Clouds kill FPS in 2.3

[theskyisthelimit 48]

Same.

 

I use OpusFSI which doesn't have layers, just certain options. In this case "Auto Cirrus Clouds" which makes pretty brilliant looking blankets of overcast. I wish I had taken screenshots from the test last night.

 

Using default weather in P3D with the special weather situations you can select doesn't seem to look as good, or run as good for whatever reason.

 

I'm not using NI at all, my drivers are 337.88, my shadows are something like 60,000km, high quality shadows, full view distance, full scenery. "selective" texture/mesh settings (not full, but close to full)

 

Titan GTX running at 16x, with PCIe 2.0. CPU is 2600k with 4.7 OC, Memory is GSkill 32gb

 

I'd be curious to know if any OpusFSI (beta) users have this issue?

 

Also what EXACTLY causes the issue, as I may not have even done it right.

 

Curious what resolution you run at.. i think that is also key.. i run at 5760x1080 and fps can dip into the mid teens near a "heavy" city addon coupled with thick weather/rain.  I did find one thing interesting fps wise in general.. i saw no real difference, in fact an increase (on a paused saved flight) of about 2 fps by raising water up to the max (not reflections, but water quality)

[nathanpinard 0]

Ah, well I only run at a standard 1920 x 1080 on an HDTV.

[WarpD 824]

Curious what resolution you run at.. i think that is also key.. i run at 5760x1080 and fps can dip into the mid teens near a "heavy" city addon coupled with thick weather/rain.  I did find one thing interesting fps wise in general.. i saw no real difference, in fact an increase (on a paused saved flight) of about 2 fps by raising water up to the max (not reflections, but water quality)

Based on your chosen screen resolution versus your video card memory... I'd say you're pushing yours to it's physical limit of display capability and anything you do beyond that is going to result in reduced frame rates because it doesn't have enough memory.

 

A single card's memory footprint is expected to provide memory for a single display... which is typically in the 1920x1080 range. To run the resolutions you're at... you truly need multiple cards and monitors to get good performance.

[theskyisthelimit 48]

Based on your chosen screen resolution versus your video card memory... I'd say you're pushing yours to it's physical limit of display capability and anything you do beyond that is going to result in reduced frame rates because it doesn't have enough memory.

 

A single card's memory footprint is expected to provide memory for a single display... which is typically in the 1920x1080 range. To run the resolutions you're at... you truly need multiple cards and monitors to get good performance.

Yea I figure once there is true sli support with prepar3d ill add another 780ti or maybe the oculus rift:)

[DannyR 35]

Same.

 

I use OpusFSI which doesn't have layers, just certain options. In this case "Auto Cirrus Clouds" which makes pretty brilliant looking blankets of overcast. I wish I had taken screenshots from the test last night.

 

Using default weather in P3D with the special weather situations you can select doesn't seem to look as good, or run as good for whatever reason.

 

I'm not using NI at all, my drivers are 337.88, my shadows are something like 60,000km, high quality shadows, full view distance, full scenery. "selective" texture/mesh settings (not full, but close to full)

 

Titan GTX running at 16x, with PCIe 2.0. CPU is 2600k with 4.7 OC, Memory is GSkill 32gb

 

I'd be curious to know if any OpusFSI (beta) users have this issue?

 

Also what EXACTLY causes the issue, as I may not have even done it right.

Well, I am using OPUSFXI as well.  I do see the FPS drop in heavy cloud cover/storms.  I am also using an HD TV for display.  I can minimize the FPS drop by decreasing the cloud draw distance or by changing the clouds to "simple" from detailed.  Of course, using the simple cloud setting does not present a very good clouds cover or image.    My graphics card, a FTW 670 with 4G of memory, is getting hit pretty hard with the FPS drops.

 

Danny

[Alan_A 1,650]

As I just posted over on the LM forums, I ran some more tests tonight and for reasons I can't understand or explain, was able to drop my NI profile without causing anything more than very minimal shimmering on my A2A gauges.  Without NI, I was able to get good performance with only occasional microstutters in heavy cloud, using both FS Global Real Weather and ASE.  I can't explain the differences in gauge rendering between this round and my previous attempt to go without NI.  As for the microstutters, they weren't anything close to being showstoppers, and I suspect I could have gotten rid of them altogether by turning off some of the shadows.  This is by no means the first time my PCs have done unexplainable things.  I wish I could tell you what accounts for the difference but I can't - it's simple submitted for your consideration.  I'll keep exploring and I'll let you know what, if anything, happens differently next time.

[FLAPS15 8]

The one thing that would really help us all and LM is Nvidia developed a profile for P3D.

 

Not saying it would solve anything but it might assist on areas, most of us nowadays have pretty good systems and can FSX very well, P3d is far in advance with graphics shadows etc, however it should not be a case of tweaking and turning things down or off to make it run.

 

Micro stutters in the scenery are a no no and should not happen not sure what has happened between the last version and the current, I noted and informed the LM that the Aero setting does have an impact, have it on you get micro stutters every few seconds turn it off micros stutters gone but very bad screen tearing back......VSynch issue? Again NVidia?

 

I do know that LM takes this seriously who wants a product that the targeted audience is not happy with it is not in their interest, it is all a balancing act and sometimes I just wonder if patches upgrades are forced out to quickly before very through Beta testing.

 

We all want this to work...I am still on the 60 day buy or return and I like what I see but if the performance does not improve I will return to the old war horse FSX and stick with it because it works it is smooth and I throw any amount of cloud at it with minimal impact.

 

If woks in FSX there must be a way forward for p3d come LM work your magic!

 

Cheers all

[WarpD 824]

Some of you seem to think that if there's something FSX displays that Prepar3D should do the same.

 

They do not, and please... really... this is significant in the entire discussion... they do not use the same rendering engine. At all.

 

FSX can not do antialiasing in windowed mode worth beans. In full screen it does ok, but requires NI settings to do a great job.

 

Prepar3D has decent antialiasing in windowed mode, without the need of NI or anything like it.

 

FSX can not do HDR lighting. It takes a hacked version of the DirectX DLLs to get that to happen as well as tweaking of settings for it to get it to look 'right' (something totally subjective).

 

Prepar3D has HDR built into it and it's fairly decent without any tweaking whatsoever.

 

I could go on and on... but, bottom line.. pointing to FSX and saying that Prepar3D should be just like that... is asking for the new rendering system to be ripped out and replaced by the original FSX rendering system.

 

There are things that DirectX can't do. Doesn't matter who's software is using it. There are things that the core video drivers can't do... doesn't matter who's software is using it.

[spirit_66 2]

 

 

Some of you seem to think that if there's something FSX displays that Prepar3D should do the same.

In this case it wouldn't be a need to move to P3D v2, isn't it? One should accept P3D v2 as the new FS which is under a permanent developing and improving situation. If somebody prefer the older stuff so be it. There are many guys out there which like their FS9 so much. The main point is to be satisfied which what one get. Whining and ranting doesn't get us further.

Spirit

[geolpilot 88]

Well taking the suggestions from many contributors here I have got my P3D2.3 running smoothly even with heavy clouds and rain. The settings I am using are;

 

Shadow Quality - Medium

Cloud Coverage Density - Medium

Max cloud draw distance - P3D 80nm

Max cloud draw diatnce ASN - 80nm

Min cloud draw distance ASN - 60nm

ASN - 3 layers of cloud.

Cloud Shadow Cast Distance - 60000m

 

and now the critical settings;

 

Scrapped my NI P3D profile - No NI Transparency Supersampling. Yes a slight degradation in sharpness of lines and some minor shimmering compared to using 4xSGSS. But as soon as I use SGSS I get a real FPS hit with clouds (something I feel a NI profile for P3D from Nvidia should address, though LM have indicated they have some plans of their own to assist further with fps impact of clouds in 2.4).

 

And now for the one setting that most will not like - especially the frame rate watchers! I use FPS limited in P3D at 25fps. Sure I can run unlimited, and with low cloud cover see FPS hitting 60-90 at times. That drops to 20-35 if I have heavy cloud cover. Watching the load on my GPU when this happens, I see it tanks at 99%, the stutters start and FPS drops.

 

I found a sweet spot where my FPS remains steady at 25fps even in heavy cloud, and GPU usage never exceeds 75%, provided I limit FPS at 25. The result is smooth, stutter-free flight in all situations (cloud-free and heavy cloud).

 

So I for one am not going to be bull-headed anymore, and insist on high FPS settings, simply because  that is THE measure of performance (for many it seems), and my graphics card "should" be able to get better than 25fps. It clearly cannot, and the 99% usage at FPS above 25 is telling me that quite clearly. On the other hand, by limiting FPS to 25 I only notice smoothness, and less load, lower temps on my GPU, and to me that is what counts as performance.

 

Flying low and smooth under a heavy cloud deck, with the light and cloud shadow play on the scenery, is just awesome. So for me its off with the FPS counter, and enjoy the flying.

 

Rob

[spirit_66 2]

 

 

Cloud Shadow Cast Distance - 60000m

Did you try 80000 because it's said that this is the best for them related to FPS?

Spirit

[theskyisthelimit 48]

Well taking the suggestions from many contributors here I have got my P3D2.3 running smoothly even with heavy clouds and rain. The settings I am using are;

Curious what resolution your success was at

Edited August 28, 2014 by n4gix
Removed excessive quote. Again!

[FLAPS15 8]

And was this with a payware aircraft? just try the AEX and a payware airport plus clouds and AI...at reasonable setting... if you could let us know then what the smoothness is like it would be helpful.

 

cheers

[Arwen 289]

To you guys here who state that they have v.2.3 running totally without stutters . . . if you're running a NVidia GPU, could you please do me a favor?

 

Try using Shadowplay to record a short flight in heavy clouds (when micro stutters are the most noticeable), while looking out the side window, or in spot view, lined up with the wings.

 

I would be VERY surprised if you do not notice micro stutters . . . even when your FPS might remain fairly decent and consistent. 

[Guest]

 

To you guys here who state that they have v.2.3 running totally without stutters . . . if you're running a NVidia GPU, could you please do me a favor?

 

Try using Shadowplay to record a short flight in heavy clouds (when micro stutters are the most noticeable), while looking out the side window, or in spot view, lined up with the wings.

 

I would be VERY surprised if you do not notice micro stutters . . . even when your FPS might remain fairly decent and consistent. 

 

 

WarpD posted something sometime ago about those 'stutters' you see while looking out of the side window and according to him and a few other experts those are not actual official stutters. A real stutter is when the entire screen halts for a split second. What you see out of the side window is specific objects passing by a bit jerky while the rest of the screen is still smooth. But that aren't stutters.

 

Anyway (maybe he can chime in and explain it once again) there is a clear miscommunication in this regard: the people who claim to have no stutters and smooth flight usually DO have that jerky passing by of objects when they look sideways but they don't call it stutters because officially they aren't. 

 

I think this different point of view has caused a lot of miscommunication on flightsim forums... So making a video isn't needed, I think: you will indeed see things 'stuttering' out of the side window but officially those aren't stutters and there is nothing you can do about it for now (afaik).