In his latest 7800X3D benchmark Pauls's Hardware showed the 7800x3D clearly outperforming the 7950X3D in MSFS despite the 7950X3D's slightly higher clock (by 200 MHz = 7%) and slightly more v-cache. this doesn't make sense and more detailed benchmarks from other testers show the 7950X3D indeed slightly ahead of the 7800X3D, albeit only marginal and probably not noticeably in game. these misleading higher fps numbers of the 7800X3D might be the result of a bug in AMD's core parking driver.

obviously false benchmark result:

.........................................................

here is why, and it is not Paul's Hardware fault:

Ryzen 7 7800X3D Benchmark Test Anomaly and Solution

"That core-parking technique isn't needed for the 7800X3D because it only has a single compute chiplet. However, the PPM file provisioning driver is still needed for other purposes, so it's still installed as part of the chipset driver package. Unfortunately, due to an apparent bug in the chipset driver provided by AMD (or perhaps an enumeration issue with our test motherboard), this package enabled core parking on our test system even after a fresh Windows install, thus resulting in noticeably lower performance in several game titles.

AMD is looking into the issue, and we learned that the Ryzen 7 7800X3D's cores should never park. Unfortunately, that isn't spelled out in the reviewer guide, and our cores were parking during gaming on a fresh Windows install. Luckily we noticed this as the source of the performance issues, but we haven't had time to verify it on other motherboards due to the late discovery. "

https://www.tomshardware.com/reviews/amd-ryzen-7-7800x3d-cpu-review/3

 

Core Parking Fail

https://www.techpowerup.com/review/amd-ryzen-7-7800x3d/28.html

"When I installed the Ryzen 7 7800X3D in my CPU test system, which had the 7950X3D installed previously, the performance numbers didn't look right. Cinebench and other non-gaming benchmarks showed the expected performance, only games were slow."

pcgameshardware.de conducted a newer test which shows the 7950x3D slightly ahead of the 7800X3D, which was to be expected considering it's 200 MHz higher clock speed and slightly more v-cache.

Edited April 7, 20233 yr by turbomax

Hmm. Now I am even more leaning towards the 7950x3d. 😅

40 minutes ago, Swe_Richard said:

now I am even more leaning towards the 7950x3d.

7% increase for $ 250. nobody will feel the difference in MSFS. but less hassle without that faulty core parking trouble maker.

Edited April 7, 20233 yr by turbomax

The benchmark is absolutely flawed

FYI, the fenix external process and FSHUD keeps my CCD1 busy enough.

Edited April 7, 20233 yr by vincentrouleau

57 minutes ago, vincentrouleau said:

FYI, the fenix external process and FSHUD keeps my CCD1 busy enough.

Did you apply any tweak for cpu scheduling? 

5 hours ago, Swe_Richard said:

Hmm. Now I am even more leaning towards the 7950x3d. 😅

It makes me lean AWAY from the 7950X3D.    I do not want to tweak my simrig, I want to fly with it.   People can say the provisioning software is set-and-forget, but, experience tells me otherwise.  And then there is the added layer of complication of Windows update and game bar updates altering the scheduler.   Yeaaaahh. No.

Far too many extra layers that I have no control over (or at least, have no easy control over...) for $900 when I can get the same performance for $450.

I'd be all over a 7950X3D for a play-around rig, or if I wanted to develop for MSFS and then test on the same machine.  But for a sim-only machine, for me anyway, I'm leaning towards 7800X3D.

 

So, I assume this is just an issue if ”down”grading from the 7950/7900X3D to the 7800X3D? Hopefully I don’t run into this going from the 7600X to the 7800X3D. 

"this package enabled core parking on our test system even after a fresh Windows install, thus resulting in noticeably lower performance in several game titles."

1 hour ago, turbomax said:

"this package enabled core parking on our test system even after a fresh Windows install, thus resulting in noticeably lower performance in several game titles."

Use at your own risk, as I haven't done this, but there is a way to prevent Win11 from parking cores (with 7800X3D you would not want any core parking):

You can activate and enable ultimate performance power mode in Windows, for which one of the effects is to disable core parking. For Windows 11 at least, you run the following command ‘powercfg -duplicatescheme e9a42b02-d5df-448d-aa00-03f14749eb61’ in a terminal window with administrative rights. Ultimate Performance will then show up under power options in control panel.

Of course this may have other, unintended effects so its best to research "ultimate performance" power mode fully before implementing it.

12 hours ago, turbomax said:

7% increase for $ 250. nobody will feel the difference in MSFS. but less hassle without that faulty core parking trouble maker.

Yup. That might be true, however multithreaded performance in the 7950x3d seems to absolutely stomp the 7800x3d though. 

4 hours ago, Swe_Richard said:

multithreaded performance in the 7950x3d seems to absolutely stomp the 7800x3d though. 

True, but doesn't help MSFS a bit, and that's not what we are discussing here. just like the i9-13900K is better suited for pure multithreaded apps (handbrake, winzip, Blender, Cinebench) than the 7800x3D.

Edited April 8, 20233 yr by turbomax

Hi,

Linus Tech Tips has a review out for the 7800X3d with a very good explanation concerning the performance differences in gaming vs. other multithreading applications. The use of only 8 cores in the 7800X3D basically allows for the different physical chiplet design with a direct V-Cache access for all 8 cores. That is not the case for the 79X0X3D. The higher core count of the 7950X3D has its benefits the moment more than 8 cores are used, but gaming usually maxes out at 8 cores - so the extra cores are not utilized. SO - here the better benchmark-results are the result of the different chiplet design of the 7800X3D. (at least that is my understanding here...)

 

"This driver steers threads by parking what AMD determines is the least performant CCD, leaving only the more performant CCD initially active. This keeps all of the threads for a game on the same CCD, reducing (if not eliminating) the need to reach across the IOD to access the L3 cache on the other CCD, and thus improving the cache hit rate and resulting performance.

Technically, the PPM driver can park either CCD. But in practice it's going to virtually always be the vanilla CCD, pushing games on to the V-Cache CCD. Though should a game ask for more CPU cores (and technically, threads) than a single CCD can provide, then the PPM will allow the other CCD to go active. Parking the CCD doesn't prevent its use – so all 16 CPU cores are available  – it just discourages using more than 1 CCD (8 cores) when at all possible."

https://www.anandtech.com/show/18747/the-amd-ryzen-9-7950x3d-review-amd-s-fastest-gaming-processor