Hi Ray,Clouds in general are a FSP hog in FSX. Our clouds are the same size as default clouds when using 256. Also, running at 19 X 12 is stressful. Limit your cloud layers to 5 in ASX and see if that helps.Thanks,Jimhttp://www.hifisim.com/banners/hifi-community-sigbanner.jpghttp://www.hifisim.com/

Hi Jim,So for whatever reason you weren't able to optimise the cloud bitmaps for FSX? That's a first isn't it?Yes, 1920*1200 is high but the card returns similar frame rates at lower resolutions so I'm sure it's not the high res that is causing the issue. It is a powerful card with 1Gb of VRAM available.I'll limit the layer numbers as you suggest but earlier today I experimented with a single layer cumulus and once it was sufficiently thick the fps dropped so I'm not hopeful.It seems there's a lot of things that are less than perfect in FSX. I think many of us will welcome the arrival of FSXI and hopefully a better flying experience.Thanks,

Hi Ray,"optimise the cloud bitmaps for FSX?Optimise clouds are only related to resolution not others things.Clouds in fs9 was 512x512 32 bit as default, Asg for fs9 we provided 512x512, 256x256, 128x128 and 64x64 in 32 bit and dxt3.In fsx default is already in 256x256 dxt5 (similar to dxt3), For fsx, we provided 512x512, 256x256 and 128x128 in both format 32 bit and dxt5. Default fsx is 256x256 dxt5, there was not alot of room to increase fps compare to fs9, simply because FSX fps use many ressource in others area like 2x water and others things like this, this is why Microsoft Aces used 256x256 in dxt5 to increase their performance for others area.In fact many users use 512x512 or even 1024x1024 in 32 bits imagine, 256x256 or 128x128 dxt5 is the most you could use for fsx.ThanksChris Willishttp://www.hifisim.comhttp://www.hifisim.com/banners/hifi-dev-sigbanner.jpg

Hi Jim / Damien,In FS9 using AS6.5 the cloud textures are rendered very efficiently on my system and even in heavy cloud the frame rate isn't affected too much.However, in FSX with ASX/ASG I'm seeing quite a frame rate hit with moderate amounts of cloud. Have you optimised the cloud textures for FSX in the same way you did for FS9?I ask because I feel my graphics card should be able to render clouds more efficiently in FSX. Is it simply a driver issue or are the clouds in FSX designed substantially differently to those in FS9? The cloud texture bitmap is at the default of 256 (from memory).Thanks.

Hi Chris,Thanks for explaining the process. I'll play with water settings and see if it helps with heavy cloud performance.I'm still a little perplexed because owners of nVidia 8800 cards who previously used the ATI 3870X2 are reporting much better performance in heavy cloud.Can you suggest why the 3870X2 might struggle when the 8800 series apparently doesn't?Thanks,

Ray,I share your frustration.I have a system that is similar to yours except I have an ATI HD2900 (512M). By far, clouds have the biggest FPS hit for me. For me, nothing makes it better (e.g. low res textures, less AA). It

Hi Tyrodes,I'm glad I'm not alone. :-) I did quite a bit of research on my components before taking the plunge and the graphics card is the only one where I seem to have fouled up. I say 'seem' because according to Phil's blog my card should perform better than it does. But unless I'm in thick cloud the fps is fine and the quality of the display is excellent. Better than nVidia cards apparently. :(The memory bandwidth is a key area. The 8800Ultra's is 103.7Gb/Sec. The 3870X2's is 115.3Gb/Sec so in theory it should perform faster so why doesn't it? The main area where they differ is memory speed. The 3870X2 uses DDR3 whereas the 8800Ultra (and the 3870) uses DDR4. But that shouldn't kill fps in cloud should it?The memory clock of the 8800Ultra is 1080MHz but the 3870X2'a is 1802MHz so it should be significantly faster if these numbers actually mean anything.Looking at the technical spec of the 3870X2 I can't see any obvious area where it would be so weak rendering cloud textures. I feel very frustrated that these super-dooper cards can perform absolutely brilliantly with one game and quite poorly with another. How on earth are we to decide what's a decent card for FSX when so few reviewers use it in their tests?I look at Phil's blog and the spec of my 3870X2 and I'm still convinced the card is in the top category. Perhaps it is. Maybe it's the drivers that are responsible. This is a relatively new card so maybe the drivers still need tweaking but without FSX being on their list of high-profile games we have little or no chance of getting ATI to look into this problem. I have asked the question on their forum so hopefully they might respond.Your technical knowledge is impressive. Mine is far less so but if choosing a graphics card for FSX is such a lottery maybe Aces should go as far as listing their recommended cards. That might be contentuous but if it's going to save people money it would be welcomed. If the graphics card companies complain that their cards aren't being recommended they know what to do don't they! :-) Actually as there are only two of them it might provide an incentive to extract their digit and start supporting something worthwhile instead of all these silly shoot 'em ups bought by kids!! :-fume PS. Have a read of this review http://www.hexus.net/content/item.php?item=11398&page=2 in which they say "the 3870X2 on paper should have the beating of all single GPU cards". I feel like taking someone to court over this! :-eek

If it makes you feel any better, I would have purchased the same card. I like the image quality associated with ATI. I don

Hi Tyrodes,<>Thanks. At least it shows I still have reasonable judgement.<>That was my first question in this thread. Hi-Fi were able to improve cloud performance by reducing the ridiculously high default cloud bitmaps. The same opportunity wasn't available with FSX but it's disappointing that they are so GPU dependent. And I find it incredibly frustrating that many games perform differently with different GPUs.<

I have just completed a feedback document suggesting the drivers could be faulty with the 3870X2. Hopefully they'll look into it. If anyone else wants to supply feedback on AMD cards go here http://support.ati.com/ics/survey/survey.a...ID=486&type=webI hope Damian and Jim don't mind us discussing this here. :-)

No, not at all as long as it stays on topic!Jimhttp://www.hifisim.com/banners/hifi-community-sigbanner.jpghttp://www.hifisim.com/

Ray,I tested a few things last night. Almost all graphic settings (FSX or ATI driver) did not improve performance in heavy clouds. However, screen resolution had a big impact. At 1680x1050 I would get FPS in the single digits with overcast clouds (e.g. thunderstorm), but at 1280x800 FPS stays above 18. I am not crazy about the lower resolution, but I am going to give it a try for a bit (try ASX again!). I know you said resolution did not matter, so this may not help you.

Hi Tyrodes,Thanks for the update. I'll try some different resolutions tonight and see if it makes any difference but I'm doubtful. That second GPU and its 512Mb probably allow me to run at 1920*1200 without penalty.More later... after the football - Liverpool v Chelsea.

Ray,Have you tried the lastest ATI driver (8.4) with low (2X box) antialiasing set in the ATI control panel?With this combo I get very good cloud performance and the image still looks good (1680x1050 res).I am almost sure this did not work pre 8.4.I suspect something has changed in the driver (I hope). BTW, last night not only did I test resolutions, I also loaded the 8.4 driver at the same time(not a systematic way to test things). This or the now acceptable image at 2x antialiasing may be why I get good FPS in clouds all of a sudden)???Or, it could be something else I tweaked? Either way, I think this cloud issue we have can be fixed without a new card!

Hi Tyrodes,Good detective work! I am running with the 8.4 and Ray Adams ATI Tray Tools - not the awful CCC. Res is 1920*1200.I placed my aircraft at EGLC Rwy 28 pointing towards the City of London. I setup a t/storm layer at 3000ft base and top at 13000ft. Overcast. I took the fps when on the ground, the minimum during a slew up to 13000ft at 100 feet per second and finally the fps at 13000ft. I switched off all settings under traffic so as not to affect the frame rate with moving objects.Firstly, with 4X Box...25 - 16 - 27Next with 2X Box...29 - 25 - 45With 8X Box...13 - 9 - 21App Controlled (no AA in FSX)25 - 16 - 27App Controlled (AA enabled in FSX)24 - 16 - 28I double-checked each of these tests and I'm happy that they're pretty accurate. The AA setting in FSX equates to 4X.I still have the 8.2 drivers so I'll try the same test and see how they differ. I changed no other settings to ensure this was purely a test of AA.More later and of course the real test will be an actual flight but so far it's very encouraging. The fact that fps are hit so hard with higher AA may well be due to a driver bug. The card is powerful enough to cope with anything chucked at it really.Well done!