
I play in vr and I am expecting my 1440p monitor soon.

Your thoughts please.

[Hiob]

51 minutes ago, xoxen said:

Many thanks Hiob,

I had the idea of switching to the 3D from my 5800X. But I´m running 3 displays at a resolution of 3x 2560x1440. Based on your experience I will not go into that direction. At that resolution I already believed that I´m GPU bound. So many thanks on saving me a lot of money and disappointment. I owe you at least one beer. Whenever you are in my area (CLP), let me know 

Haha, no problem, you're welcome.

You are almost certainly GPU limited most of the time. Mind, that there are certainly specific scenarios in which the 3D would shine. The question is, how much and how often they apply.
I would certainly save the money for a more meaningful upgrade down the road.

Let's see what todays update has to offer for improvements.....

[DmitriKozlowsky]

On 10/20/2022 at 3:53 PM, shastasm said:

Hey all.

I've been considering getting a Ryzen 7 5800x3d for DCS & MSFS2020.  I am currently running a Ryzen 9 5900x and getting good performance, but would like to understand if there is a big performance gain for DCS if I switch to the R7 5800x3d.  

Appreciate any feedback from anyone that has made this switch.  

I have the same dilemma. So I went to unbiased objective source. YouTube!  With same memory at 3600, same GPU RTX 3080, same NvME+SATA3 SSD, same settings (non DCS) in MS2022, and a bunch 1st/3rd person shooters. FPS were nearly the same. Some games higher, some lower, and Blender Cycles+Vray close enough that both are statistically nearly same. I'd say what you got , 5900X, is excellent. IMHO, you may not see significant increase or degradation in FPS with R7 5800X3D. However mileage may vary. If your budget allows, upgrade GPU. But what you have now is already fantastic. 

As appalling as it may be, R9 7900X with OC DDR5 6000 ram, PCI Gen4 NvMe, & 4090. Is maybe 15% higher then 5900X with 3600 DDR4. Either its the code efficiency, or Nvidia drivers, or whatever else, there is little to none justification to jump to AM5 DDR5 RTX40XX platform. Given the expense. But my opinion is derived from YT posted duels. I downselected only 2560X1440 duels. At 3840X2160 is where AM5 and 4090 pulls ahead.

[shastasm]

On 10/26/2022 at 5:21 AM, DmitriKozlowsky said:

I have the same dilemma. So I went to unbiased objective source. YouTube!  With same memory at 3600, same GPU RTX 3080, same NvME+SATA3 SSD, same settings (non DCS) in MS2022, and a bunch 1st/3rd person shooters. FPS were nearly the same. Some games higher, some lower, and Blender Cycles+Vray close enough that both are statistically nearly same. I'd say what you got , 5900X, is excellent. IMHO, you may not see significant increase or degradation in FPS with R7 5800X3D. However mileage may vary. If your budget allows, upgrade GPU. But what you have now is already fantastic. 

As appalling as it may be, R9 7900X with OC DDR5 6000 ram, PCI Gen4 NvMe, & 4090. Is maybe 15% higher then 5900X with 3600 DDR4. Either its the code efficiency, or Nvidia drivers, or whatever else, there is little to none justification to jump to AM5 DDR5 RTX40XX platform. Given the expense. But my opinion is derived from YT posted duels. I downselected only 2560X1440 duels. At 3840X2160 is where AM5 and 4090 pulls ahead.

Appreciate the feedback.  I think I may skip a generation or two before my next updgrade.  Looking forward to DCS 2.8!

[DmitriKozlowsky]

51 minutes ago, shastasm said:

Appreciate the feedback.  I think I may skip a generation or two before my next updgrade.  Looking forward to DCS 2.8!

Once 2.8 ships, the whole performance testing begins all over again. As all these cool 2.8 stuff, dynamic weather, rolling clouds, will suckup CPU+memory and GPU power I suppouse. I havent seen it, there are some early insider build posts on YT  So I have no idea how much more horsepower 2.8 will need over 2.7 to maintain desired FPS. My requirement is 144+ on panels, and at least 180FPS with VR set, not using any anti-aliasing at 2560X1440. Now my current rig still delivers above 120 most of the time,  but its struggling, with F-16 and AH-64D performing poorly, and bogs down in combat when I strike convoys. F/A-18C siting cold on supercarrier at 1440P no anti-aliasing, is 70-90 FPS. Far too low for me. FPS increases after T/O and altitude gain. I loose on average 15-20 FPS if I turn on mirrors in cockpit view. Thats on current 2.7. I am going to loose FPS with 2.8 so it may time to budget a new system. Current one will not benefit from higher GPU, and I doubt that my PS can handle more then it has to right now. I replaced RTX 2060Super with RTX 3060Ti, and saw little if any FPS gain. I am really disappointed with that letdown.

My rig: XEON E5-1630v4 3.7-4.0(turbo), 64gb ECC 2666, HP Turbodrive NvMe, RTX 3060Ti. DCS is 1 screen 2560X1440.

[pegasus1]

I know not 100% appales for apples but i went from a 3800x to a 5800x3d gaming at 4k.

The 5800x3d is a huge jump in performance when a lot is going on but ultimately the GPU still bears the brunt of the work at max eye candy and 4K.

If you do go for a 5800x3d then make sure you undervolt it, ive shaved off 50mv and load temps have dropped significantly, boosted my CB23 score and also enabled faster and more sustained max single and all core clocks.