[WIP|1.7.10] FastCraft 1.25 (Note: All Posts of new Members need to be approved first, so if you press the submit button but no new post appears here, it is in the folder of posts to be approved first, and Player will receive them a few hours later)

  • Hi Player,


    I've been using your fantastic mod on my server ever since you posted it here and up until now the performance has been amazing.


    However since a few days the server just randomly started lagging like crazy. (I have been running it with FastCraft for a while now, but I have not made any updates since you released 1.19).
    It just randomly started, and now at times it gets so bad you get kicked off the server. At times the console says it's skipping 200+ ticks! It was all fine a 3-4 days ago. Apart from "Can't keep up" warnings, the server doesn't crash and there is NOTHING in the log.


    I've tried using your sampler and looked at the file through VisualVM but I'm not sure what I need to be looking at.


    I also tried using spiketrigger 60 and in a matter of seconds I had ~30 spiketrigger files and the server crashed itself.


    It would be awesome if you could take a look at them:
    (they are hosted on my servers website)


    Longer sampler run~ 5 minutes


    spike1 , spike2 , spike3 , spike4


    (These 4 are some of the spiketrigger files that were created in under a second)



    If you need anything else, please let me know.

    • Official Post

    bochen415: I blame gradle, it seems to not refresh all files properly -_-


    DreamMasterXXL: It's an issue with IHL Tools, it's accessing IC2 core functionality which has changed.


    CreeperShift: The long run, which is indicative for the average load, only shows EnderIO's TileConduiBundle, more specifically the redstone variant, being somewhat high.
    The spikes are from Mekanism refreshing its energy network in a very inefficient manner, i.e. loading plenty of cascaded chunks in the process and probably restarting the refresh for every block added all over. Don't use its cabling for anything physically extensive and/or chunk load the whole area containing such a continuous network.
    It's recommended to use something like spiketrigger 200 or higher btw, 60 would already trigger once the tick is exceeded by even 10 ms, which isn't really a problem. Also note that the 1st number a in the original spike-a-b.nps filename contains the spike/tick length in ms, files with higher values are of increased interest of course.

  • bochen415: I blame gradle, it seems to not refresh all files properly -_-


    DreamMasterXXL: It's an issue with IHL Tools, it's accessing IC2 core functionality which has changed.


    CreeperShift: The long run, which is indicative for the average load, only shows EnderIO's TileConduiBundle, more specifically the redstone variant, being somewhat high.
    The spikes are from Mekanism refreshing its energy network in a very inefficient manner, i.e. loading plenty of cascaded chunks in the process and probably restarting the refresh for every block added all over. Don't use its cabling for anything physically extensive and/or chunk load the whole area containing such a continuous network.
    It's recommended to use something like spiketrigger 200 or higher btw, 60 would already trigger once the tick is exceeded by even 10 ms, which isn't really a problem. Also note that the 1st number a in the original spike-a-b.nps filename contains the spike/tick length in ms, files with higher values are of increased interest of course.

    Thanks for the Information! I don't personally use any mekanism cables, but I know a few users who I can blame >_>. For now, I have removed all chunkloader from the map temporarly (as a test) and the lag is still present. The only loaded area at that point (my base) contains not a single mekanism cable/pipe. It contains maybe ~15 mekanism machines, all connected via EnderIO conduits tho. Are they still likely the cause of that lag? The server is on a dedicated host, and right now even when no one is online, with only the spawn loaded, it's skipping ticks like crazy.


    Also sorry about changing the filename of the spiketrigger :/ I'll do a proper one with 200 or higher once I get home from work.


    Still, thanks for the quick reply!



    EDIT:
    Derp, I can do it from the console:
    0 Players online, all chunkloaders are removed, nothing except "skipping ticks" error in console:
    Getting a spike every ~5-15 seconds.


    spike-1131-2438273084
    spike-1134-2438228094
    spike-1213-2438318177
    spike-1249-2438093240
    spike-1253-2438183229



    These should be correct then I assume?

    • Official Post

    I'm not exactly sure if it's actually cables, the stack trace just indicates some "TransmitterNetwork" and multiparts, which could as well be the machines themselves, although unlikely.


    Removing chunkloaders won't help unless you get every part covered and then restart the server to remove any chunks stuck loaded. It's obviously wasted effort if the spawn area, a player or another type of chunk loader (quarry etc.) loads a connected part anyway..

  • ! IMPORTANT FOR ALL !there are big incompability problem with fastcraft <>macpatcher+MOJO x256 - !gigantic! lags


    DONT INSTALL FASTCRAFT IF YOU USE THIS PACK !
    (or same type packs may be )

  • I'm not exactly sure if it's actually cables, the stack trace just indicates some "TransmitterNetwork" and multiparts, which could as well be the machines themselves, although unlikely.


    Removing chunkloaders won't help unless you get every part covered and then restart the server to remove any chunks stuck loaded. It's obviously wasted effort if the spawn area, a player or another type of chunk loader (quarry etc.) loads a connected part anyway..

    Hmm. It's weird, it was running fine for a month and now suddenly this. We haven't actually updated any mods in the last 2 weeks (except FastCraft) so I'm really out of ideas what to do now. The spikes are getting to a point where you can no longer play on the server. Do you have any other ideas I can try out? Else I might just switch over to our new pack prematurely.

  • Quote

    Very large resource packs like this are a problem in general unless you have 8 or better 16 GB of RAM.

    please do not enter players in misleading!


    with fastcraft i have 17 fps and freeze every second, without -117+ fps


    all good with pack and RAM - only 1800 mb memory usage with 100+ mods

    • Official Post

    A pack change won't help it besides the map reset, you just have to remove all EnderIO redstone conduits. Someone apparently just happened to build something problematic with them recently or modified an existing setup. It's by far not the first time I'm seeing EIO rs conduits cause severe lag spikes, they have some inherent implementation problems.

  • A pack change won't help it besides the map reset, you just have to remove all EnderIO redstone conduits. Someone apparently just happened to build something problematic with them recently or modified an existing setup. It's by far not the first time I'm seeing EIO rs conduits cause severe lag spikes, they have some inherent implementation problems.

    Ah no, with the pack change I meant switching over to a new pack we've been making, new map included. Nothing being copied over.


    I have removed the EnderIO mod and loaded up the world, and now the lagspikes are down to 215-300.
    I'll put the map back in and talk to a few people and remove their conduits.


    Thanks for the help and awesome replies! :)


    oh by the way, by redstone conduits we are talking about the conduits that transport redstone signals, not the low-tier energy conduits right? If thats correct, I'll get everyone to remove their cables and disable the crafting recipe.

    • Official Post

    please do not enter players in misleading!


    with fastcraft i have 17 fps and freeze every second, without -117+ fps


    all good with pack and RAM - only 1800 mb memory usage with 100+ mods


    You need to look at the memory usage in the task manager, the textures and other OpenGL objects aren't being stored on the heap and thus don't show up in F3 etc.

  • its info from task manager - 1700mb on my modpack ( i have xmx4g and 8 gb ram so all ok )
    but low fps its not 1 problem, i feel freezes during play even on mc with 0 other mods with fastcraft and mojo -mcpatcher

  • Quoted


    Very large resource packs like this are a problem in general unless you have 8 or better 16 GB of RAM.
    please do not enter players in misleading!


    with fastcraft i have 17 fps and freeze every second, without -117+ fps


    all good with pack and RAM - only 1800 mb memory usage with 100+ mods


    If you are having stellar performance without FastCraft, why are you even trying to use FastPack?

  • Just use it if you fulfil the requirements :)

    I'm not exactly replying to this, but people are still being stupid and not reading the pages before, even though they saw my earlier one stating to read your post on page 21 if you're trying to put it in a pack, and i'm just stating this again so they will hopefully see it this time.

  • Hi Player


    Just thought I would let you know that I intend to use Fastcraft in a private modpack of my own creation. Its worked wonders for me in FTB infinity (first modpack without stutter ty) so I thought I would add it to my pack also. I have followed your guidelines. If you want any more info feel free to PM me

  • Hello I am reporting a crash when using Fastcraft 1.19 and Mystcraft 0.11.00. It only occurs when generating the libraries that Mystcraft adds to it's dimensions. The crash report mentions Fastcraft so I thought I should mention it. Hopefully it's nothing to big :)