Have a 9950x3d / 5090 rig and love itđ
Very smooth
2 Likes
For pure gaming, Ryzen 7 9800X3D is the king of gaming - its 3D V-Cache will deliver equal or even higher FPS with an RTX 5090 and it offers a superior price-to-performance ratio.
If you also need heavy multi-threaded performance for professional applications, rendering or other compute-intensive tasks, then 9950X3D would be the better choice.
So, unless you plan to do heavy multitasking during gameplay -such as rendering, encoding, or running CPU-intensive background tasks - you donât really need 9950X3D. The extra cores of the 9950X3D wonât provide a noticeable benefit for gaming.
In March 2025, I wrote:
I have Ryzen 7 9800X3D and Iâm not even considering switching to the 9950X3D. The 9800X3D is so good that Iâm waiting for its true successor, which the 9950X3D definitely isnât.
Check out AMD Ryzen 9 9950X3D CPU Review by Gamers Nexus.
https://gamersnexus.net/cpus/amd-ryzen-9-9950x3d-cpu-review-benchmarks-vs-9800x3d-285k-9950x-more
Iâm glad to see that Gamers Nexus shares my opinion - it reassures me that sticking with the 9800X3D was the right choice.
2 Likes
Canât argue with your assessment of the 9800x3d as a gaming cpu. Heck I love my 5800x3d.
I splurged on the 9950x3d but the additional cost over the 9800x3d pales when you consider what I paid for my RTX 5090đ
I just upgraded my 7900x3d to a 9950x3d a couple of weeks ago and have been tweaking & testing. I wanted to know if I need to further tweak the 9950x3d cores/CCDâs (i.e. Process Lasso) at 4k resolution to get optimal performance with the sim. I set up a short test with SU2 and the FA/18 Hornet sitting on the runway at Kennedy, taking off straight out leveling off at 4000ft, cut throttle to 50%, fly over Manhattan, bank left and land at Newark. The test encompassed everything from sitting on the runway at Kennedy to coming to a full stop on the runway at Newark.
I first used the Win 11 at 4k using the Balanced Power Plan to run the sim exclusively off the v-cache cores with the other cores parked:
I then merely changed to the Win 11 Ultimate Power Plan which loaded up the sim to run on the v-cache cores but did not park any cores and allowed the sim to use all 16 cores/32 threads (which it did):
So as you can see, both charts show great results and did not uncover any issues with the two CCDâs. The charts for the frame times also mirrored each other with great results. I repeated the tests with SU3 Beta v1.5.3.0 with the same results so I am concluding that no cpu CCD tweaking is necessary on my 9950x3d with MSFS 2024 (MSFS 2020 may be another story). Process Lasso may squeak out a few more fps but with the smooth fps already being achieved I donât think it would be worth it. You can click on my gamer pic for my PC specs.
2 Likes
Could you do the same test though disabling in BIOS the CCD that not has the X3D memory?
Wow amazing tests thank you so much. Yes they are promising than there is no worry to go with 9950x3d cores not parking in msfs24. I want to be able to have fast video renders but aslo play msfs in the same time.
at the moment my 13700K is not good enough for my 4080 and system show cpu bottleneck. It is great on anything else but with msfs and it is add ons I can not reach 30*40 fps average. with my system I hope I will 
Do you think there would be any benefit to running other apps like LittleNavMap, Volanta, BeyondATC, etc. etc. on the non-vcache CCD though? i.e. keeping them OFF the vcache CCD.
In other words, would Win11 Balance power plan do that automatically?
I think Windows Game Mode with the Balanced Power Plan prioritizes that, but doesnât guarantee that scheduling the way Process Lasso does with CPU Affinities and CPU Sets, which are hard locks in a way.
1 Like
So I was reading here today as always, and took a look myself at the MS powerplan I was using. Didnât know about the Ultimate power plan. I been using High Performance with Game Mode on. I switched to the Ultimate Performance setting. Iâve noticed a difference myself. Iâm running some CFX data now, to compare to your chart, but the .01% lows have went down, and Vram usage as well. Holding pretty steady @60fps. No frame gen. Iâm a PC kindergartner, but learning, 
(Thank goodness for these forums)but there is a difference on my setup too.78003d 4070 super 32gb ram@4k 2d. Great day overall with the sim for me. Iâll post back for comparison. Glad I read your post WingwarperâŠKudos 
2 Likes
Reran the tests adjusting my resolution from 1280x1024 to 1080p - see below
3 Likes
Nice testing, thanks for doing it.
Am I misinterpreting the data, arenât your tests showing that Ultimate power plan gives better 0.1% lows? Or is this close enough to be almost imperceptible. Because thatâs all that matters I suppose â weâre not normally staring at an fps counter, we just want it to feel smooth.
Yes, you are technically correct and you are also correct in that this is pretty much imperceptible. Note also that this small advantage does evaporate at 4k.
I also want to mention that I run my PC at all times under the ultimate power plan in 4k. Cinebench and other benchmarking software will run a tad faster under the ultimate plan as the cores are kept primed and at the ready at all times (and never parked). It does take a very small amount of time to unpark a core and that can make a small difference when youâre running at full throttle.
Youâll notice the extra zip as @TogetherSpark69 mentioned above as in addition my cores will boost more aggressively .5 - .75 GHz higher under the ultimate plan.
Edit - from CapFrameX sensor data you can see under the ultimate power plan the cpu runs at a higher frequency
Ultimate power plan:
Balanced power plan:
1 Like
Good stuff!
1 Like
Thanks for the heads up TenâŠIâll be more carefulâŠFixed⊠
1 Like
With all due respect, Iâm not sure if you noticed, but this thread is about the Ryzen 9800X3D and 9950X3D, not the 7800X3D.
That said, itâs great to hear MSFS is running well for you after the changes - if it feels better on your setup, thatâs what matters most!
1 Like
Out of curiosity, what settings were you using during the run? For example - LOD, resolution, DLSS or TAA, FSR? FG? etc.
Would be great to get a full picture for comparison.
Not long ago, during some of my own testing, I was wondering why I was suddenly getting lower max FPS - your test reminded me exactly why 
Thanks for that!
BTW, when you switching back and forth between the balanced and ultimate power plans, did you leave them at their default settings, or did you tweak anything within the plans themselves?
Iâd like to know this as well, so I can run the same test with the same conditions.
In addition to TenPâs requests: Weather setting? Traffic setting? PG on or off?
Reran my tests with the 9950x3d this time making sure my resolution was at 1080p .
To recap, I just upgraded my 7900x3d to a 9950x3d a couple of weeks ago and have been tweaking & testing. I wanted to know if I need to further tweak the 9950x3d cores/CCDâs (i.e. Process Lasso) to get optimal performance with the sim. I previously ran the test in 4k (see above) however @BegottenPoet228 recommended I run the test in 1080p as this resolution puts more stress on the CPU.
Test Flight Set Up in 1080p:
SU2 and spawn on the default runway at KJFK with the FA/18 Hornet, take off straight out leveling off at 4000ft, cut throttle to 50%, fly over Manhattan, bank left after crossing Hudson River and land at Newark. The test encompassed everything from sitting on the runway at Kennedy to coming to a full stop on the runway at Newark. I sat on the runway for about 30 secs before starting CapFrameX to allow for everything to load in. All tests were run with time set to 12 UTC and empty rolling cache and no 3rd party add-ons.
Tests were run with a Sapphire Nitro+ 7900xtx at default settings (no OC). 9950x3d PBO was turned on with the following settings:
PBO: Advanced
PBO limits: Motherboard
PBO Scalar: 10x
Max Boost Override: +125
Curve Optimizer: By core (undervolted each core using CoreCycler utility)
Other misc settings:
- Win 11 Memory Integrity Off
- Game Mode On
- HAGS On
- ReBAR On
- Windows Gaming Optimizations On
- Monitor refresh rate 144Hz
- Freesync On
- Vitualization Off
- Hyperthreading (SMT) On
DDR5-8000 (CL36) memory OC:
All tests were run with an empty rolling cache (I would delete the cache file prior to starting the sim for each test). Other MSFS 2024 SU2 settings:
Test 1: 1080p w/ultimate power plan
Test 2: 1080p w/balanced power plan
As you can see above, the balanced power plan averaged 101 fps versus 89 fps for the ultimate power plan. So while this does show the balanced power plan as providing more fps at 1080p, the advantage disappears at 4k as they both averaged 68 fps (as per my previous 4k test above).
Again, I am concluding that no cpu CCD tweaking is really necessary on the 9950x3d with MSFS 2024 at 4k resolution (MSFS 2020 may be another story). All tests in 4k and 1080p ran as smooth as butter with little to no micro stuttering. Process Lasso may squeak out a few more fps but with the smooth fps already being achieved, I donât think it would be worth it. I think for the most part, Process Lasso was being used by simmers to mainly eliminate the micro stutters caused by the 2 unoptimized CCDâs. I believe AMD/Microsoft have come a long way to optimize the 2 CCDâs of the 9950x3d.
You can click on my gamer pic for my full PC specs.
4 Likes