It's scary how I finally made 5.1 working so good...Now I am thinking what to do. Only reason I haven't updates is lack of FSL and that keeps me out for now. And for the life of me I can't re-trace what exactly I did to get this performance and EA working so good in 5.1 to replicate to 5.2 if needed.

 

33 minutes ago, PATCO LCH said:

That is my suspicion. They moved more work load back to the CPU to help the VRAM overload so many experienced.  

That means, my super duper GPU is useless now ...

1 minute ago, Nemo said:

 

That means, my super duper GPU is useless now ...

Consider this: some of the people in this thread might not know what they're talking about. 

23 minutes ago, Chapstick said:

Consider this: some of the people in this thread might not know what they're talking about. 

Yea, me for one. Sorry.

1 hour ago, GSalden said:

Could it be that making less use of VRAM resulted in more use of the main thread core ? …

My VRAM usage is about 2/3 of what it was with v5.1 and I wondered how they pulled that off.

My thoughts exactly. I’ve just spent a couple of hours taxiing around various airports I know are quite demanding of VRAM such as PILOT’s LEAL and VRAM usage is definitely down but RAM usage is significantly up. Hitting 15.7GB at several large airports with plenty of Ai. Previously it topped around 13Gb.

If that’s the case I wish they’d been open about it because I probably wouldn’t have gone to 5.2.

My somewhat modest 1080Ti is being used far less and my CPU is being driven quite a bit harder. It’s like going back to FSX days. Not a good move I’m afraid.

2 hours ago, Luke said:

I’d be interested if there’s more AVX usage in the new build. Do you have an AVX offset enabled in the BIOS?
 

 

1 hour ago, pgde said:

https://linustechtips.com/topic/805187-what-exactly-is-avx-and-do-i-need-it/

I haven’t looked and neither have I seen any mention of that in all the discussion about P3D. I haven’t touched the BIOS in over 18 months. I suspect this is not going to make any difference at all. The only thing that has changed is going from 5.1 to 5.2.

Dumb question, but is this excessive load on the CPU going to do damage? If so this definitely concerns me.

Cheers, Pete

6 minutes ago, PilotPete99 said:

Dumb question, but is this excessive load on the CPU going to do damage? If so this definitely concerns me.

Cheers, Pete

Who knows but when usage of CPU1 hits 100% that’s when P3D cannot deliver the performance you have set resulting in stutters and pauses. These are far more frequent than with 5.1.

Reducing Scenery Complexity to Sparse still caused it to hit 100%. Not good.

Just now, Ray Proudfoot said:

Reducing Scenery Complexity to Sparse still caused it to hit 100%. Not good.

There is no way that can be correct. they must have messed something up big time if that is the case. People will be frying their CPU's now. @mpo910

9 minutes ago, B777ER said:

There is no way that can be correct. they must have messed something up big time if that is the case. People will be frying their CPU's now. @mpo910

Something doesn't make any sense. Almost impossible that non of the Beta testers reported such an issue and was ignored. Although not sure how testing is done. If they had all addons on they should notice. No?

Edited June 13, 20214 yr by cyyzrwy24

I have checked my cpu usage each time I have gotten ready for a flight and every time it was extremely high. This is puzzling, hope they find out what is causing this because I don’t think I’m capable of figuring it out.

Cheers, Pete

It’s very easy to get a thread to take 100% of a CPU core by accident. It doesn’t actually mean it’s doing much of anything. I could write an empty loop and make a mistake such that it runs forever, and it would pin a core at 100% but all it’s doing is going round and round and achieving nothing. Seeing high CPU usage does not mean that P3D is suddenly inherently more CPU intensive somehow. It is most likely a bug. 

For example, I’m working on an add-on at the moment which runs tasks on several threads, and each of those tasks sits and waits for something to do, and in the meantime they just loop forever to avoid the thread terminating prematurely. But at the end of each iteration of the loop I have an instruction that indicates to the Windows scheduler that it can give my time to another process, and the CPU usage for my thread drops to 0 when it’s not actually doing anything useful.

The point being, this could be caused by any number of things that are simple bugs and would be very easily fixed. Or it could be caused by a combination of factors that don’t necessarily apply to all systems. It may not even be P3D itself which is the cause. 

Edited June 13, 20214 yr by neilhewitt

15 minutes ago, B777ER said:

There is no way that can be correct. they must have messed something up big time if that is the case. People will be frying their CPU's now. @mpo910

I’ll double-check tomorrow. But why switch processing to the CPU when the GPU has so much spare capacity? It needs to be more intelligent in how it decides what component needs using.

17 minutes ago, cyyzrwy24 said:

Something doesn't make any sense. Almost impossible that non of the Beta testers reported such an issue and was ignored. Although not sure how testing is done.

They would have to use Afterburner and RTSS to monitor usage. I found it so useful when setting complexity levels after getting v4 I’ve kept on using it.

Without it I’d know performance wasn’t as smooth but wouldn’t know why.

The old "Affinity Mask trick" cannot solve it? No possibility to try it for myself at the moment...