Gkirmathal

Thanks for the feedback TotheMun. Already thought that could've been the root cause and I was merely describing its symptom I guess. Not found a post describing this symptom though. Btw, someone an other thread (memory leak I thought) wrote even his Linux x64 system came to a crawl, after KSP had flooded his 16Gb ram and fully filled his swap. It worries me so many users reporting these several memory management issues and I have read nothing about any technical, engine related, developments in the Dev Blogs.

Hey there community. I have been having an odd issue while playing KSP since I got it and never had this issue before I had KSP installed. Don't know whether it is mod related or not, or if it is an Unity4.x limit, as I have not found any info about my issue here on the forums. So that's why I post here. The problem. After KSP runs for more than 30 minutes, I start noticing that my system has instances in which it becomes less responsive, these instances show very high disk activity. Having a dual screen setup and I run a browser on my second monitor which also started to lag significantly when these instances happened. Ran Task Manager (standard practice) to see what is going on I/O wise after some ingame time. Results were processes: KSP.exe and CSRSS.exe, had I/O Writes and Reads counts that went through the 1 to 3+ million mark. Every time high disk activity occurred, were after scene changes and or when 'alt-tab' to look something up/mod some cfg's. My system: W7 x64 Intel X5460 @3.8 6Gb ram 6Gb swap partition on a different drive from OS + game partitions. Ram with KSP + FF + system still has 1.5Gb free, but swap seems to have increased to it's max size. Has anyone ever ran into something similar, or can anyone perhaps corroborate this?? Thanks Edit: I am just about to end another session, ksp is at 3.3Gb, now ~1.5 hours in and KSP.exe I/O Writes count stand at 14.97M counts and CSRSS.exe at 2.16M counts. FF is slow and somewhat unresponsive in browsing.

Even as a new player, I can report the same thing here. With ATM aggressive + DDSloader + forced d3d11, KSP starts out with ~2.78Gb ram on main menu and ~2.85Gb on Space Center. While building/loading a craft in VAB, my ram usage shoots up to around the ~2.9Gb mark. When launching it increases again to ~3.0Gb +/- 0.7 depending on the craft complexity. Now the worrisome part: reverting back to VAB/Space Center/quickload, does not seem to unload the previous scene from ram and relaunching does not seem load this previously cached scene still in ram. There are more instances in KSP where this happens. To me this seems, what I heard and understood from friends who do programming: is that this Unity4.x implementation has very poor and non intelligent garbage collection. That, or parts/textures/data etc are not correctly flagged in the code garbage collection, thus the garbage collector does not recognize them and that results in a steady climb of memory usage. Albeit: a memory leak. As a newly player to KSP, I was shocked they wanted to go to a 1.0 release. Without the above being fixed (good working garbage collection), even without good x64, 1.0 will be devastating for the game.

Thanks for the imput! <snip lengthy post> Edit! Overlooked one small typo...need to remind myself not to edit 4 config at once while tired.

Huy guys, great mod! But got a problem getting it to work with tech tree mod 'Mod-Oriented Tech Tree' and coulnd't find a solution in this thread. The mod has custom techID nodes for fairing mods and their fairing sizes. The PF base rings are all placed in their correct modded nodes and I can unlock them! The issue: the PF size functionality (for above 1.25) is only unlocked when you unlock the tech node with the stock required techID (aerodynamicSystems for example) and does not correspond to the custom fairings node from MOTT (techID = fairings). Now I had a look in all PF part.cfg's and thought changing most RequiredTech techIDs would do the job, but changing them to the corresponding custom techID nodes didn't work. Now I am sort of at a loss on how to get it to work. Could anyone help me out?? I'd really like using PF with MOTT tech tree mod.

Svod, thanks for this tech mod! This techtree really makes the KSP+mods experience, so much more logical in comparison to stock and I started playing KSP not too long ago only because of the NF mod I was following since 2014. This really saved my KSP experience and makes me look forward to start in Career mode. Anyway, some balance things I did ran into regarding NF fuel tanks, stock science + DM orbital science and some nuclear engine mods (FTmN+AtomicAge). First of. It felt like the LARGE NF hydrogen fuel tanks, especially, were unlock-able too early on. (Made a personal edit so the NF hydrogen-25/125/rad-25 are the only tanks available at Nuclear Power node. The rest come after.) Second thing. About most stock science equipment + DM Orbital Science. Some parts are not placed 'logically' for when playing with DM Orbital Science parts. Especially the parts from nodes 'Science' to 'Even More Science'. Experimental Science node look ok. FTmN parts seems changed, the 280 for instance is no longer in the pack? If any want to know what I rearranged in detail let me know. Svod, again my thanks!

Hey odd place to make my first post but. Just out of curiosity installed this (along with NearFuture I have been using) to see what it's about and liking what I have seen thus far. One, huge, snag though. I constantly run into a nasty kraken bug problem on quickload and focus changes: every ship build with it, encounters the kraken/disintegrates in orbit, moments after the drives 'load animation' for absence of a better term. Other ships (build in sandbox for testing) in the 70t range do not suffer this. Can anyone advice me what this might cause and or if it is fixable? I use Kerbal Joint Reinforcement, which came to mind, but I have too little experience with KSP modding to know if it could be a culprit. Edit: it would have been logical to check all my mods, seemed there was a new KJR version released, it fixed the issue.