Hi all, finally took the plunge and switched from a 13600K with DDR4 to a 9800X3D and DDR5. GPU is still the same 4080 and running 4K. What to say? Performance increase is disappointing, not to say a total waste of money. Glad MSFS2024 is not the only game I play... 

BUT: maybe some remnants are hindering the X3D to show its full potential? I deleted shader caches, rolling cache and scenery indexes. Something else to rebuild/delete that might hamper X3D performance? 

Dont get me wrong, outside big hubs with AI traffic and ini A350, performance is perfect (was already with the 13600K though...), sim seems notably smoother. But e.g. on MK Studios LSZH with AIG traffic and ini A350, I do not see any performance gain, max. 2-3FPS (60-65FPS with FG on). And as the dev mode overlay showed «limited by main thread» all the time over there with the 13600K, I expected more tbh...

Keep in mind that these CPU differences are only noticeable in lower resolutions. My dad went from a 7800x3D to a 9800x3D in 4k and the difference is about... zero. I'm the lucky one because at 1440p and going from a 7600x to his old 7800x3D for a fraction of the price did increase performance for me (and even then not that much).

The 13600K is a great CPU, i had one and only changed to AMD because the thing is an oven and draws way too much power for what it delivers.

55 minutes ago, AnkH said:

switched from a 13600K with DDR4 to a 9800X3D and DDR5

https://cpu.userbenchmark.com/Compare/Intel-Core-i5-13600K-vs-AMD-Ryzen-7-9800X3D/4134vsm2356328

"the 13600K and 14600K still deliver almost unparalleled real-world gaming performance for around $200 USD. Spending more on a gaming CPU is often pointless, as games are normally limited by the GPU."

 

at 4K, going from a 4080 to a 4090 would make a bigger difference

Edited May 29, 20251 yr by turbomax

1 hour ago, AnkH said:

Hi all, finally took the plunge and switched from a 13600K with DDR4 to a 9800X3D and DDR5. GPU is still the same 4080 and running 4K. What to say? Performance increase is disappointing, not to say a total waste of money. Glad MSFS2024 is not the only game I play... 

BUT: maybe some remnants are hindering the X3D to show its full potential? I deleted shader caches, rolling cache and scenery indexes. Something else to rebuild/delete that might hamper X3D performance? 

Dont get me wrong, outside big hubs with AI traffic and ini A350, performance is perfect (was already with the 13600K though...), sim seems notably smoother. But e.g. on MK Studios LSZH with AIG traffic and ini A350, I do not see any performance gain, max. 2-3FPS (60-65FPS with FG on). And as the dev mode overlay showed «limited by main thread» all the time over there with the 13600K, I expected more tbh...

If performance was already perfect, why did you feel the need to upgrade??

11 minutes ago, Christopher Low said:

If performance was already perfect, why did you feel the need to upgrade??

" Intel's marketers remain asleep (terminally?) at the wheel. Nevertheless, the 13600K and 14600K still deliver almost unparalleled real-world gaming performance for around $200 USD. Spending more on a gaming CPU is often pointless, as games are normally limited by the GPU. Without significant improvements in social media marketing: forums, reddit, youtube etc., Intel now face the very real risk of bankruptcy (third worst-performing S&P500 stock from Jan to Aug 2024). Since this summary was published just two days ago, hundreds of twitter threads, thousands of “pcmasterrace” reddit posts, multiple magazine articles, and several youtube videos have emerged in unanimous support for the $480 USD 9800X3D. All of these supposedly disinterested actors are working the weekend to convince you to pay their favourite billion-dollar brand an extra $280 USD this holiday season. ^{[Nov '24 CPUPro]"}

 

^{https://cpu.userbenchmark.com/Compare/Intel-Core-i5-13600K-vs-AMD-Ryzen-7-9800X3D/4134vsm2356328}

Edited May 29, 20251 yr by turbomax

Userbenchmark call AMD 'Advanced Marketing Devices' and claim their users are an army of Neanderthals, so I wouldn't take them seriously.

 

I am not interested in a single userbenchmark comment. but in 27.000 vs. 2.000 uploaded test results, so I would take only their results seriously.

Edited May 29, 20251 yr by turbomax

Update was done primarly because I like working on my computerhardware from time to time, because I can sell the old components to a friend of mine and because I play other games where I expect more benefit from the X3D. 

But my question remains: what else is to be deleted to start all over again with MSFS2024? Is in the usercfg.opt something that might influence performance after switching the CPU?

And yes, a 4090 would be great because of the VRAM issue, but I am still primarly CPU limited, that is why I expected more from the X3D...

What's the refresh rate of your monitor?  

What kind of frame rate do you get with both FG and VSync turned off?

Something I noticed when running MSFS 2020 on my 13900KS with x2 FG on (into a 4K 60Hz TV), is with FG on, the video frame rate (the rate at which frames are sent from the GPU to the display) synced up to the display's refresh rate at 60 fps and MSFS showed CPU limited with it's frame rate capped at 30 fps.  The MSFS frame rate was considerably higher with VSync and FG off, but not as smooth, and with screen tearing.

When I cap frame rate using RTSS, it will also show CPU-limited...so I'm wondering if the "CPU-limited" you're seeing is a red herring, and you're actually capping frame rates at 1/2 the refresh rate due to FG hitting the display's max refresh rate.

 

Good point, I indeed limit the FPS to 36 via usercfg.opt, might be the reason for my readout. Thanks for pointing that out. But then, I do not even reach those 36FPS on my heavy load test scenario, so at least in this, the limit is real...

2 hours ago, AnkH said:

Good point, I indeed limit the FPS to 36 via usercfg.opt, might be the reason for my readout. Thanks for pointing that out. But then, I do not even reach those 36FPS on my heavy load test scenario, so at least in this, the limit is real...

What is your Cinebench r23 score?

No idea, but 3DMark Timespy CPU score is roughly 16500, which is perfectly normal for a 9800X3D...

A few days late to this party, but I have to ask if you kept your storage (i.e. did not format / install OS), or are you running off the same OS install that you had with the 13600K?

Something "in between". I did not reinstall W11 and simply reactivated it again using the key. Then I forced a W11 24H2 full update using the Windows installation tool, took 2.5h and is basically a new OS installation. 

BTW: in the mean time I did some further testing, the rig is ok, the 9800X3D performs as it should and I guess I simply had too high expectations from this upgrade. But with the exception of inibuilds A350 on MK Studios LSZH with AIG traffic, the performance I get is fine now, I guess I also did not properly test the same settings because I rebuilt the usercfg.opt after switching the hardware, resulting in a full reset of my settings. 

On 5/29/2025 at 11:32 PM, Bob Scott said:

What's the refresh rate of your monitor?  

What kind of frame rate do you get with both FG and VSync turned off?

Something I noticed when running MSFS 2020 on my 13900KS with x2 FG on (into a 4K 60Hz TV), is with FG on, the video frame rate (the rate at which frames are sent from the GPU to the display) synced up to the display's refresh rate at 60 fps and MSFS showed CPU limited with it's frame rate capped at 30 fps.  The MSFS frame rate was considerably higher with VSync and FG off, but not as smooth, and with screen tearing.

When I cap frame rate using RTSS, it will also show CPU-limited...so I'm wondering if the "CPU-limited" you're seeing is a red herring, and you're actually capping frame rates at 1/2 the refresh rate due to FG hitting the display's max refresh rate.

 

This has been my observation as well. I too was a little underwhelmed by my upgrade fron a 9900K to a 9800X3D I expected more considering the silicon quality indicated by the bios.  However when I consider the minor performance uplifts we have got with each CPU generation over the past 6 years I'm probably getting about getting what would be expected and any deficit might well be due to taking the short cut method and not doing a full reinstall of the OS and the sim.

Bruce