Are you getting even usage across all Cores?

[SteveW 1,792]

 

 

 

14 minutes ago, Sethos1988 said:

Steve, I just spent about an hour testing out various solutions on my CPU, which is identical to OP's. I saw your AM 5460, which I believe renders the game on the 6 "middle" physical cores, leaving the first and last unused. On one hand, it actually provided an overall FPS boost or rather, an FPS stability. Flying over Copenhagen, alongside FlyTampa's EKCH, there's a lot to render and a lot going on. On the other hand, for the first time, v4 gave me the dreaded blurries. 

- Running completely default, 8 cores, 16 threads, FPS would be choppy in the 30-38 region in a jet around Copenhagen.

- Using Process Lasso to remove the first logical core, leaving the first physical core to deal with the game alone, was a boost in smoothness and around 34-39 around Copenhagen.

- AM 5460 would be very close to my constant locked 40 (with Rivatuner), I reckon an extra stutter when it loaded huge chunks of EKCH but I saw quite a lot of blurries. 

- I then suspected the age old problem of moving rendering off the first core and thread is the reason for the blurries. So I tried AM 65532, which leaves out Core 0 / 1. Again, I saw blurries but also lower FPS due to logical cores being used, again seeing down to 30. 

- I tried AM 21845, which is basically turning HT off via masks but that was neither smooth nor gave me high framerates. Bit choppy and FPS was sailing from 32 up to 38.

- Then I tried completely turning off HT and no AM, which surprisingly seems to be the best solution. No blurries and lowest FPS was 38, almost kept a constant 40. 

So I think I might run with HT off from here on out. 

 

 

 

Means only one thing to me;

 what are you doing such that six cores can't give you a good sim?

Most around only got four cores.

[Sethos 4,538]

Six cores can give me a good sim but not when it's moved off the main thread, which is known to cause blurries and it also did in my case.

[ComSimPilot 623]

@SteveW, your input is very much appreciated and you are someone who knows a lot about how processors work. I have read many of your posts here, however my experience always shows that Ht-off produces  better performance. I have a 4core 7700K processor. I have a question about heat and performance after reading your posts:

Why heat necessarily means that the processor is better utilized? Couldn't be, that you force the same amount of workload into 4 LPs, instead of 4 cores in the case of HT-off? I mean, couldn't the extra heat be explained because P3D tasks are forced to be processed by smaller logical units, and so increased heat? In other words, does the processing "bandwidth" of a processor increases because of HT?

[SteveW 1,792]

Mathematically and physically HT enabled does more work and makes more heat. If you want to utilise HT then moderate the overclock to accommodate it.  Most systems at not used in full on stressed mode and operate in a more relaxed fashion so that HT enabled takes up the loss of a few MHz easily.

If I hear HT off produces a better result that's because the AM is not right and the way the addon exe apps run is not right.

We can't just turn on HT and expect P3D type apps to work well without the proper tuning.

 

Guys, look at this, the sim works the same with these two setups:

HT Off 8 cores =11111111=0=no AM

HT on 8 cores=01,01,01,01,01,01,01,01=AM=21845

However HT On version performs better - that's a mathematical certainty or something else is wrong! 

 

 

[Sethos 4,538]

1 minute ago, SteveW said:

Guys, look at this, the sim works the same with these two setups:

HT Off 8 cores =11111111=0=no AM

HT on 8 cores=01,01,01,01,01,01,01,01=AM=21845

However HT On version performs better - that's a mathematical certainty or something else is wrong! 

Funny that I can consistently disprove that notion on my own setup. I'll then proceed to ask, have you actually tested all the claims you put forth with P3D specifically or is it just all theory?

[SteveW 1,792]

Say you take your car in for a service and say it's a bit poor braking from above 70mph but otherwise not too bad. When you get it back you find it's been restricted to 70mph. It's a cure but not a fix.

 

4 minutes ago, Sethos1988 said:

Funny that I can consistently disprove that notion on my own setup. I'll then proceed to ask, have you actually tested all the claims you put forth with P3D specifically or is it just all theory?

I've built software to test the 'theory' presented to me (and you) by Microsoft and Intel - and it seems to work exactly as they say. Funny, don't know why your HT mode performs worse for you.

 

 

 

[loafer87gt 0]

I tried some experimenting using Project Lasso as well as disabling hyperthreading, and I found that the best performance was with the default Windows affinity mask settings and running Windows 10 game mode to give the sim priority. That being said, V4 is still a no go on my system as I just can't get acceptable performance no matter what the detail settings. The stutters are just horrible, and far worse than anything I ever experienced with 3.2 or 3.4 . Regardless of the detail settings, about every second the game has a millisecond freeze which ruins the immersion as smoothness of the sim. I even tried installing it on my system SSD drive to see if it was a disk performance issue. Still stuttered like crazy though. I am running an older 6 Core / 12 Thread 3930K at 4.8 Ghz with 32 GB or ram, and a GTX 1070 video card. Sadly, I will be uninstalling this version of the sim until I get a newer system capable of running it. Kind of disappointing as this is a vanilla install of the sim, and it runs much significantly worse than V3 with FTX Vector, FTX Global, and Active Sky.

[Sethos 4,538]

Just now, SteveW said:

I've built software to test the 'theory' presented to me (and you) by Microsoft and Intel - and it seems to work exactly as they say. Funny, don't know why your HT mode performs worse for you.

So I take that as a no.

[SteveW 1,792]

Why does that mean no to you? I've tested the theory extensively and have feedback from dozens of guys:

http://www.codelegend.com/forums/viewtopic.php?f=2&t=532

[SteveW 1,792]

HT off is where poor HT enabled setups end up - in the dustbin - and that expensive HT mode is wasted.

[Sethos 4,538]

2 minutes ago, SteveW said:

Why does that mean no to you? I've tested the theory extensively and have feedback from dozens of guys:

http://www.codelegend.com/forums/viewtopic.php?f=2&t=532

Oh look, another snooze fest wall of text from SteveW that is all theory. Surprise.

Not seeing any personal test of 8 core / 16 thread CPUs with V4 in that thread from 2014. So that's still a no.

[ComSimPilot 623]

3 minutes ago, SteveW said:

Why does that mean no to you? I've tested the theory extensively and have feedback from dozens of guys:

http://www.codelegend.com/forums/viewtopic.php?f=2&t=532

Steve, I trust you. Would it be easy for you to give me your so far "best recipe" for AF setting for an i7 7700K, overclocked at 5Ghz? I'd be very pleased. I want to do more tests.

[lsummer101 1]

WOW you guys just screwed up my marrige,, Now i have to buy New 8 core cpu and spend the next few weeks testing my set up.  I am running 4790K at 4.7Mhz stable was using only 4 cores (ht OFF) due to my testing but somthing always told me that with the proper settings the HT would help.  Its time for new MB and CPU anyway to complement the 1080. 

back to the Drawing board any lack of woopie due to an angry Wife!!! 

 

Lee 

[TechguyMaxC 516]

16 minutes ago, Sethos1988 said:

Funny that I can consistently disprove that notion on my own setup. I'll then proceed to ask, have you actually tested all the claims you put forth with P3D specifically or is it just all theory?

I doubt that very much as you have not provided any frame time evidence to support your statement, just vague "I get around X FPS" statements.  This topic has been discussed ad nauseam.  

Here's ACTUAL evidence that supports the theory: https://www.avsim.com/forums/topic/477057-a-frame-time-analysis-of-p3d-v3-effects-of-cpu-affinity-frame-lock-and-ht/

 

2 minutes ago, Sethos1988 said:

Oh look, another snooze fest wall of text from SteveW that is all theory. Surprise.

Who are you again, person that has 452 posts and has been a member here for all of a year?

[him225 51]

In quadcore giving 4 cores and one thread each the two loading jobs touch max frequently but dividing them into 4 and sharing them in LPs plus 5th shared with the second low demand job none of the jobs max out except the main thread and looks smoother.

In answer to the topic nothing has changed from v3 in this regard, increasable lod radiuses without oom may allow for more busy secondary cores giving even usage in some places.

All settings and addons same except autogen radius increased one notch I see at least 33% increase in fps and on average 50% higher fps than v3.