Padishar

Does everyone else's output_log.txt have the following written after every call to MonoBehaviour.print? (Filename: C:/BuildAgent/work/ea95e74f6e5f192d/Runtime/ExportGenerated/StandalonePlayer/UnityEngineDebug.cpp Line: 54) That's 1 line with just a space, 1 with the filename and linenumber stuff and 1 empty line. This greatly increases the size of the file and makes it a lot harder to find the output you are looking for. If it does happen for everybody and isn't just a strange quirk of my installation (I presume not) then we should ask Squad to get rid of it. Debugging plugins without a sensible debugger is hard enough without having to wade through megabytes of spam to find your info or regex replace it so you can fit more than 20 lines of your output on the screen... There are also a couple of messages that come out as: WheelCollider requires an attached Rigidbody to function. (Filename: Line: 167) So, no filename and a different line number. Presumably that particular bit of code is using a different routine to write to the log than everything else does...

You haven't exactly given us a lot to go on. Which version did you used to have and which version are you having the problem with? Did you delete the old version before installing the new one? What does "it stops" mean? Does KSP quit or does it hang? Have you looked in the KSP_Data\output_log.txt file for any information?

The compiled version in this thread is basically my github branch with an extra few tweaks to add (Pad) to the window titles. The way the simulation works is considerably different though it is still called in the same way. What makes you think that fuel lines or staging aren't working properly? I've tested quite a few asparagus and onion staged craft and the numbers looked pretty good to me... Can you send me an example craft (preferably stock) that you think the numbers are wrong for? I have already looked at your new version and I don't think there should be any problem integrating it.

A quick update on progress for anyone not following the Test of new KER deltaV simulation code thread. The new simulation code appears to be much better but a couple of issues have cropped up with the Horizon Zenit mod. First, this mod sets the electric charge on the engines to be tweakable even on those with zero capacity and this seems to trigger a bug in KSP that, when you open the tweak panel, it shows NaN/0.0 and this breaks the mass calculations (with knock on effects on other values). I have worked around this and reported it to the author who will fix it in the next release of his mod. I will also report it to Squad as a KSP bug as it should just refuse to make it tweakable if the capacity is zero. The second issue is that KER is not currently handling engines with multiple engine modules correctly. It only uses one of the modules for the calculations which results in wildly wrong thrust (and hence TWR) values and (usually) somewhat less wrong ISP (and hence deltaV) values. This will require considerable work to fix "properly" as I think different engine modules could theoretically use different propellants which may mean sub-engines could burn-out part way through a stage, changing the effective ISP (as well as effectively simulating the individual engine modules as separate engines, this also needs the simulation to calculate the effective isp of the stage backwards from the stage's deltaV). Hopefully I will be able to look at this over the next few days. There are a number of more intricate tweaks that could be done such as taking account of the orientation of engines which may be useful for landers with splayed out radial engines or shuttle type designs with angled main engines where some of the thrust is wasted but they need more thought to work out how to do them properly (e.g. this example actually needs the deltaV to be calculated as the final magnitude of the vector sum of the individual engine contributions throughout the stage to account for both early burn out and the direction of the thrusts).

I am currently investigating a couple of issues in KER when using the engine parts from this mod. It appears that allowing the electric charge of an engine to be tweakable when it has a capacity of zero causes a problem when you open the tweak window in the VAB. Adjusting the slider in the tweak panel makes it change to 0.0/0.0 and KER starts calculating correctly (but see below) again. I have worked around this in my updated version of KER but the problem should really be fixed in the parts as the NaN/0.0 showing in the tweak panel looks very dodgy. None of the stock engines that have 0 electric charge capacity allow the value to be tweaked so it should be fixable by a simple tweak in the part configs... There is also another issue in KER with engines that have multiple engine modules where it only counts one of them for the thrust and isp calculations. I will be fixing this over the next few days.

This appears to be a problem with the engines in the Horizon mod. My KER window shows correct results until I bring up the tweakable window for an engine. After this it behaves as you say, the stage with that engine in and all stages below it in the rocket stop displaying mass, deltaV and TWR values. This appears to be because the tweakable window shows NaN/0.0 for the electric charge. If you drag the slider it changes to 0.0/0.0 and the KER window starts working again. I can probably work around this in KER by not counting any resource in the mass calculations that has a value of NaN but it should also be fixed in the Horizon mod... Edit: I have reported this issue in the Horizon mod's thread. Edit 2: There is also an issue in KER with the multiple engine modules in that part. It is currently only using one of the modules for calculating the thrust and isp of the engine so the isp, thrust, TWR and deltaV values come out wrong. I will work on this over the next few days but to fix it "properly" will require that the simulation keeps track of the individual modules independently in case they actually use different fuels (this could cause them to burn out at different times, changing the effective isp of the engine part way through a stage). I don't know if there are any mods out there with engines like this but I really don't like making half-fixes...

32 bit processes on Win 64 can actually use the full 4 gig of address space but some is still reserved for some things. You should be able to take KSP up to about 3.4 gig before you start getting out-of-memory crashes...

Most of those messages aren't coming from KER as far as I can tell. I know that KER wrote the "BuildEngineer: OnSave" message and it probably wrote the Lock and Unlock ones too but none of the messages after that look at all familiar to me and I can't find the text in the KER source. It looks like KER may have locked the editor UI just before something else got badly upset. KER only locks and unlocks the UI in the window drawing function by checking if the pointer is in the window, so if the core game stops calling KER to redraw then it can't unlock the UI. Runge-Kutta refers to an iterative method for approximating solutions to differential equations and there isn't any of the sort of thing in KER. I suspect this came from the core game or from another plugin that does do more complex maths...

I'll try to build another example from scratch that uses the same basic parts as yours and see if I get the problem...

Thanks, I'll try to take a look at it this evening. When I tested the Horizon parts I only tried a few very simple stacked arrangements but I have another build configuration that includes lots of useful logging so it should tell me pretty quickly what is going wrong... Edit: I managed to install the AIES pack and take a quick look over lunch but I get this in output.log when I load your craft. Cannot load Module ModuleSPU because the associated module on the part doesn't match the saved module Part probeStackLarge_4294410052 cannot load module #4. It only has 4 modules defined Part probeStackLarge_4294410052 cannot load module #5. It only has 4 modules defined Part probeStackLarge_4294410052 cannot load module #6. It only has 4 modules defined Part dockingPortLarge_4294409956 cannot load module #1. It only has 1 modules defined Part RCSBlock_4294409670 cannot load module #1. It only has 1 modules defined Part RCSBlock_4294409670 cannot load module #2. It only has 1 modules defined Part RCSBlock_4294409628 cannot load module #1. It only has 1 modules defined Part RCSBlock_4294409628 cannot load module #2. It only has 1 modules defined Part RCSBlock_4294409586 cannot load module #1. It only has 1 modules defined Part RCSBlock_4294409586 cannot load module #2. It only has 1 modules defined Part RCSBlock_4294409544 cannot load module #1. It only has 1 modules defined Part RCSBlock_4294409544 cannot load module #2. It only has 1 modules defined Part solarPanelsuncat1_4294408838 cannot load module #1. It only has 1 modules defined Part solarPanelsuncat1_4294408754 cannot load module #1. It only has 1 modules defined Part solarPanelsuncat1_4294408670 cannot load module #1. It only has 1 modules defined Part solarPanelsuncat1_4294408586 cannot load module #1. It only has 1 modules defined Part HA3SLStage1Decoupler_4294408378 cannot load module #1. It only has 1 modules defined Part HA3SLCenterDecoupler_4294407984 cannot load module #1. It only has 1 modules defined Part HA3SLStage2FuelTank_4294407840 cannot load module #0. It only has 0 modules defined Part HA3SLRD120_4294407808 cannot load module #5. It only has 5 modules defined Part HA3SLStage1Decoupler_4294406784 cannot load module #1. It only has 1 modules defined Part HA3SLRP2FuelTank2-1_4294406684 cannot load module #0. It only has 0 modules defined Part ASDecoupler_4294406300 cannot load module #1. It only has 1 modules defined Part fuelTankmr1_4294406244 cannot load module #0. It only has 0 modules defined Part launchClamp1_4294406212 cannot load module #2. It only has 2 modules defined Part fuelTanksuperior3_4294406104 cannot load module #0. It only has 0 modules defined Part fuelTanksul_4294406072 cannot load module #0. It only has 0 modules defined Part HA3SLRD120_4294406040 cannot load module #5. It only has 5 modules defined Part ASDecoupler_4294405946 cannot load module #1. It only has 1 modules defined Part fuelTankmr1_4294405890 cannot load module #0. It only has 0 modules defined Part launchClamp1_4294405858 cannot load module #2. It only has 2 modules defined Part fuelTanksuperior3_4294405750 cannot load module #0. It only has 0 modules defined Part fuelTanksul_4294405718 cannot load module #0. It only has 0 modules defined Part HA3SLRD120_4294405686 cannot load module #5. It only has 5 modules defined Part ASDecoupler_4294405592 cannot load module #1. It only has 1 modules defined Part fuelTankmr1_4294405536 cannot load module #0. It only has 0 modules defined Part launchClamp1_4294405504 cannot load module #2. It only has 2 modules defined Part fuelTanksuperior3_4294405396 cannot load module #0. It only has 0 modules defined Part fuelTanksul_4294405364 cannot load module #0. It only has 0 modules defined Part HA3SLRD120_4294405332 cannot load module #5. It only has 5 modules defined Part ASDecoupler_4294405238 cannot load module #1. It only has 1 modules defined Part fuelTankmr1_4294405182 cannot load module #0. It only has 0 modules defined Part launchClamp1_4294405150 cannot load module #2. It only has 2 modules defined Part fuelTanksuperior3_4294405042 cannot load module #0. It only has 0 modules defined Part fuelTanksul_4294405010 cannot load module #0. It only has 0 modules defined Part HA3SLRD120_4294404978 cannot load module #5. It only has 5 modules defined Part HA3SLRD171_4294404884 cannot load module #4. It only has 4 modules defined Part LRDecoupler_4294407538 cannot load module #1. It only has 1 modules defined Part tank1300cl_4294407482 cannot load module #0. It only has 0 modules defined Part liquidEngineconstelacion_4294407450 cannot load module #5. It only has 5 modules defined Part LRDecoupler_4294407356 cannot load module #1. It only has 1 modules defined Part tank1300cl_4294407300 cannot load module #0. It only has 0 modules defined Part liquidEngineconstelacion_4294407268 cannot load module #5. It only has 5 modules defined Part LRDecoupler_4294407174 cannot load module #1. It only has 1 modules defined Part tank1300cl_4294407118 cannot load module #0. It only has 0 modules defined Part liquidEngineconstelacion_4294407086 cannot load module #5. It only has 5 modules defined Part LRDecoupler_4294406992 cannot load module #1. It only has 1 modules defined Part tank1300cl_4294406936 cannot load module #0. It only has 0 modules defined Part liquidEngineconstelacion_4294406904 cannot load module #5. It only has 5 modules defined Lots of the part masses are then coming back as "NaN" so the deltaV and TWR calculations don't work but the code doesn't hang or otherwise crash and it starts working again if I start a new craft and don't use the parts. I only have v1.1. of the Horizon Zenit pack so I'll update that to 1.2 and try again a bit later... Edit 2: Actually, looking at the craft file it appears that you have other mods installed that piggyback onto existing parts. RemoteTech and TweakableEverything(?). This may be unimportant but I'll fiddle later and see why the masses are coming out as NaN...

Do you mean that the window stopped updating or did it give obviously wrong values? I tested a few combinations of the Horizon parts and didn't notice any obvious problems but I haven't tried the AIES ones. Can you post a craft file that breaks it please?

Yes, that station is put together from 4 vertex craft and 6 edge craft all docked together. KER used to get stuck in an infinite loop which would stop all update of the VES tab until you restarted KSP (or with the older single-threaded versions of KER, like 0.6.1.5 it would totally lock up KSP). The new simulation code works in a significantly different way and correctly accounts for parts it has already visited when scanning for fuel sources so this station works now...

I've not actually done much in the last week or so because I've been too busy rewriting the deltaV calculation code in KER to fix some serious issues. Happily, that is now done and available for testing in this thread. It finally allowed me to move this problem station to a higher orbit to improve the framerate without making the KER vessel tab stop updating until you restart KSP (with an older version of KER it would hang KSP completely). As an amusing aside, it all went well until I deorbited the RCS tanker/tug. It is designed to land back on Kerbin using chutes and a little blip of engine throttle to cut the speed just before touchdown. There was still a lot of RCS fuel onboard (probably 80% of the total) so it was much heavier than it was during testing. Everything was going fine, the chutes semi-deployed and put me on course for a landing very close to KSC, and then the chutes fully opened and the probe core (where the chutes were attached) broke off from the topmost RCS tank leaving just the core and the docking port floating down under the chutes and the rest of the ship fell the remaining 500m and exploded. Guess I'll have to tweak it and test it again with a full load of fuel and RCS. I don't really want to have to strut the probe core to one of the lower tanks to prevent it coming apart but if that's what it takes...

Thanks for your report. Glad to hear it is working with your craft that used to break it... P.S. You win a cookie (and rep) for being the first to respond...

Please post other suggestions for KER in the main thread that I linked to in the first post.

I have just posted a thread asking for people to test the new code here. Please post any follow-ups about my changes to that thread instead of this one...

This thread is for testing and discussion of my on-going development of KER 0.6.x.x. This development version is currently based on KER 0.6.2.9 and should only be installed over that version. Download development version. (Updated 12/08/2014 20:57 GMT) Note: This is compiled against the 0.24.2 dlls so should only be used with that version (though it should work with any version >= 0.24). Change Log: 12/08/2014 20:57 GMT Fixed possible exception in SetupAttachNodes. 03/08/2014 14:55 GMT Incorporated 0.6.2.9 updates from the official version. Modified simulation staging logic to fix issues with empty stages. Removed pointless calculation of "active" stage when no engines are active. 25/07/2014 21:22 GMT Incorporated 0.6.2.8 updates from the official version. Removed new tweakable PartModules and added tweaks to settings window instead. Added settings button to build engineer. Fixed an exception thrown by the Log Sim code. Fixed compact button to keep right side of window static. Fixed some null ref exceptions in the Blizzy toolbar code. 24/07/2014 11:03 GMT Added error handling to config file loading. 21/07/2014 21:31 GMT Added check for NEAR to disable atmospheric readouts. 20/07/2014 19:41 GMT Fixed FormatSI to not run off end of array. 20/07/2014 09:54 GMT Another little tweak to the toolbar handling. 19/07/2014 21:06 GMT Incorporated 0.6.2.7 updates from the official version. Fixed toolbar issues in SPH. 19/07/2014 13:52 GMT Incorporated 0.6.2.6 updates from the official version. 18/07/2014 17:03 GMT Incorporated 0.6.2.5 updates from the official version. 17/07/2014 22:50 GMT Fixed for 0.24 changing part costs to float. 01/07/2014 22:35 GMT Fixed struts allowing fuel flow during flight simulation. 28/06/2014 10:29 GMT Only calculate ground slope when readout is being shown. Added explicit new WaitCallback. 23/06/2014 22:02 GMT Implemented slope angle and heading display in SUR tab. 16/06/2014 22:05 GMT Added biome and impact biome readouts to SUR tab. 15/06/2014 00:33 GMT Fixed ports docked during flight not feeding fuel correctly Fixed simulation of parts using NathanKell's CrossFeedEnabler Fixed VES tab to not blank most values when "active" extra stage is displayed Added "Log Sim" button to build and flight engineer 08/06/2014 21:17 GMT Fixed issue with fuel tanks with fuel crossfeed disabled Small optimisation to GetSourceSet to eliminate pointless recursive calls 31/05/2014 22:03 GMT Added Dump tree button to build engineer 18/05/2014 22:02 GMT Implemented correction for engine thrust vectors enabled via tweakable toggle 17/05/2014 16:33 GMT Implemented velocity thrust correction for jets 15/05/2014 22:43 GMT Fixed log spam with non-RF engines when RF is installed 14/05/2014 20:34 GMT Added checking for third type of RealFuels engine module Fixed AllowedToStage to not stage if nothing is decoupled and any engines are still burning 11/05/2014 21:28 GMT Added button to flight engineer tweakable that causes the next run of the simulation to dump the part tree to output_log.txt Tweaked RealFuels thrust correction to detect and log some potential causes of blank outputs 09/05/2014 22:31 GMT Added extra stage to in-flight simulation that uses the currently active engines This is currently a compiled version of the "staging" branch of my fork on GitHub. I haven't updated the version number as this is not an official release of the plugin and there may still be bugs but I would welcome other people to test it, especially if they currently have problems related to the areas I'm working on. The zip file only contains the two dlls so, to "upgrade" to this version, you should: Quit KSP. Back up your current GameData\Engineer folder. Extract the two dlls and replace the ones in your GameData\Engineer folder. Run KSP. If you are installing it for the first time then simply install the standard 0.6.2.9 version from the main KER thread and then replace the DLLs with these. Please report your findings in this thread to avoid cluttering up the main KER thread with development discussion. Many thanks, Pad.

Doh! I've just sent you a link to the zipped dlls in a PM. I should have checked the thread first...

Thanks for working this out. I suspect the tanks from that mod have something "wrong" with them as far as KER is concerned and this is causing an exception in the simulation code causing the calculations to be aborted and possibly leaving KER thinking that the simulation is still running (so it doesn't run it again). I'll download it and have a look if I can see what is going wrong... Edit: I've just downloaded the pack and my new code manages to calculate the deltaV with no problems for all the engines and fuel tanks. I'll give it a try with the 0.6.2.3 version in a little while... Edit 2: By the way, that mod was updated to v1.1 yesterday so if you haven't got the latest version of it then I would give that a try. They may have fixed something in the definitions of the parts. Edit 3: I've just tried it with 0.6.2.3 of KER and, sure enough, the simulation code is dying, though I haven't dug into it to see exactly why (not much point since my new code fixes it).

Some of the issues with the calculations freezing are definitely caused by bugs in the simulation code. I believe that my modified version of the code has fixed these issues. Crazor, I've not noticed any issues like you describe with fuel lines. I have quite a few asparagus and onion rockets and the deltaV values were always correct (and still are with the new code). As for the typo in the rendezvous window, I've noticed it several times and have meant to say something about it but always forgot. I have added another commit to my pull request that fixes this.

Well, it took me a bit longer than I hoped but I've finally got my changes working... I then proceeded to move the station from its 140km orbit up to 170km (framerate gets much better above 160km) with the VES window updating correctly during the burns. I've also loaded about 80 different craft in the VAB and switched to 20+ vessels in flight and the numbers appear to check out. Given the substantial changes to the simulation algorithm (it now explicitly uses the fuel flow rules that Kasuha posted and the simulation now generates its own data structures from the core game structures in the main game thread so the background thread doesn't access anything in the core game structures that could change during the simulation) there is a reasonable chance of a bug or three but it is looking good so far... Cybutek, I'll send you a pull request for you to take a look at shortly.

Nothing that isn't running physics can collide with anything. So, as long as you aren't flying something within 2.5 km of the "collision" the ships will just pass straight through each other...

Does the 5830 have more memory (or a wider internal memory interface)? I use a 5 year old laptop with a Core Duo T2600 at 2.16GHz, 3Gb of RAM and Nvidia Quadro NVS 110M graphics with only 128Mb of dedicated memory. Even with all the settings highly optimised I still often see framerates below 5. Is it still fun? You bet!

Only approximately by looking at the G meter on launch or timing (by the MET clock) how long it takes to ascend a certain distance (e.g. 100m) and working out the acceleration (should be fairly constant over the first short distance) from that... Alternatively just install KER and load the vessel in the VAB and attach the build engineer part for a screenshot and then remove KER again if you want to stay stock... Edit: Duh! Yeah you could do that too, he did say how many engines it had...

You think that's bad... Also, there is a single thread benchmark on that site too... http://www.cpubenchmark.net/singleThread.html I fare a bit better on there... only about 4 times slower than the top-dogs...