MSFS Core Utilization intel CPU

[ttbq1 347]

I would like to check if other users of intel cpu have a similar CPU core utilization on MSFS:

My core 4 and 5 are always at 100% and the rest of P-cores around 20% average. I have HT turned off and the last 4 e-cores are also off. I am using process lasso to assign other apps running in the background only to the e-cores.

I don't have any stutters or performance hit, but I am just curious if it is only me or other users are also having the same behavior

[Luis Hernandez 572]

I understand the behavior you see is normal: MSFS uses barely two cores. That's one of the improvements that we're expecting for MSFS2024: better multi-threading.

What I can't explain is the fact these cores are hammered harder than with any synthetic benchmark (i.e., Cinebench), and I end up getting higher temperatures in MSFS.

[bobcat999 8,484]

1 hour ago, ttbq1 said:

I would like to check if other users of intel cpu have a similar CPU core utilization on MSFS:

My core 4 and 5 are always at 100% and the rest of P-cores around 20% average. I have HT turned off and the last 4 e-cores are also off. I am using process lasso to assign other apps running in the background only to the e-cores.

I don't have any stutters or performance hit, but I am just curious if it is only me or other users are also having the same behavior

14900k here coming from a 13900k previously, and that situation is exactly the same on mine, and I believe normal on both chips. 

MSFS is still main thread limited (although this main thread runs on two cores! ) and these chips tend to select cores 4 and 5 for any high activity or demanding task, as they are normally the best performing cores on the chip for some reason.
When overclocking (admittedly not really worth it on these chips) these cores tend to perform the best.  Not sure if it is because they have better chance of cooling due to placement on the chip die, but there it is.

In some hardware diagnostic programs, including Intel's own overclocking software, these cores tend to have an asterisk (*) next to them.  If ever they auto overclock to 5.8 GHz (0r 6.0 on the 14900k), then it tends to be just these two cores.

PS.  Similar to you, I run with hyperthreading off, and only 8 active E-cores.  I find performance is exactly the same, but a lot less heat!  
Also, for people who don't know, the first core is classed as Core 0 by Intel, so it looks like cores 5 and 6 on the graph, but it isn't.

Edited February 20 by bobcat999

[SierraDelta 1,342]

Hmmm, mine's a bit different.

HT off, only 8 E-cores in use, load seems more evenly spread across the P-cores for me.

This is a IniBuilds KJFK with PSXT/RealTraffic.

[G-RFRY 1,712]

3 hours ago, SierraDelta said:

This is a IniBuilds KJFK with PSXT/RealTraffic.

And your GPU is only idling at 37% .

[AnkH 2,271]

I use ParkControl (for StarCitizen though), disabled Core Parking and I am happy with the result. But only having a 13600K, but HT on and all E-Cores active.

[bobcat999 8,484]

1 hour ago, AnkH said:

I use ParkControl (for StarCitizen though), disabled Core Parking and I am happy with the result. But only having a 13600K, but HT on and all E-Cores active.

I have never understood this.  Why would the cores park when it is flat out in a game?

[Lonesome Cowboy Burt 161]

23 minutes ago, bobcat999 said:

I have never understood this.  Why would the cores park when it is flat out in a game?

 

I wrote a load of confused rubbish that I have now deleted...sorry

Edited February 21 by Lonesome Cowboy Burt

[AnkH 2,271]

1 hour ago, bobcat999 said:

I have never understood this.  Why would the cores park when it is flat out in a game?

No idea, eventually because load is unevenly distributed and Windows thinks that some cores are not used and parks them? If I am not mistaken, this happens in milliseconds, yet it leads to stuttering in StarCitizen... Effect on MSFS is unclear, but it does no harm.

[vonduck 99]

21 hours ago, bobcat999 said:

14900k here coming from a 13900k previously

Hi Bob, quick question... notice much difference in upgrading your CPU? Am on the fence whether to upgrade from mine... hmmmm

[cavaricooper 2,615]

If MSFS2024 spreads over all available cores, that alone will be worth the price of admission!

[ttbq1 347]

20 hours ago, SierraDelta said:

Hmmm, mine's a bit different.

HT off, only 8 E-cores in use, load seems more evenly spread across the P-cores for me.

This is a IniBuilds KJFK with PSXT/RealTraffic.

I also saw a similar screenshot somewhere and triggered my concern, why my core 4 and 5 are always at 100%, will I get more smoothness if my cores are equally distributed?

I have tried the following:

1. Uninstall Process Lasso

2. Change from DX12 to DX11

3. Vsync on/off

all of those changes and the same results...I am out of ideas.

[SierraDelta 1,342]

57 minutes ago, cavaricooper said:

If MSFS2024 spreads over all available cores, that alone will be worth the price of admission!

Look at my screenshot - MSFS (2020) already does - for me at least 😉

 

[Cpt_Piett 4,711]

SU14 DX12 hyperthreading on. 

[cavaricooper 2,615]

How are you accomplishing this??