Thanks for your review my mobo supports this CPU and I'm tempted to go there.  My only question would be if my Noctua air cooler could handle it.  Right now temps in either sim are around 69C, 65C in colder weather but will get up to 72 or 73C during the summer months here.

By your best estimate what kind of increase in frame rate where FPS-fg peaks at say 70 might I see with the new CPU over my current 7800x3D?  

Tried my hand in the graph stuff. This was off of runway 7C, not 7L.

As you can see from the charts, the lower FPS numbers are a bit higher in FS2020 which equates to smoother performance. Obviously, the numbers don't lie, but anecdotally, both sims "felt" very smooth with FS2020 feeling much smoother overall, despite having lower FPS. This isn't the most "realistic" scenario since most people are probably using AutoFPS, Vatsim, and/or some type of traffic injection. Why not start at the top shelf and hope for the best? 😅 I'm going to post more charts of the same scenario while using AutoFPS, and again with AutoFPS and BATC.

 

 

33 minutes ago, Noel said:

By your best estimate what kind of increase in frame rate where FPS-fg peaks at say 70 might I see with the new CPU over my current 7800x3D?

I can't say for sure because the 7800 is a great CPU. But if I were to guess, it'd probably be somewhere in the neighborhood of 10-15% FPS increase, however I'd imagine you'd have better overall smoothness which, historically has been your focus. The "50" versions in either generation of processors would provide you more room for background tasks leaving more raw power for processing the core sim.

Changed the graphic format up a bit, as this one makes more sense to me, personally.

The TLOD settled to 280 on the ground using AutoFPS and NO Traffic.

 

 

The TLOD Settled to 230 on the ground WITH Traffic. There were 57 visible A/C at the time of recording.

Edited March 13, 20251 yr by V1ROTA7E

This is with FS2020.

TLOD Settled at 230 on the ground, BATC is OFF.

 

TLOD for some reason wouldn't go below 200 on the ground, so it has a lower base FPS...and I'm also not sure what that frame spike was about, I honestly didn't even notice it while going through the scenario.

13 hours ago, AnkH said:

You need the Windows GameBar being installed for this "feature" that Windows puts the game only on the CCD with Cache. No GameBar, no automatic assignement. Big show stopper for me. 

I dont believe that's the case with the 9950X3D using the new chipset drivers... not from what I'm reading anyway. It seems the drivers themselves are enough to manage the allocation of threads without needing GameBar on which would be great if true.

 

5 hours ago, V1ROTA7E said:

This is the case with FS2020, FS24 I’m noticing far more overall cpu usage. I’m going to run some tests because even with the increase in usage, performance in FS24 appears to be quite a bit worse compared to FS20 and it was the opposite before the switch. 

I noticed this on my 7950X3D When using 2024. 2024 seemed much better at spreading the load over multiple cores than 2020 did...and in many instances I'm not seeing a huge increase in FPS by just using the X3D cores which I always found was the case with 2020

Deleted my previous post here and opened a new topic for the 9900X3D

 

Edited March 14, 20251 yr by mpo910

On 3/13/2025 at 3:26 PM, V1ROTA7E said:

Changed the graphic format up a bit, as this one makes more sense to me, personally.

The TLOD settled to 280 on the ground using AutoFPS and NO Traffic.

 

 

The TLOD Settled to 230 on the ground WITH Traffic. There were 57 visible A/C at the time of recording.

Do you still have these graphs? They seem to have been deleted.

I'm trying to decide whether to upgrade to a 9800x3d or a 9950x3d. Did you ever try assigning 3rd party apps to run only on the NON 3d cache ccd using a program like process lasso? I'm wondering if forcing things to be split up like this would have any benefit.

1 hour ago, RichConn81 said:

Do you still have these graphs? They seem to have been deleted.

I'm trying to decide whether to upgrade to a 9800x3d or a 9950x3d. Did you ever try assigning 3rd party apps to run only on the NON 3d cache ccd using a program like process lasso? I'm wondering if forcing things to be split up like this would have any benefit.

I’ll have to re-upload them. For the background tasks, yes. 

On 3/13/2025 at 5:02 PM, V1ROTA7E said:

 if the game is in focus, you’ll see CCD0 at near full activity and CCD1 relatively idle. 
 

Houston, we have a problem !
🤕
I've just received my 9950X3D.
I've followed usual recommendations (AMD chipset update, last Asus mobo Bios, CPPC on "driver", game mode ON, game bar ON, power setting on "balanced") but ingame, CCD0 is parked when CCD1 is showing full activity!

Why??? It should be the opposite !

 

Edited March 26, 20251 yr by David Roch

5 hours ago, David Roch said:

Houston, we have a problem !
🤕
I've just received my 9950X3D.
I've followed usual recommendations (AMD chipset update, last Asus mobo Bios, CPPC on "driver", game mode ON, game bar ON, power setting on "balanced") but ingame, CCD0 is parked when CCD1 is showing full activity!

Why??? It should be the opposite !

 

make sure you reset process lasso if you're using it. sometimes process lasso will override balanced mode. you could also try setting CPPC to 'prefer cache'

6 hours ago, David Roch said:

but ingame, CCD0 is parked when CCD1 is showing full activity!

Be aware:

As soon as you take a look at another piece of software (aka Ressource Monitoring, Taskmanager or whatever) MSFS loses focus AND THEN the calculations are shifted to CCD1 and other cores may look as "not doing anything  at all".

MSFS needs to be in focus to run on ccd0.

 

Edited March 26, 20251 yr by mpo910

Thank you both. 😊

10 hours ago, V1ROTA7E said:

make sure you reset process lasso if you're using it. sometimes process lasso will override balanced mode. you could also try setting CPPC to 'prefer cache'

In order to test the new CPU I first want a set and forget option, so I let the Windows Game Mode take care of things without Process Lasso. Later, I will switch CPPC to "Frequency" and use it.

Yes, I've seen that Cache is using the right cores but I wonder if it is a safe solution. I am affraid it'll be forcing everything on the cache CCD, which depending on how many programs you run in the background, could choke your games of the precious 8 v-cache cores. Am I wrong?

9 hours ago, mpo910 said:

Be aware:

As soon as you take a look at another piece of software (aka Ressource Monitoring, Taskmanager or whatever) MSFS loses focus AND THEN the calculations are shifted to CCD1 and other cores may look as "not doing anything  at all".

MSFS needs to be in focus to run on ccd0.

 

Yes, you are right and I do split the screen in 2, MSFS on one side and Ryzen Master on the other one, putting the focus on the simulator. 😉

********************************

After tearing my hair out for most of the night, I managed to put the cores back in place.
In fact, the chipset drivers were already installed and Ryzen Master configured when I replaced my "old" 9950X with the 9950X3D, and AMD Adrenalin didn't require reinstalling them. Hence my problem.
After uninstalling and reinstalling the AMD drivers and Ryzen Master, CCD0 is now fully utilized during MSFS and rests for all other tasks, leaving the little CCD1 to do the work.

This automatic solution is, of course, less efficient than an intelligent use of Process Lasso (excellent tool), but it's still a great "collaboration" between AMD's Ryzen and Windows. I like it.
Coming from a 9950X I don't really notice any terrific improvement in terms of fps and I didn't expect any, but MSFS2024 is really smoother and I couldn't be more satisfied!
 

12 minutes ago, David Roch said:

smoother and I couldn't be more satisfied!

Yes.....that is a remarkable step up indeed as I see this too. 

I find especially MSFS2020 is a huge huge difference with this chip. 

But msfs2024 also . .just the sim needs some work from Asobo. 

These are the original graphs that got deleted. should all be in the same order. I can, if anyone is interested, run a new test with the OC stable.