Are you getting even usage across all Cores?

[GSalden 4,655]

Steve, my 5820 ( 6-core ) is running great with AM=340 in P3Dv3.  

Can I use the same setting in v4 or has something changed and I need to use another value now ?

Thanks,

 

[DaveCT2003 1,820]

2 hours ago, glider1 said:

the old outdated archaic single core processor FSX code still existing in P3D"

Not sure what your talking about when you reference FSX being single core design.  In my post above, I provide a link to an article which contains a quote from Phil Taylor, the member of the ACES Team who helped create the multicore/thread compatibility.

My very best wishes.

 

 

 

[SteveW 1,792]

6 hours ago, GSalden said:

Steve, my 5820 ( 6-core ) is running great with AM=340 in P3Dv3.  

Can I use the same setting in v4 or has something changed and I need to use another value now ?

Thanks,

 

So far I've not had time to do a proper thorough analysis of v4, but 340 is a good starting point Gerrard. 64bit does slightly less work per thread so there will be differences, but the same theory should hold.

 

 

[SteveW 1,792]

7 hours ago, glider1 said:

Does the core usage graphs measure actual physical usage of the core or just allocation of processor time?

What I'm saying is I think the fact that core-0 is pegged at 95%-100% doesn't mean the core is actually working at 100%, just that it is allocated 100% to distribute load to the other cores as needed. This makes sense to me. You always need one logical process that is continuously allocated for the purposes of shedding load.

If the theory is correct, core-0 should not run hotter than any other core even if the performance graph indicates it is loaded to 95%-100%.

Would be good to get clarity on this, because the story about "the old outdated archaic single core processor FSX code still existing in P3D" continues to this day.

The reason I would like some clarity, is that it would be wrong from an engineering point of view, to physically stress only one core on the silicon die more than other sections of the die, because this would reduce the mean time before failure (MTBF).

@SteveW or anyone else, got a comment on this?

Core usage graphs in Task Manager show the percent of current capacity of that Logical Processor. I mentioned earlier that two LPs of the same core showing 100% are both at 50%, and then the core may only be at 50% turbo throttled.

When the CPU nears overheat it throttles back.

Depending on what the multi-threaded software has to calculate there can be really good advantages splitting the load over as many cores as possible. However it may not be possible to do that with some algorithms can't be worked out in parallel until some other calculation is performed. With these sims the overriding performance factor is running the renderer which we want as many fps as possible it goes flat out to do that while assimilating data collected by the other threads.

As Dave showed us with Phil's blog there's a lot of work gone into FSX so that it can do the same job on one core as it does on as many cores we can give it. It 'sees' the number of cores and splits itself out over as many as it can. However there's only so much it can split it out to, at four parts the rendering stage is at the leanest, we can add more cores but only the background data gathering can split out more from there.

It does this by use of monolithic processes that can be deliberately assigned separate cores - not because the code is lame as we often see suggested.

There's a physical limit what the hardware can do, so although Phil talks about hundreds of cores a possibility, in practice after six or eight there will be no improvements. With more parts to arbitrate that's another overhead worth avoiding so we only want to give it the moderate number of cores that work well enough.

[Bert Pieke 7,584]

7 hours ago, glider1 said:

Would be good to get clarity on this, because the story about "the old outdated archaic single core processor FSX code still existing in P3D" continues to this day.

The reason I would like some clarity, is that it would be wrong from an engineering point of view, to physically stress only one core on the silicon die more than other sections of the die, because this would reduce the mean time before failure (MTBF).

@SteveW or anyone else, got a comment on this?

It may be Wrong, but that is what it is.

[Boeing or not going 578]

19 hours ago, SteveW said:

FSX and P3D are similar in many respects but are quite different whereby P3D can have more intense scenes than FSX and so poor loading bandwidth can cause problems such as blurriness. An AM used with FSX may provide a good balance with background performance on FSX but the same setup might not be good enough for P3D scenes. P3D operates in a slightly different way to FSX with less impact on the main process when adding more than four LPs, more than four add to background performace.

I have an i7 6 core overclocked to 4ghz. and HT on. Can you recommend an AM for me to try for p3d v4?

[SteveW 1,792]

2 hours ago, 777200lrf said:

I have an i7 6 core overclocked to 4ghz. and HT on. Can you recommend an AM for me to try for p3d v4?

Try 340.

[garrett_frank 150]

I hate to be that tl;dr guy (though I have actually read not only 80% of this thread but many related ones as well) but it seems difficult to find a consensus whether or not those of us (i.e. many of us) that have 4 [physical] core CPUS (e.g. a 6700k) should be using HT or not with the later iterations of P3D (v4, specifically). I gather that all of our systems and setups are different but how much difference does that make? It seems like most of us are using 4 physical core processors and that there's also a smaller segment that has the 6 core (not knocking the latter of course, they're always over my budget and wasn't sure what to make of them tbh).

[Bert Pieke 7,584]

I doubt that you will ever get a consensus on this question.

First of all, neither FSX nor P3D were designed for "even core usage"..

FSX/P3D is based on a single core based initial design with later efforts to offload as many tasks to the other cores as possible.

If you read Phil Taylor (of Aces fame) carefully, you will note that he understands that if the airplane is standing still and no scenery is being loaded or re-lit, a single core will carry most of, if not all of the load.

Only when the airplane (or the view) starts moving, do the other cores kick in.

As to HT on or HT off.... there are as many opinions as there are users

[bean_sprout 1,201]

Cores 0 and 2 @90%, 1 and 3 @ 70%.  Very kewl

Cheers 

bs

[garrett_frank 150]

1 hour ago, Bert Pieke said:

I doubt that you will ever get a consensus on this question.

Hi Bert. Thanks for your reply. Not to get philosophical but, telling me that there is no consensus, to me, is a good consensus to work with... for now, at least...

I did note earlier that Steve said something to the effect that HT off configurations are where all HT on (but with AM tweaking) configurations wind up. I did plenty of testing back in P3Dv2 and 3--and as scientific as I could get, i.e. repeatable results and completely mimicking various scenarios--and found that my i7 4-core always did better with HT off. And until only recently when VAS was an issue, apparently HT was running address space out into the nether regions so it seemed like yet another reason to leave it off...

[SteveW 1,792]

Don't forget that hyperthreading looks like straight regular cores to the apps so VAS issues can't be attributed to HT on or off, but rather it could be attributable to the number of cores the app runs on. What four core HT Off guys are suggesting is that eight cores is bad.

[SteveW 1,792]

...and then you have the eight core guys saying that four cores is bad.

[331BK 96]

Soon we will have the 18/36 core guys saying all the others are bad ;-)

 

Would be interessting what such a beast can do with v4.  But i think they are clocking to slow for p3d

[lsummer101 1]

To all, FIrst my appoligies for a minor hijack but i am now desprate,  SO after reading this thred i decied to mess around with My CPU settings in Bios, the first thing i did was to change the OC up one noch (greedy) i was at 4.7Mhz stable for over a year now. ( i know this was dumb) then i turned on HT to see the differance. GOt the blue screen did not want to mess with Voltage so i just set it back to its origonal OC at 4.7..  Booted up but locked up in less then 5 min. so went back in and turned of HT. No differance kept locking up. went back in and set OC down to 4.5 Mhz it running now but all four cores are showing 100% that never happend before.. and getting strange feedback sound thru speakers. 

Is it possible i cooked the CPU?? with no voltage ajustment. also noticed the cpu dropping to 280 Mhz for second then right back up to 4499. 

Please if you can explaine to the ignorant dummy that messed with a perfectly working system what might have happend.