August 29, 201411 yr There may be other issues contributing to this. If you want to test things with volumizing disabled you can turn off the volume-effect that uses the noise texture by commenting out one line in ShadersHLSL\Cloud.fx. You are looking for this line: I took a quick look at the HLSL code pertaining to this and what it seems to do is overlay a sample of this "new" (new to 2.3) 2D texture file: <drive>:\Program Files (x86)\Lockheed Martin\Prepar3d v2\texture\CloudNoise.dds on top of the cloud texture. The file looks like a random mix of RGB colors, which probably is used to give each part of the cloud a slightly different look, so that the textures don't repeat exactly. One unrelated thing that I just noticed is that ASN alters two of the shader files, swarmcloud.fx and swarmcloud40.fx, but the unmodified P3d 2.3 versions of the two files appear to be identical to those in FSX.
August 29, 201411 yr Moderator "VOLUMIZE " edit dose nothing on my end so removed it - I was not having any issues to begin with so maybe thats why - anyone else finding that this edit does nothing ? I noticed the same Richard. Vic RIG#1 - I9 14900K MSI Pro z790 RTX 5070Ti 40" 4K Monitor 3840x2160
August 29, 201411 yr One unrelated thing that I just noticed is that ASN alters two of the shader files, swarmcloud.fx and swarmcloud40.fx, but the unmodified P3d 2.3 versions of the two files appear to be identical to those in FSX. Oh, great find ... this is important, I need check this out. It may explain a few things why I seem to get much better performance with P3D weather vs. ASN even when the sky is relatively cloud free. Also ASN has a couple of P3D specific features: Repeating overcast texture pattern fix P3D low cloud offset Checking .... please wait ... And I do indeed see SwarmCloud40.fx.beforeASN and SwamCloud.fx.beforeASN ... however I did a file compare between ASN version and P3D version and they are identical. However, if you check the "Repeating overcast texture pattern fix" in ASN the SwarmCloud.fx gets modified ... date time stamp changes, BUT nothing is actually changed in the file - content is identical ... it's as if ASN opens it, reads it, writes the same file back. If ASN were to modify this file, they would have to trigger a shader rebuild ... I don't see any SHADER_CACHE_VERSION entry in my Prepar3d.cfg so that isn't happening. So in other words, ASN is doing something, but it doesn't seem to be changing anything other than date/time stamp for the file. Perhaps the ASN folks could shed some light? Cheers, Rob. EDIT: I do run ASN remotely on a separate PC but it has full access to the directories on my main sim PC. I noticed the same Richard. Yep, same for me, did nothing so I went back to "normal". Cheers, Rob.
August 29, 201411 yr I should have used Google ... The ASN change only affects FSX DX9. Why it creates the new but identical files in p3d is odd.
August 29, 201411 yr But I do know that I'll take a 60% FPS hit when using DXT5 on clouds while flying thru or close to them that I don't suffer when using 32bit. Interesting. And odd. I don't use REX but the HDE v2 clouds. I almost automatically went for the DXT5 files and not the 32bits ones but maybe I should use those instead and see what happens. It's not that I am having real problems with performance near or in clouds but every little bit helps, of course. Anyone else already gave this a try?
August 29, 201411 yr 3. It seems DXT5 clouds have a huge performance penalty also, I'm not exactly sure why, so stick with 32bit clouds Have to report that the opposite happens. Loosing about 8 - 10% FPS with 32bit clouds. That's what we should expect, isn't it? Spirit
August 29, 201411 yr Did you try 80000 Have not tried that, but if it is better I will. Think Rob Ainscough said it really makes little differenece, so may as well up that setting. Curious what resolution your success was at Resolution of? I have textures at 1024x1024 Main screen at 1920 x1080 (ruuning a small second monitor off same card for monitoring apps, ATC window etc) Have reduced cloud textures to 512x512 (REX OD) Using the cfg file setting to reduce cloud shadows to 512 And was this with a payware aircraft? Default and payware GA, I get these favourable results. With the Carenado Sr22 for example (FPS hoog with both G1000 screens on), I still get steady 25fps. With the iFly 737 for P3D, yes, the FPS drops to 17-19 (but no added hit on GPU - so looks like this is a CPU limitation - additional systems processing, not sure why?), but once flying this it was not a spoiler. Some stutters on the ground in turns, but taxi and take-off straight ahead views were smooth. After takeoff, has a brief look to the side before entering solid cloud at 2000ft AGL, and yes in side view there were stutters, but not show stoppers. Once in cloud could not notice these anyhow. Above the solid cloud deck at 9000 ft things were smooth again, and FPS picked up steadily till 25fps at cruise. I can fly this OK if this is really only what happens with solid cloud. Any scenario with better weather is only going to be more fluid. All tests by the way same historical ASN weather for 27th August 2014 10h00Z., at Cape Town (FACT). Reminds me of the one reason why living there was sometimes not so pleasant! Rob Robin Harris
August 29, 201411 yr Its so funny reading though some posts.At one point here at work I thought, cannot wait to get home to do the VOLUMIZE tweak, and I also run rex4 with DXT5, so get them all over to 32 asap.... 60% boost wow. Then someone does that and get a FPS hit by going to 32 from DXT5....lol who knows :rolleyes: :rolleyes: :unsure: :blink: David Murden. MSFS • Fenix A320 • PMDG 737 • MG Honda Jet • 414 / TDS 750Xi • FS-ATC Chatter • FlyingIron Spitfire & ME109G • MG Honda Jet • • Fenix A320 Walkthrough PDF • Flightsim.to • DCS • A10c II • F-16c • F/A-18c • F-14 • (Others in hanger) • Supercarrier • Terrains = • Nevada NTTR • Persian Gulf • Syria • Marianas • • [email protected] All Cores HT ON • 32GB DDR4 3200MHz • RTX 3080 • TM Warthog HOTAS • TM TPR • Corsair Virtuoso XT with Dolby Atmos® • Samsung G7 32" 1440p 240Hz • TrackIR 5 & ProClip •
August 29, 201411 yr It seems to me if your running Titan or higher (if there is one) you will see no difference with this volumize fix as the high end video cards seem to perform great in P3D. So I think the issue is that the majority of people do not run a Titan so this fix does do wonders for those less fortunate video cards. Now I run a 580Ti and am able to maintain 30fps with most sliders at high settings so I'm happy until my old outdated video card croaks for now and the fix just improved the smoothness of my sim. The moral of the story start saving your spare change and buy the $$$ Titan or greater video card. This hobby just a got a lot more expensive. Sean Green
August 29, 201411 yr "VOLUMIZE " edit dose nothing on my end so removed it - I was not having any issues to begin with so maybe thats why - anyone else finding that this edit does nothing ? Well, if you're not having the cloud performance issue, it's expected that the "fix" for it won't do much if anything for you. ^_^ As for me, on a 660ti locked and sustaining 30fps except dropping to 20 near clouds in v2.3 (it did hold 30fps near clouds in v2.2), the VOLUMIZE edit then SHADER rebuild didn't noticeably change anything for me, either.
August 29, 201411 yr Well, I did this sequentially with 2.3 and got back to the performance that I had in 2.2: 1) Disabled HDR - this made a little bit of difference in heavy weather conditions... nothing new. 2) Made the volumetric cloud shader fix - this improved performance almost back to 2.2 levels. 3) Put the REX clouds at 512 and 32 bit. - all three together boosted everything back roughly to 2.2. Would you mind pointing me to the volumetric cloud shader fix? If it's in this thread I've somehow missed it. Thanks.
August 29, 201411 yr Greetings All, Daniel, what resolution are your cloud textures in/from REX4? Somewhere, perhaps the Opus forum, I read that the 512 textures actually look better and run much smoother. It seems to work for me - I did a clean install of 2.3, not an upgrade.. Respectfully, Brian
August 29, 201411 yr Have to report that the opposite happens. Loosing about 8 - 10% FPS with 32bit clouds. That's what we should expect, isn't it? Spirit Curious how you tested.. ie: in a thick cloud, screen paused, note fps or fraps fps and repeat and rinse etc? Or some other method.. It seems to me if your running Titan or higher (if there is one) you will see no difference with this volumize fix as the high end video cards seem to perform great in P3D. So I think the issue is that the majority of people do not run a Titan so this fix does do wonders for those less fortunate video cards. Now I run a 580Ti and am able to maintain 30fps with most sliders at high settings so I'm happy until my old outdated video card croaks for now and the fix just improved the smoothness of my sim. The moral of the story start saving your spare change and buy the $$$ Titan or greater video card. This hobby just a got a lot more expensive. 780ti + 4930k at 4.4ghz here and i saw no change using the fix.. at least in my one test.. i may do another from within a few clouds to see if the paused screen fps/smoothness changes any. So perhaps 780ti and up ? (3gb vram) Asus Strix z790-e; 1000 watt evga SuperNova Plat; 14900k AC_LL 0.55 adp -0.050 253/253/355 CEPoff (CB-1pass 39200 80c, msfs peak 92,avg 60-78c, astrorender 95c,room76F); 64GB(dual 32) cl32 6400 at 6400 xmpII F5-6400J3239G32GX2-TZ5RK, Asus Ryuo III 360mm; Thermaltake v51 Case; Gigabyte 4090 OC; VR-Crystal; Dofreality H6; Astrosite
August 29, 201411 yr Would you mind pointing me to the volumetric cloud shader fix?Go back a page or so in this thread to Arwen's post where she quotes Beau from LM. The quote explains how to make the change.
August 29, 201411 yr Interesting. And odd. I don't use REX but the HDE v2 clouds. I almost automatically went for the DXT5 files and not the 32bits ones but maybe I should use those instead and see what happens. It's not that I am having real problems with performance near or in clouds but every little bit helps, of course. Anyone else already gave this a try? I tried all of the HDE v2 versions. Very little difference in FPS on my system. Still appalling framerates with any weather. Chillblast Core i5 14600KF Liquid Cooled RTX 4070 SUPER 32GB RAM. Internet: 1 Gig Fibre. HoneyComb Throttle & Flight System. UK PPL since 2006 current on PA-28, C-152, C172, Decathlon, C-42 based at EGHP.
Create an account or sign in to comment