couldn't see any difference in overall performance so went back to HT disabled.

 

HiRes terrain loading is the only difference I see (no fps change) ... slightly better and more able to keep up.  But like I posted earlier 4.2 Ghz or 4.8 Ghz = no difference in fps, so it appears my weakest link now is GPU.

 

 

weakest link now is GPU.

 

With a Titan GPU. They make better? I suppose with a couple three thousand somolies.

HiRes terrain loading is the only difference I see (no fps change) ... slightly better and more able to keep up.  But like I posted earlier 4.2 Ghz or 4.8 Ghz = no difference in fps, so it appears my weakest link now is GPU.

 

The most frustrating thing for me is just how variable the GPU load is.

 

And that I can't get MGPU working in any mode.

 

My ~1.4 Titans worth of GPU power should be able to run all the bells and whistles.

 

I'm dying to see that in action.

Well it has been changed no doubt, though I'm not sure it is necessarily a bad thing.

 

I remember seeing much better runway lighting effects et al, but I also remember it was pretty easy to completely wash out the virtual cockpit in certain conditions, with no option but to turn off the effect if you actually wanted to see anything.

 

Integral lighting also had a tendency to bloom a bit too much at times. I personally feel that the biggest issue it had, was that from the start it darkened the scene a bit too much, pushing the contrast over the top. Now, it seems that in order to fix the over-bloom/washout, rather than lessen the amount of darkening, they've instead lessened the amount of bloom.

 

I would have rather they went the other way, leaving bloom as it was and keeping the scene a bit brighter to begin with, but there may very well be a reason they weren't able to do that. 

 

I wonder if the bloom on the runway lights was separate from the real HDR lighting effect or just part of the scene getting washed out. It definitely wasn't like that in the DX9 bloom, so I can't see how they couldn't reduce the effect of HDR while keeping the same bloom level.

The most frustrating thing for me is just how variable the GPU load is.

 

 

 

I agree.  In urban areas and OVC my gpu takes a beating and framerates drop significantly enough to be rather annoying, especialy with trackIR. If I kill cloud shadows, and reduce a few sliders then some semblence of fps sanity resume, but then this is no better graphically than fsx. It's  unfortionate, but FSX is still running smoother in the same conditions with many more addons, SweetFX, SSGS, AA and is generally much more stable (less variable) in the fps department, atleast for me.  

 

I am starting to get a little skeptical of the philosophy of letting the gpu handle almost everything.

 

 

With a Titan GPU. They make better?

 

I thought the 780 OC  was supposed to be a little faster (and less costly)?  Let me qualify, with my P3DV2 graphics settings, it seems the GPU is the weakest link, not a "bad" link.  But point being, with my settings and hardware, I need more GPU (actually I should say, I need support for more than one GPU, I have another Titan sitting idle). 

 

 

actually I should say, I need support for more than one GPU, I have another Titan sitting idle

 

I am considering to get an SLI setup in my new computer but when I read 'SLI' I also read 'stutters'. Now I've heard that nVidia is doing great work on getting rid of those stutters but since P3D, like FSX, is a sim that has a rather strong tendency to stutter by itself, I really, really doubt if SLI is such a good idea for P3D, even when LM (and nVidia) fully supports it... I guess we will have to wait and see. Hopefully the wait won't be too long. (I plan to get that new PC when the 4790K has been released, which may be in a month or two but it can take longer.)

...actually I should say, I need support for more than one GPU, I have another Titan sitting idle. 

 

YES. YES. YES.

 

  

I am starting to get a little skeptical of the philosophy of letting the gpu handle almost everything.

 

I hear ya, but on this last point I'd just say that being GPU limited is always preferable to being CPU limited.

 

Not only are GPU FPS drops more tolerable (assuming we are within VRAM limits), but it is also much easier to simply upgrade GPUs, and/or with future versions, drop in a second (or third ? Hey, one can dream) GPU and see better performance. 

 

Granted, with enough GPU power the bottleneck will move back to the CPU, but we're still better off as the CPU is doing less than it has done historically, and should have more headroom. 

 

They are also still likely working on threading improvements, and that is certainly another way to go, but is much harder to do. The sim is still very reliant on single threaded performance, and with IPC bumps being as underwhelming as they have been these past few gens, the more they push to GPU the better. 

 

Either that or we stand still waiting on intel.

I am starting to get a little skeptical of the philosophy of letting the gpu handle almost everything.

 

At least GPUs are advancing very quickly, both NVIDIA Maxwell and NVIDIA Pascal are going to provide big benefits over the current generation (the former will offer much better efficiency and the latter 5x the bandwidth). On the other hand, CPUs have barely progressed since Sandy Bridge, and at the same time they're being massively crippled by the overhead caused by DirectX.

 

If they keep moving stuff to the GPU, in a few years the results of this decision will show.

I am considering to get an SLI setup in my new computer but when I read 'SLI' I also read 'stutters'. Now I've heard that nVidia is doing great work on getting rid of those stutters but since P3D, like FSX, is a sim that has a rather strong tendency to stutter by itself, I really, really doubt if SLI is such a good idea for P3D, even when LM (and nVidia) fully supports it... I guess we will have to wait and see. Hopefully the wait won't be too long. (I plan to get that new PC when the 4790K has been released, which may be in a month or two but it can take longer.)

 

Both camps got into a big contest over who had the best frame times over the last 1 to 1 1/2 years or so, and as a result, both are able to do MGPU with frame times and frame time variance as good if not better than single GPU now. It was a problem, but is now mostly solved. 

 

~18 months ago, AMD was pretty bad in his regard, but thanks to the great competition we've seen between both sides lately, both brands have greatly improved MGPU performance, not just in raw FPS, but also smooth frame delivery. The GTX 570 was the last "bad" card from nv with frame time performance, and AMD cleaned up everything from the 6-series on up with about a year or so of driver work. 

 

I wouldn't worry about this specific issue much if at all.

I am starting to get a little skeptical of the philosophy of letting the gpu handle almost everything.

 

Your 680 was good for a GPU limited FSX, however it's not so good for P3DV2.x ... you really do want the GPU to be doing as much work as possible - GPU's scale well and perform better.

 

Upgrading your GPU's is easy, remove old GPU, insert new GPU.  CPU not so easy if your CPU choice requires a new motherboard.  Then you have to rebuild and re-install everything (OS, applications, etc.). 

 

GPU's also offer the MOST performance benefits, CPUs don't do shadow calculations and many numerous other calculations as fast nor as efficiently as the GPU.  When building 3D applications you can do "software" emulation (meaning don't use the GPU) to do graphics calculations and processing (you see this process in some of the benchmark software).  If you let the CPU do all the work you go down from 60 fps to 2 fps.  You can also double CPU performance but not realize much benefit in graphics performance.

 

Unlike CPUs, adding another GPU (provide application supports multiple GPUs) you can realize almost double the performance.  P3DV2.x is threading out tasks on all cores (if one elects to use all) yet it makes little to no difference in my frame rates when I go from 4.2Ghz to 4.8Ghz.  I don't even run at 4.8Ghz any more, just no point ... I've dialed it back to 4.2Ghz.

 

Cheers, Rob.

I agree, and most certainly P3Dv2 is future proofed. We will certainly being seeing much more performance down the road.  

I have a GTX 670 and running 3 X 24" Dell monitors

 

and this is my GPU workload. Not sure if I would benefit upgrading.

 

I never though I would say that my next build will like be quad SLI. :lol:

 

 

I wouldn't worry about this specific issue much if at all.

Thanks for the information! I will definitely have a look at SLI now when I start shopping for that new PC!