Pete Dowson

Sorry, you are right! I just installed the very latest P3D update (HF2) on the PC where I tested the Traffic Limiter, and you are correct. The deletion process isn't working. FSUIPC is assuming the call for deletion happened, so it doesn't need to wait for a Simconnect message, and counts it so. But it isn't actually deleting! That's weird. I'm looking into it now. [LATER] Found it! Can you confirm you are using FSUIPC version 4.959m? Because a part of the debugging code I added since 4.959 is making FSUIPC actually by-pass the call to delete aircraft! Duh. Stupid. It was part of the process I was using to narrow down the reasons for the crashes on exit folks were getting, and it should have been removed when I fixed it. I'm getting too old, missing things like this. :-( One trouble is that with something with as many functions as FSUIPC, it becomes impossible to re-test everything after making each change. You only tend to re-test the areas you think you've changed! With the investigation into the crash on exit problems over 10 days hard work there were lots of little changes made to narrow things down. I've now re-checked all those I made notes of or remembered -- I hope I've not missed any others! I'll post an interim update (4.959n) on my website later today. Version 4.960 full installer should be releasable early next week, when I've managed to do the few updates to all the documents. Pete

Strange, it works perfectly here, in P3D. The number isn't supplied by P3D as such, it is only obtained by FSUIPC conting the AI Traffic "Added" messages received from SimConnect, and subtracting those notified in the the "Removed" messages . I thought AITrafficManager deleted traffic using the IDs and via FSUIPC? FSUIPC can only know about traffic for which it receives traffic messages from SimConnect, and it cannot delete traffic it doesn't know about, so I find this all very mystifying. Does AITrafficManager work the same now without FSUIPC installed? The count FSUIPC comes up with always matches that shown by P3D's own Traffic Explorer. -- bar one, as Traffic Explorer includes the User Aircraft which FSUIPC doesn't. Check with that please. It's part of the Tools menu created by the TrafficToolbox.dll from the P3D SDK. Would be better for me and my users. Pete

.. traff) Sorry, but isn't that rather obvious by inspecting the others? The .. just joints thingsup into one string for the display. Pete

It is similar, but uses the "wnd" Lua library, which is only available in WideClient: w = wnd.open("VAS Monitor", WND_FIXED, 100,100,425,185) wnd.backcol(w, 0x000) wnd.textcol(w, 0x6c0) wnd.font(w, WND_ARIAL,-4) -- Update the display at 500 msec intervals (see event at end) function mytimer(time) wnd.clear(w) fr = ipc.readUW(0x0274) if fr ~= 0 then fr = 32768/fr end fr = math.floor((fr * 10) + 0.5)/10 mem = ipc.readUD(0x024C) memmb = math.floor((mem/1024) + 0.5) mem2 = ipc.readUD(0x0290) memmb2 = math.floor((mem2/1024) + 0.5) traff = ipc.readUD(0x025C) wnd.text(w, "Frame Rate " .. fr .. " fps\nVAS free " .. memmb .. " Mb\nMax block " .. memmb2 .. " Mb\nTraffic " .. traff) end -- Adjust timing to taste: 500 = 1/2 second event.timer(500, "mytimer") Pete

Yes. The count is provided in FSUIPC offset 025C, as a 32-bit (4-byte or "UD") integer. If you already use a VAS Monitor display, like the one around as a Lua plug-in, you could add a line to show this. Here's a Lua plug-in which will show frame rate, VAS total free, VAS max block size and the AI traffic count, all in a Lua display window in FS/P3D. I'' add this to the Lua examples package installed by FSUIPC on the next full release (which will be soon). -- Update the display at 500 msec intervals (see event at end) function mytimer(time) fr = ipc.readUW(0x0274) if fr ~= 0 then fr = 32768/fr end fr = math.floor((fr * 10) + 0.5)/10 mem = ipc.readUD(0x024C) memmb = math.floor((mem/1024) + 0.5) maxmem = ipc.readUD(0x0290) maxmemmb = math.floor((maxmem/1024) + 0.5) traff = ipc.readUD(0x025C) ipc.display("Frame Rate " .. fr .. " fps\nVAS free " .. memmb .. " Mb\nMax block free " .. maxmemmb .. " Mb\nTraffic " .. traff) end -- Adjust timing to taste: 500 = 1/2 second event.timer(500, "mytimer") I also have one which shows these values in a separate Window , but it will only run under WideClient (I use it on a Network PC). Pete

I'm using ProATC/X all the time and Manchester is my 'home' airport so I fly from there a lot, and I have never yet come across any such discrepancy. Are you using an add-on Manchester, or the default? The latter may indeed have incorrect frequencies. I don't know. The difference between what ProATC/X says it wants and what it really wants seems very odd. I don't know where it would get two different values for the same facility from. Your best bet is certainly to pose this problem on the ProATC/X support forum. It is very good and responsive these days. Just make sure you attach a log from a flight with this problem, and sign your real name. Pete

No, no Orbx active. My vectors are UTX. My FPS is okay, it's only the pauses which are a problem, and only in P3D. Pete Thanks. I'll see whether the pauses near EGLC disappear if I disable UK2000 Gatwick. I don't know what the performance it like flying into Gatwick is tese days. Not gone there for a while! Pete

It's ok here with little traffic, the differences come with realistic amounts of traffic for the area. I don't really have lower frame rates at or near EGLC -- they are higher (in fact, at my limit of 30) than at EGLL. And I certainly don't use London X, at all, let alone at max. The frame rates aren't the problem, it's the pauses -- one of which must be the best part of a second! I was using the original FSX version (V1.0) of EGLC in both FSX and P3D, but I've since installed the one labelled for P3D (V2.0) and it makes no difference. And it isn't the traffic -- there's very little there in any case, and the pauses are still there without any traffic enabled. I'm also happier with the sharper (better anti-aliassed) image I get in FSX-SE. Although it can be more flickery than P3D, at least straight edges are straight, etc. This is with scenery and traffic -- I don't have any on-screen cockpit parts. Pete

I've got both Aerosoft and the new UK2000 Manchester, so I'll be comparing them soon, when I get some time. P3D is a very different animal with some scenery building methods. Gary's methods for UK2000 have changed recently to try to deal with P3D nuances better, but certainly some of them are dreadful performers in P3D compared to FSX. I'm thinking particularly of EGLL (Heathrow) and EGLC (London City). On my system, where P3D is an excellent performer on the whole, UK2000 EGLL is really stuttery and slow compared to Aerosoft's EGLL with half the number of AI. So for EGLL I still to the Aerosoft version (which also has the Concorde near runway 27L threshold where it is in reality -- it is missing on the UK2000 version). EGLC is not just stuttery on my system -- approaching it or flying near it (eg on approach to EGLL runway 27L or 27R) there are actual pauses! This alone has made me return to FSX-SE as my main cockpit sim. (I always keep both sims maintained and flyable for my cockpit in case P3D one day is better). This may be partly down to my desire for realistic traffic levels, for which I use both UT2 and MT6. Even so, with no AI Traffic at all, UK2000 EGLC still causes stutters/pauses. With each P3D update I try again, but so far, though P3D3.4 looked hopeful, it still isn't good enough. Pete Without SGSS my P3D screen is horrible, the anti-alias doesn't work properly and I get too many flickery objects. I can only go up to 1080P resolution max (Projection), so are you on 4k or, if not, how do you make a lesser resolution look good? Pete

Actually version 2 of UK2000 EGCCextreme is just released, and at half price for version 1 owners, so I'd recommend that you update! Pete

Until at least a record of the details of an added aircraft are available, it cannot be detected or deleted until it is more or less present, with ID and Location. And in any case none of the management programs or FSUIPC are merely deleting the latest ones to be added, they either delete the furthest, or use some other algorithm. FSUIPC has three criteria, with a randomised choice system, each one having a user-settable likelihood. Recovery of VAS would occur with correct deletion by the allocating functions in P3D freeing the memory. Freed memory will be returned to the pool where it may be a fragmented loner, or concatenated with an adjoining free block to make it larger. Incidentally, the next incremental update for FSUIPC provides not only the total free VAS but also the size of the largest free block within it. It is surprising how close these become over time, though initially they are quite different. Windows automatically joins adjacent free blocks to make larger ones each time any new block is requested. There's apparently no function to defragment lone blocks, but as I said, this doesn't appear to be a big problem. Pete

Okay, thanks. I'll take a look. What should it be? (Yours is "Level 5"). Pete

What's "LLC" and where do I find it? The 1.47 peak was only very occasional and momentary. Mostly 1.43/1.44. I think AISuite3 set 1.44 as max. Pete

I've now run RealBench, and get these figures: Image Editing 235011 Encoding 189299 Open CL 83627 Heavy Multitasking 180064 System Score 155274 17 6700K @ 4.7 GHz Corsair 3200 MHz GTX 1080 With Benchmark, are higher numbers better or worse? Under RealBench, using RealTemp, my temperatures were MAX 80/76/71/73 (on the 4 cores), but it only hit highs like 80 very briefly now and then. It doesn't show average but I would say it was about 74 on Core 0. The top voltage I saw was 1.43/1.44 mostly, occasionally hitting 1.49 briefly. Pete

Thanks! I'll probably try this in the New Year. Pete

Would I be better off overclocking the processors more, above my current 4.7 to your illustrated 4.9, or replacing the Crucial memory with, say, GSkill TridentZ 3600MHz rated as 15-15-15-35? I can get the latter easily enough from Amazon. From several adviser's and results I see that often the memory is more likely to be the bottleneck, once you have a good processor speed (like 4.7 GHz?). I've got a copy of the BIOS setting screens you posted, so thank you. Regards Pete

I had a good look at your 4 results. It seems that the differences are really all in the individual memory, as follws: 54.3 GB/s, 34.9nS with Membus 1800, ratio 54.3, 13-13-13-28-CR1 51.9 GB/s, 36,7nS with Membus 1800, ratio 54.3, 14-14-14-55-CR1 46.9 GB/s, 39.2nS with Membus 1600, ratio 48.3, 14-14-14-35-CR1 45/0 GB/s, 42.4nS with Membus 1600, ratio 48.3, 20-16-16-35-CR2 I assume this was all with the same memory sticks. Was the last one above (top right in the picture) the "mormal" or default for them? What can't be seen in the pictures is the make of memory, which might be important for overclocking, and the assorted voltages which affect the memory and which I assume had to be increased to get the better performance. Many of the things I've read recently seem to be saying that overall better performance arises from better memory performance even if it means reducing core clock speeds to do this without damaging the memory from too high a voltage. Can anyone here confirm that? Should I be looking for an idiot's guide to overclocking memory on a Z170 motherboard? Ah, I needed that too! ;-) I don't remember at present and it's doing some other tests, in Prepar3D, for me at present. I'll check temperatures and voltage later today. I think the max temp was around 72. Voltage wasn't that high. Both seem lower than the last time I got to 4.7 -- the time which gave me blue screens. The cooling is air except for the Corsair H100 water cooler on the processor. I once had a system built with full circulating water cooling (in fact it is still in use as my delevlopment and testing PC in my office, the one I'm typing on now, but not overclocked any more). I decided I didn't like it, too many pipes in the way and very difficult to change components. Good. Time perhaps to look elsewhere for my Prepar3D stutters and hesitations. Must be one of the other processes. Elmination is the order of the day today ... and then trying to decide what to do about it once the culprit is identified! I already prune them as much as I can, and move much onto other PCs (there are 7 involved in my cockpit operations, though only 4 heavily). Thanks Pete

Hmm. It looks interesting though rather technical. Not sure I'll understand the things that seems to display. there's a lot of them! I was really only looking for 1 figure for bandwidth and one for latency, and then only because it was suggested I should check them, and when I did and showed them it was further suggested that I need to try to do something about improving them before messing more with Prepar3D to get rid of the stutters and hesitations. Unfortunately, whilst various things like more overclocking and better memory settings have improved the original figures, this seems not enough. I looked at that, but thought I'd try the utomtic method again first. This time, because I think of different choices made before starting it, it seems to have done very well -- and even the temperatures seem fine. I can't manage 4.8 like you, but 4.7 was more than before. The only thing it did wrong was reset the XMP option so my memory was back to 2933 instead of 3200 where it should be. I've rectified that anyway now. Yes, thanks. At least it is interesting. I still would love to know what bandwidth and latency figures you managed to get, though for comparison I suppose they'd need to be from the same tool -- i.e. Sandra or AIDA64, or the one you just pointed me to if you also tel me which figures I should look at or what parameters I should use. Regards Pete Thank you! I see you have the full version of AIDA64. Might be useful too. Gives me more to think about! ... Regards Pete

Are your results better than mine? If so, how did you do it, please? Ah, but I'm a programmer, not a hardware expert. I used to understand the hardware before it got so complicated and full of new terms I don't know. Actually I just reinstalled it and re-ran it, but changing the strting settings a bit as another poster mentioned above. It went quite well, an it's given my what looks so far to be a stable 4.7GHz. I had to go back into the BIOS and reenable XMP mode again to get the memory speed correct, but apart from that I'm quite happy with it -- except for that non-removeable icon it sticks on the screen! I reran the benchmarks and got further improved results. Still not sure whether that's it though, r whether I've still got something wrong. The supplier is supposed to do that! I think they did. Regards Pete P.S New benchmark results (at 4.7GHz and 3200 MHz RAM): Sandra memory benchmark multi-threaded: 37.8 GB/s memory benchmark single-threaded: 30.2 GB/s memory latency: 18.8 nS AIDA64 memory benchmark (read): 44.4 GB/s latency 47.7 nS For latency they surely must be measuring different things?

Ah, thanks for that. My mobo is the ASUS MAXIMUS VIII HERO, and, yes, I did install and use AiSuite 3. I let it get as far as it wanted -- eventually it fails and reverts. I think that was at 4.8. So it left it at 4.7. it was only later, when running FSX or P3D that I got blue screens occasionally, and reverted to 4.5. (I could have gone to 4.6, but I was playing safe -- anything to avoid blue screens! I uninstalled AISuite because I couldn't find any other way of getting it's damned fax coontrol selection graphic off the screen -- even stopping its services loading didn't seem to help! But it is easy enough to reinstall and let it try again. Hmm. That's interesting. I might try 4.6 that way before installing AIsuite again. Ah, good. That's a relief then. I might try that too -- first perhaps, after your last statement. Thank you very much! That's a very helpful post! ;-) To all posters: After all the reports and responses above, for which thank you all very much, I have discovered a few more things. I found that my Sisoft Sandra was woefully out of date, so I downloaded the latest version and reran the tests. The results are a bit better: Memory Latency : 19.0ns Bandwidth Single-threaded, Aggregate Memory Performance : 30.18GB/s Bandwidth Multi-threaded, Aggregate Memory Performance : 38GB/s I also ran AIDA64 on the recommendation in one of the posts above. That’s different again. It seems to be giving 44.3GB/s for bandwidth (reading only, but better), but 48.1nS for latency. Now I'm not sure which is okay or whether I've still got a memory bottleneck! Pete

Thanks. I'll try that. I do find Sandra quite difficult to make sense of at times. Hmm. Maybe I'm reading Sandra wrong. I didn't even realise there were multiple choices in each category. I just selected the benchmark from the first screen then pressed to icon looking like two arrows in a circle to make it do it. Sorry -- typo here (n and m are next to each other!). I did have "nS" in my emails. I still get stutters in P3D with traffic off. Maybe not as bad or as often. But still there. It looks like I need more info yet to improve things. Regards Pete Both suppliers I've now had 6700K systems from said that 4.7 is achievabl with a smallish sample of the chips they get in. 4.6 is easy, and I tried 4.7 but got some blue screens. If you are using n ASUS Z170 mobo, could you possibly share your BIOS settings? I don't know idf you can save then to a file, write them all down, or perhaps even take pictures. I realise there's a huge number of menus and submenus, selections and subselections. Currently I've no idea what might affect things though. Regards Pete

Found it. The menu for it was set to "Manual". I changed it to XMP, and the correct memory details appeared in the selection below., the one now labelled XMP. With that it manages an increase in memory bandwidth from 28Gb/s to 29.4 Gb/s, and the latency decreases from 63.7nS to 60.4nS. Not really much different but an improvement -- I think the manual settings they did must have been fairly close. And it manages 4.6 okay, and I might go back to that.. It has been at 4.7.but I had a couple of blue screens with those settings, so I toned it down. 4.5 is okay, and a lot faster than my previous PC (3yrs old) at 4.5. Thanks for that pic, it helped me find XMP. In fact it got to 4.7 using their Windows-based app, with does it for you. But I ran with 4.7 for a little while, but as I say above I got a couple of blue screen crashes so went down a notch. Actually two notches to be safe. No blue screens at all since then (2 weeks). I think the Win8/10 setting could be a problem. It enables something called secure boot, I think, which prevents much of the software I need to run my cockpit from running. and I use Windows 7. This PC is the second build with the same hardware, but from a different company. The first only supports Windows 10 and installed that which killed my cockpit setup. I tried to change to Win7 on it but had problems which they were unable/unwilling to help with, so I got my money back and got it from a source who installs Win7 or Win10, your choice. There was just the one XMP profile, but it reads correctly for my memory. The Auto overclock utility bumps the multiplier -- it's at 45 at present on all 4 cores. I think it starts increasing the base clock too if you let it carry on, but then it soon reaches a point where it resets because it hangs or crashes the machine. It seems to adjust the voltage to suit too. And, yes, CPU-Z shows 1600, with 16/18/18//36. I'll check against TomsHardware reports, but the performance, in terms of things like frame rates, is fine, much better than my previous system. It's the stutters I get in Prepar3D which I'm trying to get rid of, and Rob sems to think this is due to memory bandwidth and latency. Still not sure how to solve that, having tried pretty much all combination of settings in P3D itself. Thanks everyone. Still seem to need something ... Pete

No, other than what is reported by by the BIOS optimising utility which came with the mtherboard, an Asus. I'll check again in the morning but I think that showed it set at just under 3200. What's "XMP profile"?. It is labelled as such in the BIOS? Pete

For FSX and P3D, CPU clock speed counts for a lot. A 12.5% improvement allows me to have denser scenery, more AI traffic, the lot. It's a 12.5% improvement in results. FS is virtually completely CPU bound. The overclocking isn't the problem, and it's done by the mobo's own utility. The PC is perfectly reliable and no sign of overheating at that speed. Pete

I'm a bit concerned. I have purchased (at relatively great expense) a new PC which I hoped would be the last upgrade I'd need for a while. It has a Z170 motherboard with a 6700K processor overclocked by 12.5% to 4.4GHz, matched with 2 x 8Gb Corsair 3200MHz DDR4 memory sticks. It was supposedly setup and optimised for me by the PC builders who supplied it, but I am being led to believe it might be considerably under-performing. The measures used to judge with were 1. The memory bandwidth, as measured by SiSoft Sandra's benchmark. I got 28 GB/s but have been told I should get about 45 Gb/s. 2. The memory latency, again according to Sandra benchmarking. I got 64nS, which I have been told is rather high. Is there anyone here who can help me work out what EFI BIOS settings I should be setting in order to improve matters? THis modern PC technology has me baffled. Perhaps someone who gets better results can tell me their settings in the BIOS so I can compare? It would need to be a Z170 motherboard of course, but I think the memory doesn't need to be exactly the same, though it would help. Thanks, Pete