Nvidia 352.86 ...!

[dgraham1284 92]

i'm using rex 4 and saw streaky lines at altitude. Didn't have them before.

Lightning or lighting? There was lightning in the weather when I saw them.

 

Lightning.  When using REX textures, the whole package, when there is a storm. Forget it. I just shut it down.  Perhaps they are the issue? 

 

 

I have REX 4 installed, but I only use the soft clouds aspect of it.

 

 

I will remove everything but the clouds and work my way from there.

[WebMaximus 372]

Many thanks Paul!

 

Will be very interested to hear about your experience once you're sure you have SLI properly working in P3D because what you say about running 4xSGSS and how that will result in poor performance sounds familiar to what I've been seeing using my single Titan card so either

 

- SLI is not working for you

- SLI won't give the needed boost required to use 4xSGSS

 

I hope it's the first one and that you will be able to run 4xSGSS without any issues once you have SLI up and running in P3D.

 

Still waiting for a definite confirmation on whether the newly released driver is the SLI driver we've been waiting for.

[birdmanmike 6]

Paul, re switching SLI bar, see my earlier reply intended for you page  8  . . .  :rolleyes:

Else, with my earlier 50-80 in the clag I was, and am, using SGSS x8 to stop barber poles at Manchester, with very little impact   Was OK at 8 without SLI as well - following advice from Steve to try it

 

(I can't get quotes working again - but have sorted the BSODs on SLI start . . .)

[mikea76 184]

It's not a "placeholder" there is no such thing in a driver.  But the proof is simple ... when you enabled SLI in NVCP go over to the Manage 3D application settings and you will see SLI selected by default.  If this wasn't "the" driver, then it would NOT default to SLI.  Since it does, this IS the driver.  Now that doesn't mean future optimization will not happen on both sides (LM/nVidia) but like I said, I'm now 99.9% certain this IS the driver.

 

As for the performance difference ... well ... I'm still testing that (will provide some data later).  Testing really needs to be done against a BASE v2.5 12945 as some add-ons cause stutters regardless because of how they design specific elements (i.e. runway lights).

 

I'm not seeing any of the visual issues in SLI mode like others have reported, all visually looks good.

 

The Titan X's do overclock well "as is".

 

Before OC (Ultra 4K test)

 

 

AFTER OC:

 

 

Modest OC settings with no Voltage changes:

 

 

Energy used for my entire PC (including 4K monitor) is about 740-760 Watts under full load when benchmarking.

 

Cheers, Rob.

 

 

I wonder how much your monitor uses for power.   I have a AX750  power supply in my current system that I would like to reuse. 

[WebMaximus 372]

Paul, re switching SLI bar, see my earlier reply intended for you page 8 . . . :rolleyes:

Else, with my earlier 50-80 in the clag I was, and am, using SGSS x8 to stop barber poles at Manchester, with very little impact

Not sure what 50-80 means but I take it you're using an SLI setup and what about flying through dense clouds with SGSS x 8?

[Guest]

 

 

I wonder how much your monitor uses for power.

 

Oops, sorry that is load WITHOUT monitor ... I have two APC, one is dedicated to the PC unit, monitor operates on my other APC ... so the 740-760 watt is without monitor.  PSU I'm using is an EVGA 1200 P2.

 

Cheers, Rob.

[vgbaron 3,681]

Don't get wrapped up ion the green bar guys. If you see the green bar while P3D is running, all it means is that you have SLI *selected* NOT that it's working. The bar must move according to the load. It may work fine without P3D running and not work when P3D is running.

 

The best way to be sure you have a functioning SLI in P3D is to use a product like EVGA PrecisionX or MSI Afterburner to graph the GPU usage - you will get two lines, one for each GPU - when running P3D you should see the lines moving up and down in unison according to the load. If you see one line up at 50% and the other at 10% your SLI is NOT working. The load should be split close to even across the cards.

 

Vic

[birdmanmike 6]

I have no idea why I get so good results - but that earlier post re my results was 50-80 FPS and was in and out of dense cloud (clag) via Opus and REX textures, the 50 being the in. I don't have reflections or shadows set, no HDR and low AI. Bathy off.  No AF mask or HT, CPU with a 4.8 overclock. As earlier, 2x670 Windforce with factory OC.  The 50-80 is about 10% higher than earlier driver without SLI.  My settings are otherwise an eclectic mixture of Steve's and Rob's.  Airplane in test was a lowly stock Mooney.  After 2 hours still around 1.2Gb of available VAS. Haven't had an OOM since this driver and SLI, but am not using complicated airplane and EGCC add-on is toughest airport.  60 or 120 cm photo. All CPU cores running at around 60 C in flight

As a point of interest for all, I find it hard to understand why 2x970 orTitan or whatever should bring such marvellous improvements as people seem to be expecting - even one should normally do for P3D surely unless you've got a massive 4k monitor  or two, over a big add-on airport and a PMDG 777?  I mean, P3D is way better than FSX where I get 30-55 FPS in similar test, obviously no SLI but the o/c.  Back in the day, nVidia said a 480? would handle anything FSX could throw at it, it was all CPU.  Though P3D appreciably more on to GPU, it's still heavy on CPU use.  Either way nowhere near as hungry as some of the latest games using Afterburner.

I can try for some screenshots but not good at it  :( and still have to learn Shadowplay

 

PS, my green bar moves all right - gets bigger in the cloud . . . both GPU's show (and warm up) but even maxing don't tax the 6 GPU fans except under Afterburner load, not from P3D

[alulekk 3]

@birdman:

 

You are getting good results because you have no reflections, no shadows, no HDR, low AI, no Bathy and are not running 4k.

[b200 1]

I ran the new 352.86 driver in win8.1_64bit, P3Dv2.5, ActiveskyNext, PMDG 737NGX_P3D on a GTX690/i7-3930K_CPU and I went from 20-30fps to 6fps, (all in admin mode)

 

I had to roll the driver back and all working fine again now, has anybody else had a frame rate hit with the new update?

[Guest]

I had to roll the driver back and all working fine again now, has anybody else had a frame rate hit with the new update?

 

You might want to check your Event Viewer and look for any errors specific to nVidia ... my first attempt at the driver install had some issues.  I checked the Event Viewer and found the red X error reporting nvtray failure.  re-install solved it.

 

EDIT: assume you used DDU to remove older driver and prep for new driver install?

 

Cheers, Rob.

[b200 1]

Thanks for the reply,

 

what is DDU?,

 

i just used the nvidia Geforce experience updater app to update, I would assume that would do everything.

 

I will try again and have a look at the event viewer and i will google around as to what DDU does.

[Guest]

I posted my Fraps data results here if anyone is interested: http://forum.avsim.net/topic/468975-p3d-v25-sli-test-results-up-33/ - my data shows a 33% increase in SLI vs. non-SLI.

 

 

 

what is DDU?

 

DDU is here - it does a clean video driver wipe in SAFE mode: http://www.guru3d.com/files-details/display-driver-uninstaller-download.html  - it boots to safe mode, so if you're concerned about working in SAFE mode then you may NOT want to use this. 

 

nVidia Geforce updater sometimes has "issues" ... I've ran into mixed/match drivers ... especially if one goes back to older drivers from newer drivers.

 

Cheers, Rob.

[dgraham1284 92]

Well, reinstalled. Those crazy lights are gone. 

Ok, I think I narrowed it down. It came back.  FTX Global. Something with their lights. I disabled it in the scenery cfg and it went away.

[b200 1]

I ran DDU and reinstalled 352.86, all seemed fine.

 

I loaded up P3Dv2.5 with the pmdg 737ngx,

 

I opened the FMC into 2D window mode and undocked it to move it across to my other monitor and BANG fps 30 down to 6 and unflyable, so i closed the windowed FMC and everything came right again.

 

Has anybody else noticed this behaviour, it didnt happen in the previous driver.