AI Live Traffic Released for FSX and Prepar3D
Pete Dowson replied to kenz's topic in The Prepar3d ForumI don't think planes you can't see are really causing performance issues -- they aren't being drawn, whichissurely the main load. And the further away an AI aircraft is then the less frequently is it being updated. I think i recall that the update frequency goes right down to something like 4 times per second, or maybe even less. Of course that latter part might not be true for a live injection which is continually updating their positions, but I'd hope it was. PeteAI Live Traffic Released for FSX and Prepar3D
Pete Dowson replied to kenz's topic in The Prepar3d ForumThe default is only 50% as an average starting point. If they are all the same then no preferences are being expressed, so even if they were all 10% or all 90% they wouldn't have any affect. It's the differences which tell FSUIPC what you want it to do. You choose whatever suits you. Set them all to 100 except the one for planned airports -- assuming you load your plan into the Sim. Otherwise FSUIPC doesn't know them. PeteAI Live Traffic Released for FSX and Prepar3D
Pete Dowson replied to kenz's topic in The Prepar3d ForumA question really for my Support Forum, but whilst I'm here: The percentage is the chances of each category being deleted, so you need to increase those for less traffic. Apart from the main traffic limit and the frame rate target, all the others merely control deletion probabilities -- as in fact stated in the documentation (eg User Guide page 53 for FSUIPC5). PeteEnvironment Force Discussion Thread
Pete Dowson replied to vgbaron's topic in The Prepar3d ForumLooks pretty much how it looks here. Surely the darkness depends on the depth the sunlight has to travel through. Smaller (thinner) clous should be lighter, yes. PeteWhat features would you like to see in P3Dv5?
Pete Dowson replied to Jackaroo05's topic in The Prepar3d ForumI expect so. And BMP DDS and other graphics fles, maybe optionally. but not the conreollng part, the BGLs. Yes to the architecture -- splitting taks off for GPUs, differemt threads etc. But source data structure on disk is only read and cached or stored. The conversion to whatever is needed is done when loading. That can be done separately without affecting any main processes. And compressed data is sometimes faster because the loading time savings are more than the decomprression overhead in one of the cores. That's excellent! I'd love to see such a spread. What clock speed is the 9900K running at? On all cores? I tend to only get a decent spread when initial loading or loading a new scenario for a complex airport. Maybe I need to wind up some sttings. Whilst I tend to agree with you, is that not because you are looking at it? If you'd passed it already wouldn't is still be loaded, at least till it ran out of memory? Seems to be your observation shows it cleared it to make room for what is to come, so had to reload it now yo decided to look back. PeteWhat features would you like to see in P3Dv5?
Pete Dowson replied to Jackaroo05's topic in The Prepar3d ForumMy first 'PC' (before they were called that -- it was coined by IBM and they came in late) had 8kb! It was the first Commodore Pet. And I established my software business on a word processor written for it. That was the version with a calculator-style keyboard and a front mounted cassette drive. About a year later a 32kb version with a proper keyboard appeared. That was when things really took off. I left employment (with ICL) when the income from my Word Processor ('WordCraft') exceeded that from ICL. That would be in 1979. Other companies soon joined in the PC market, with IBM actually quite a latecomer. I can't remember, though, when we started talking in Megabytes! PeteFPS Loss- Unlimited vs. Locked FPS- P3dv4.x
Pete Dowson replied to tomcatman24's topic in The Prepar3d ForumOkay. I'll try 0. It did seem to work setting it greater than VS/2, but i'd rather have it off, which it sounds like '0' accomplishes. PeteFPS Loss- Unlimited vs. Locked FPS- P3dv4.x
Pete Dowson replied to tomcatman24's topic in The Prepar3d ForumSo if that is set the limit above is ignored? I can't make it go blank. For now I've set that to some high number, and the Scanline/2 limit is operating. But it isn't clear why there's always both values displayed. PeteFPS Loss- Unlimited vs. Locked FPS- P3dv4.x
Pete Dowson replied to tomcatman24's topic in The Prepar3d ForumWhen using this facility what should be done with the limit value just above. I've tried to clear it but it won't. If it is set lower than the VS/2 is that applied instead? I've only ever used the Frame limit value above before, but now I'm trying 60Hz set on my projectors, so i could try scanline x/2 for 30Hz. PeteWhat features would you like to see in P3Dv5?
Pete Dowson replied to Jackaroo05's topic in The Prepar3d ForumProgramming classes? In 1963, when I started, you taught yourself. And I started with 0's and 1's actually making bootstraps on paper tape with a Unipunch (yes, it's as it sounds -- one hole at a time). That was at Leo Computers (LEO = Lyons (bakery) Electronic Office) in North Acton. Ah, kids today. They don't know what it was like! 😃😉 Pete P.S. Oh, I do feel old ...V4.5 : what has happened to Locked framerate ?
Pete Dowson replied to GSalden's topic in The Prepar3d ForumThat's perhaps something more noticeable with UHD (or do you really have true 4K, which is 4096 x 2160 like my home cinema projector?). Currently, because wide-angle short throw UHD projectors were prohibitively expensive, and the affect on P3D frame rates would have been intolerable, I stayed with 3 x 1080p (5760 x 1080). However, with my new super-performing hardware and cheaper UHD projectors, maybe even 4K ones, I will be tempted in the future. I don't want to go to multiple PCs with Wideview. My set up is 3 windows (overlapping, for blending) on a 210 FOV curved screen. But I have no cockpit or gauges to display. Pete'coreprio' gives me BSODs with P3D4!
Pete Dowson replied to Pete Dowson's topic in The Prepar3d ForumI know why it is BSODing! But I've had this overclock running perfectly for weeks now, ever since Rob Ainscough built it for me in February, and it has done many hours and often with heavy loading, with no BSODs till now (except before I realised I needed to turn on the external cooler at least 10 minutes before daring to turn on the PC! 😉 ). With COREPRIO there must be a lot of extra load being put on something somewhere (NOT registered by Task Manager in which everything looks hunky dory). Anyway, I think I've solved it. I examined the settings in CORPRIO and saw that I had Prepar3D.exe;* set for the processes to be optimised. I hadn't noticed the * when I put the P3D entry in. So I guess it was trying to play with the core affinities for every process in the system (and there are a lot of them; 12 Apps and 57 'background processes'). With that changed to just 'Prepar3D.exe' it seems good again. I'll give it a good workout to be sure. Thanks anyway! 😉 PeteWhat features would you like to see in P3Dv5?
Pete Dowson replied to Jackaroo05's topic in The Prepar3d Forumand I don't know of any BGL compression. BGLs are programs written in 'Bruce's Graphical Language!' devised by FS's originator Bruce Artwick. The language has been much extended to do more and more things over the years, but it is basically a sound design that can do a lot. Many of the new features remove backward compatibility when used, but the original features don't take up time or processing if not used. So removing that doesn't buy you anything as long as you only add sceneries designed FOR P3D4 or P3D5 using the newer techniques. One of the problems with performance in this area now is likely simply because the DEFAULT sceneries have not been revised of re-done, probably since FS2004 let alone FSX. Textures are compressible, and mostly that's already optional. There are utilities to do it using different algorithms. But with compression you need to strike a good balance between loading times and execution times. If disk access lags behind processing speed then it would perhaps be better to keep them compresses. I think a look-ahead algorithm, running in a separate thread (using different cores) would be more productive. I think pretty much all performance and smoothness improvements has to come from making much better use of newer hardware. That may be where a different type of 'loss of backward compatibility' will have the most pronounced effect. You'd force folks to update their PCs to move to P3D5, not so much purchase new versions of all their favourite addons. Pete'coreprio' gives me BSODs with P3D4!
Pete Dowson posted a topic in The Prepar3d ForumI read in a thread here about the system addon 'coreprio', which dynamically adjusts cores for running program threads in order to even out the loadings. As one of those users who sees Core 0 maxed out most of the time I thought I'd try it. It appeared to work well. Task manager showed a good spread of cores being used, with core 0 only occasionally spiking to 100% loading. Today, however, when putting more load onto P3D4 (my normal 'stress' test scenarios at EGLL and EHAM, with up to 200 AI), I started getting Blue Screens of Death stating 'CLOCK_WATCHDOG_TIMEOUT'. These might happen during loading the scenario, or after it had been running a little while, or even twice during P3D loading. I looked at all the changes I'd made recently. I even uninstalled a couple of Windows updates which occurred over the last couple of days. I only thought of CORPRIO last, and, bingo, disabling that seems to have done it. Now I should point out that my system is heavily overclocked (and well cooled though -- 9900k CPU never more than 35C despite being clocked at 5.5GHz with cache at 5.0GHz), so such problems may well not occur in more 'normal' systems. But I thought I'd post this as a warning. It's a pity. I was hoping to get a completely stutter free experiece (I don't get many, but certain there are some little stutters over London and approaching EGLL). But little stutters are better than BSODs! PeteV4.5 : what has happened to Locked framerate ?
Pete Dowson replied to GSalden's topic in The Prepar3d ForumI'd rather be free to select the Projector VS myself. 25Hz doesn't work well on one of the Projectors (misalignments and jitter). 30Hz has been okay, but I currently use them at 60Hz so that there's no clash with my separate control screen* which can only run at 60Hz. Not using VSync allows me to limit the frame rate to whatever I discover gives me the smoothest performance. I'm happy with anything over 19fps providing it is smooth. Testing 25 in RTSS at present. * Before you ask 'why a separate 'control' screen, it is because i can never find the mouse pointer on the curved screen. The software I use (NatVis's tailored version of Immersive Pro) doesn't 'straighten' the Windows desktop, and in any case its settings for P3D4 have blended overlap areas which are difficult to navigate through with the mouse. Folks with normal vision cope better, but because of my severe tunnel vision (from advanced Retinitis Pigmentosa) I even have a job finding the pointer on my office PC's wide screen! Pete
I expect so. And BMP DDS and other graphics fles, maybe optionally. but not the conreollng part, the BGLs. Yes to the architecture -- splitting taks off for GPUs, differemt threads etc. But source data structure on disk is only read and cached or stored. The conversion to whatever is needed is done when loading. That can be done separately without affecting any main processes. And compressed data is sometimes faster because the loading time savings are more than the decomprression overhead in one of the cores. That's excellent! I'd love to see such a spread. What clock speed is the 9900K running at? On all cores? I tend to only get a decent spread when initial loading or loading a new scenario for a complex airport. Maybe I need to wind up some sttings. Whilst I tend to agree with you, is that not because you are looking at it? If you'd passed it already wouldn't is still be loaded, at least till it ran out of memory? Seems to be your observation shows it cleared it to make room for what is to come, so had to reload it now yo decided to look back. Pete
Pete Dowson wrote:The reference I gave does seem to state clearly that the cabin lights are switched by the NAV lights control. I suspect that the amount of control that can be built into any given model is limited.