No you’re not. People saying ‘the sim only utilizes one core’ are just plain wrong.
The sim IS however limited by the core that’s running the main thread (that’s why single core performance is still important).
From what I’ve seen the Sim will benefit up to a 6C/12T CPU. After that more cores/threads don’t add anything (the 5 or 10% load that’s being shifted to core 7 or 8 could easily be handled by cores 1-6, since they’re normally not really above 30-40%.
Thanks…
Okay, just thinking. That is why my upgrade from 9700 (8C-0T)
to 9900 (8C-16T) made the difference since the only other change
was from 4.7Gig to 4.8 Gig in the CPUs.
I did notice a very large increase in the scenery. Wondering why?
In my previous post I said that the sim noticeably benefits from 6C/12T. 8C/8T is 4 threads less, so it makes sense that you see a benefit (more than just the clock speed increase).
An 8 core CPU will always have 8 threads (each core has a single thread), so you can’t have 8C/0T 
Not sure what you mean with ‘large increase in the scenery’?
The Intel i9-9900K is an 8 core, 16 thread, unlocked 9th generation Coffee Lake processor.
Intel’s Core i7-9700K is an unlocked 9th generation Coffee Lake CPU. It is an 8 core, 8 thread processor.
Okay, you got me. 9700 (8C-8T). I stand corrected.
“8C/8T is 4 threads less than” (a (8C16T ))
I think is is 8 threads less (16 - 8).
“8C/8T is 4 threads less than” (a (6C12T ))
I should have bought a (6C-12T) to upgrade.
Why don’t they call it a 8 thread or 12 thread or 16 thread CPU if
number of cores don’t matter?
Don’t know how to prove the improvement in scenery. I can see it.
I can’t make you see it.
Because of Intel’s marketing department I guess.
About the scenery; I just didn’t understand what you wrote 
On the same settings I wouldn’t expect you to get more / more detailed scenery, just a more smooth experience that maybe makes you notice the scenery more since it’s less stuttery maybe?
Anyways, good to hear the upgrade paid off for you.
Thanks…!
Just testing. Original upload did not work.
[Screenshot 2020-12-21 152925|690x459]
Well, I don’t know how to make a screenshot display.
Why don’t they call it a 8 thread or 12 thread or 16 thread CPU if
number of cores don’t matter?
For many things it does matter. For, say encoding video or compiling Chrome from source code to an executable, increasing core count can often scale performance linearly because there are many small jobs to be run continuously on all available cores.
In MSFS, however, you’re limited by the fastest core (running the MSFS main thread), and there’s only enough other work to fill up about 4-6 cores.
So for MSFS, a specifiic workload that is not well optimized for multi-core/multi-thread processors, having more than 6 cores makes no difference because there’s no additional work to do.
On a different workload, with different performance characteristics, it does make a difference.
Ah also you have misunderstood the difference between cores and processor “threads” which are different from program “threads”.
Processor “threads” are logical cores. It’s an instruction handling pipeline frontend to the core. One core contains either one thread or two, depending on the processor and the configuration.
If your operating system finds that the workload of your program needs more program threads running than your computer has physical cores, then it will start to allocate additional program threads on the second CPU thread of each core, whereas until that point it will only schedule work on one CPU thread for each CPU core, which keeps things running smoothly.
You never want to run two program threads on the two CPU threads of the same CPU core unless you have already filled up the first CPU thread on every available CPU core, because when you run both CPU threads at once with workloads, they both slow down significantly because they share execution units.
But, in that case the combination of the two will more efficiently fill in idle execution units in the CPU core that’s shared by the two CPU threads. On workloads like video compression and compilation of program source code, this is good – you might get a 10% boost in total throughput.
For MSFS, additional CPU threads will make zero difference beyond 6 CPU cores because there isn’t enough work to fill the first CPU thread of each CPU core already. You can verify this by disabling “hyperthreading” (Intel) or “symmetric multithreading” (AMD) in your BIOS/UEFI configuratoin and playing more MSFS.
We really need to stop talking about ‘filling up cores’ in this context, and need to start talking ‘filling up threads’. There’s still CPU’s around with only 1 thread per core (yes Intel, I’m talking to you), and there’s a huge difference between a 8C/8T CPU and a 8C/16T CPU.
There’s still CPU’s around with only 1 thread per core, and there’s a huge difference between a 8C/8T CPU and a 8C/16T CPU.
Not on MSFS there isn’t.
For compiling Chrome? Yes absolutely! It’s 10-15% faster with the additional threads.
Note I have already confirmed this experimentally on Ryzen 3700X by disabling symmetric multithreading (AMD’s name for Intel’s hyperthreading). I’m not just making this up.
Sorry, but you’re wrong. People running a 8C/8T CPU run into issues, while those running a 6C/12T CPU or 8C/16T CPU aren’t.
edit: Maybe it’s something with Intel’s implementation of ‘Hyperthreading’, but there’s an example right here in this thread of someone upgrading from an Intel 8C/8T CPU to an Intel 8C/16T CPU, getting better performance with the 8C/16T.
Sorry, but you’re wrong. People running a 8C/8T CPU run into issues, while those running a 6C/12T CPU or 8C/16T CPU aren’t.
Why, then, do no issues appear when disabling the additional threads on an 8C/16T CPU by disabling SMT or hyperthreading?
(Please feel free to mention specific issues. I’ll be happy to test again and look for them.)
Just editing my previous post, might have something to do with Intel’s implementation of ‘Hyperthreading’.
I’m not sure offhand which post you refer to but different CPUs will also have different clock speeds, cache amounts, and other performance characteristics. Disabling hyperthreading on the new CPU in the BIOS/UEFI and comparing performance results would resolve this question very simply.
Just scroll up a bit (like 4 posts) to @MSFSRonS post. Upgraded from a 9700K to a 9900K, so same architecture / generation.
I don’t have an current gen Intel CPU to test this with, but I believe him when he posts his results.
It wouldn’t surprise me at all if AMD and Intel have different ways / implementations of multithreading, and that one might be more effective than the other.
I would do it but I can’t figure out how to display my screen shots.
No help yet.
i9-9900K, Citation CJ4 rolling down the runway on takeoff
CPU Usage 16%
CPU 1 14%
CPU 2 14%
CPU 3 22%
CPU 4 13%
CPU 5 17%
CPU 6 11%
CPU 7 48%
CPU 8 09%
I would do it but I can’t figure out how to display my screen shots.
Here’s what I do for my screenshots:
that seems like a lot of work… I have ‘ShareX’ installed, so I just make screenshot from the sim, and do CTRL-V in the forum
Are those 9700 or 9700K & 9900 or 9900K? Original post says 9700 and 9900. We can look up the specs for the exact processors and check what’s different.
Update: here’s some links
The 9700K and 9900K each have a significantly higher clock speed and bigger memory cache than the 9700 and 9900.
If someone upgraded from a 9700 to a 9900K, for instance, they would probably experience a big improvement in main thread performance – and this would have zero relation to the presence of hyperthreading.