Jump to content

waterlubber

Members
  • Posts

    675
  • Joined

  • Last visited

Reputation

368 Excellent

1 Follower

Contact Methods

Profile Information

  • About me
    F5, F9, repeat
  • Location
    Location
  • Interests
    Space! Elite Dangerous, Space Engineers, KSP (Duh!), the whole thing! Books with Space too!

Recent Profile Visitors

4,582 profile views
  1. This is exactly what I expected, and would be my normal course of action, had I not had two identical (or nearly identical) installs on my roommate's machines that did not experience nearly as high a memory usage. The problem is unique to this install, for some reason. I was actually able to greatly reduce startup memory by (I think?) removing DarkMultiPlayer -- perhaps it was causing problems. However, there's still a massive memory leak remaining as usage increases over the course of a few hours of gameplay. I'm not sure what tools exist to identify leaks (by mod would be amazing) except perhaps MemGraph. Even with this many mods, the game is still allocating 2-3x its size on disk in memory. Even accounting for the incredible waste of garbage collected languages that's still crazy amounts of usage. Maybe it is just the number and scope of the mods, but I've got a feeling that its something using far more than its fair share that I'd like to specifically eliminate.
  2. I am experiencing some profoundly absurd memory usage from my KSP install. After setting it up, the game would refuse to load since the oom-killer would get it, although I got it playable by adding 16GB of swap. During gameplay, virtual memory usage is ~24GB or so, but the game is swapping enormous amounts of data to disk every time I change scenes. I'm guessing there's some sort of mysterious memory leak, but I'm not sure to troubleshoot it and I'm not sure if it's an issue with the game or Arch Linux. Normally, I would chalk such an enormously high usage up to this particular combination of mods; however, my roommates (three of them) are running an identical install on Windows and use approximately 4-6GB of memory. Something is uniquely wrong about my install, and I'd like to work out what it is since at the rate the game swaps out when changing scenes my SSD will be dead within the month. Here's some information: Distro: Arch Linux, Ryzen 2600 on 16GB RAM KSP: 1.12.2, 64-bit. Player.log: https://mega.nz/file/LYB2kToL#USYiVe0_3Hsjmf_CkpvtF6PcJyrRVNudIei6K7ScvDc (not entirely sure why it ends with a crash, but I did actually exit the game...) Mod list: https://pastebin.com/3i4QyDR0 I'm guessing that this issue is most likely caused by some unhealthy interaction between two mods, but that's just speculation. Anyone here have any ideas, or have experienced anything similar? I've got no idea what's going on but would like to play the game without waiting two (2) minutes at every scene transition for my computer to swap 8GB to disk and back.
  3. I do have Memgraph installed, but I wasn't using any of its active (memory-poking) features at the time. It could be an issue with mono or some other weird low-level thing though, as I am playing on Linux. Thanks for the help. In the meantime I'll just spam RTGs for power. Edit: It appears to apply to all reactors, as my open cycle Safety Ray™ refuses to fire up without crashing the game.
  4. 1. UX-0 thing was Community Parts Titles. My bad, didn't realize it also did the Near Future Parts. 2. I tested with an install with just the NF mods and DBS. The issue did not occur. I was also unable to replicate it with Heat Control radiators. It appears to be caused by a really weird interaction on that particular installation.
  5. I'm experiencing some weird issues with DBS and NFE. My game crashes when I fire up the small .65m reactor, but the logs hint at DBS being the culprit. Unfortunately, the craft that this issue is 249 parts, and it takes about 10 seconds to add or remove one on my computer. Notably, when designing it in the VAB the thermal UI eventually stopped recognizing the radiators - for whatever reason, it didn't process the new radiators that I added to the thing. I'm playing on Linux AMD64. DBS is version 2.0.5.0, NFE is 1.0.2, installed via CKAN. All other mods are latest CKAN versions as of Aug 20, 2019. TweakScale is installed and a few parts on the craft are tweakscaled -- no radiators or reactors, though. There's an Emancipator engine as well. My modlist (CKAN export): Craft File, player.log, persistent.sfs, loadmeta The tail end of the Player.log suggests that it's DynamicBatteryStorage, it's almost certainly some kind of weird interaction though. EDIT: I was able to reproduce the issue with just a probe core, a NFE fission reactor, and an expanding heat radiator. It took MUCH longer after starting the reactor for the issue to manifest (about a minute). The tail end of the Player.log, which was the only thing of interest, is as follows: Finally, here's Player.log with debug on. Unfortunately, there doesn't seem to be any additional information. This appears to be a rather low-level bug. Steps to reproduce: Either load the existing save with crafts and try to start the UX-0 reactor, or just launch a craft with a probe core, UX-0 reactor, and sufficient expanding NFE heat radiators. Star the reactor and wait several minutes, eventually the game will crash. The former method should make the crash happen faster.
  6. Odd. I cloned the repo and just copied over the entire GPP_Scatterer folder, but I'm still getting black cloud errors.
  7. Speaking of things I love, GPP is working fine on 1.6 except for the aforementioned scatterer black clouds. Are there any workarounds at the moment? I'd be willing to turn off clouds or use a 2d style or something like that. I'm also excited for whatever that "labor of love" is Here's my black clouds player.log, not sure if this is under scatterer's jurisdiction or yours. I'll go over there if need be. https://mega.nz/#!DUJwUYjI!DTm3srLuv_0tO8T8RmopnQn0ZsjWUgD0RKWh3qqYLnU
  8. Yeah, no FAR, version 1.4.3. I placed them without symmetry so the flaps would deploy properly, they're attached to some big procedural wings towards the very back of the plane. I also have a set of canards up front. Pitch is working fine, but roll is inverted (which means I can't just set the actuation to -100 and invert them.)
  9. Since you might be busy, what can I do to try to resolve the roll/inversion issue? Right now they seem to be fine for pitch/yaw but roll is inverted.
  10. I have no idea. Experiment. It's a compromise between memory usage and lag spikes. I think mine is ~2048 or so
  11. Depending on your total amount of ram (and operating system: Linux handles OOM far better than Windows) you can edit the "padhead.cfg" file inside MemGraph's folder. Just increase the max value (I think I use 2048 or 4096, I can't remember) and then hit Mod + End to allocate more RAM to KSP, and allow it to do GC's less often.
  12. I'm playing with Galileo's Planet Pack. Unfortunately, when I tried to do the investor mission, the poor guys spawned out in the middle of the ocean. Is there a way, or place, for me to edit my saves so the prerequisites for the casino are fulfilled?
  13. I did, before I got this message, while researching KSC garbage collectors. For anyone reading this, "a lot" is an understatement. It brought down the lag spikes from once every ten seconds to once every two minutes.
  14. This was the prospect of the ScienceAlert mod...you can see where "Here and Now" is inspired from I'm using the modified .dll. Performance has noticeably improved, but I still get big lag spikes. They increase in frequency and duration as the game goes on, which makes me think it's a memory leak and the garbage collector is struggling.
  15. I think I gave up and used SVT instead. This was a while ago; I'm playing with GPP now. Sorry for the late reply.
×
×
  • Create New...