Limited by MainThread

[Noel 4,504]

Hopefully reports of less impact at the arrival airport with the beta may help that.

[iFlySimX 751]

Also, some people in the SU 10 Beta Forum reported that turning off HT boosts performance. Anyone try this and know if this is true?

[roi1862 602]

18 minutes ago, iFlySimX said:

Also, some people in the SU 10 Beta Forum reported that turning off HT boosts performance. Anyone try this and know if this is true?

It is true in my case. Hard to believe we back at this FSX/P3D tweak... Without HT off DX12 is horrible ! 

BUT, there are other issues with it. Graphical issues, white flickering, missing parts of runway/taxway and more... 

Hopefully a new build will be out to test  

Edited July 13, 2022 by roi1862

[MrFuzzy 2,118]

1 hour ago, roi1862 said:

It is true in my case. Hard to believe we back at this FSX/P3D tweak... Without HT off DX12 is horrible ! 

BUT, there are other issues with it. Graphical issues, white flickering, missing parts of runway/taxway and more... 

Hopefully a new build will be out to test  

Evidently there are issues with your specific configuration.

I have HT on (20 threads), DX12 is smooth and I have no graphics issues.

[roi1862 602]

1 minute ago, MrFuzzy said:

Evidently there are issues with your specific configuration.

I have HT on (20 threads), DX12 is smooth and I have no graphics issues.

So go back to MSFS forums and collect few more configurations reporting same... 

[iFlySimX 751]

Turned off HT and WHAT A DIFFERENCE!! I am simply amazed....Will need to do more testing but man this is golden. 

[RTK1972 288]

that is crazy. i am also on a 9900K but 2080TI and turning off HT gives me +10FPS at LPPT. Super smooth now under DX12

[Seth2021 291]

I never liked the idea of turning off HT. I just made some tests and looks like DX12 works better with out. I used Fenix A320 with Process Lasso and made some rules for CPU affinity.

MSFS took 11900k physical cores ( 0-2-4-6-8-10-12-14)  while FenixDisplay.exe the rest of virtual cores ( 1-3-5-7-9-11-13-15) and for the first time i didn't have the Single Main thread hammered. MSFS with HT off spreads more "equally"  the process work to Physical cores which i guess DX12 helped also. Sim was much smoother than before even in low-fps. I didn't gain fps but the whole immersion was much smoother even with tod 200 on high demanding airports.

I will make more tests as DX12 looks ultra promising

[MrFuzzy 2,118]

14 hours ago, roi1862 said:

So go back to MSFS forums and collect few more configurations reporting same... 

There must be something wrong (or peculiar) with those configs. 

I just tested a main thread limited situation (Salty 747 at EDDF) and with HT off I have approximately the same framerate, only more stutters when panning the view.

 

[bobcat999 8,484]

Other than to reduce overheating of the chip if it is getting hammered over a long period, turning off HT makes no sense, so I wonder what is going on here?

Hopefully there is technical explanation for this that will become apparent. 

I don't want to turn off HT in the BIOS, as I use a lot of video processing and music conversion tools that will take about twice as long if I do that.

It is a shame if we have to go down the Process Lasso route again.

Edit:  Just saw Fuzzy's post.  Seems to make more sense.  I can only think it is down to the individual configuration and characteristics of each PC.
I am leaving HT on, as I have no problems anyway on my PC and with MSFS.  12900k here.

[Sethos 4,358]

God, are we still beating that hyperthreading drum in this day and age. 

Edited July 14, 2022 by Sethos

[MrFuzzy 2,118]

1 minute ago, bobcat999 said:

Other than to reduce overheating of the chip if it is getting hammered over a long period, turning off HT makes no sense, so I wonder what is going on here?

Hopefully there is technical explanation for this that will become apparent. 

I don't want to turn off HT in the BIOS, as I use a lot of video processing and music conversion tools that will take about twice as long if I do that.

It is a shame if we have to go down the Process Lasso route again.

I tested Process Lasso extensively too, in the past. Snake oil as well.

Another one infesting the forums for months was this: https://devblogs.microsoft.com/directx/demystifying-full-screen-optimizations/

Imagine hundreds of professional programmers working for the most important software multinational company ignoring for years such simple youtuber tricks.

Yes, if the CPU is not properly cooled, HT may trigger substantial thermal throttling and so the higher performance with HT off is explained by higher clocks. The same for the CTDs that ruin the experience of many players and for which they blame MSFS, but honestly beside some reproducible scenarios (such as the VFR map resizing now - but it's a beta! - and incompatible addons in general), most of them are caused by unstable XMP profiles (in particular 0xc0000005).

[MrFuzzy 2,118]

8 hours ago, Seth2021 said:

MSFS took 11900k physical cores ( 0-2-4-6-8-10-12-14)  while FenixDisplay.exe the rest of virtual cores ( 1-3-5-7-9-11-13-15)

There is no distinction between physical cores and HT ones. They are all couples of virtual logical cores for the OS.

When HT is enabled you cannot choose to send instructions to a physical core rather than a "virtual" one. 

Placebo...

[roi1862 602]

22 minutes ago, MrFuzzy said:

There must be something wrong (or peculiar) with those configs. 

I just tested a main thread limited situation (Salty 747 at EDDF) and with HT off I have approximately the same framerate, only more stutters when panning the view.

 

4 minutes ago, MrFuzzy said:

There is no distinction between physical cores and HT ones. They are all couples of virtual logical cores for the OS.

When HT is enabled you cannot choose to send instructions to a physical core rather than a "virtual" one. 

Placebo...

It doesn't matter you think or what you found or that it doesn't make sense. The fact is that multiple users reporting the same, and the reports keeps pilling up on MSFS forums and with few streamers i saw. Now its Asobo job to figure out why. That's all  

Edited July 14, 2022 by roi1862

[Sethos 4,358]

1 minute ago, roi1862 said:

It doesn't matter you think or what you found or that it doesn't make sense. The fact is that multiple users reporting the same, and the reports are keep on pilling up on MSFS forums and with few streamers i saw. Now its Asobo job to figure out why. That's all  

Let's also keep in mind this is flight simmers we're talking about. The most anecdotal, "magic tweak" bunch in the games industry. They can install a new Nvidia driver and report a 62% performance gain and they never have any benchmarks or data to back anything up.