Been away for a while. When playing the game, I can either move around or stand in one place with the F3 debug screen up and watch as my memory use rises from about 25% to 100%. Of course, at that time, the game "glitches" or pauses. As one can imagine, it's very annoying.
Clean install of Minecraft, OptiFine, nVidia graphics drivers, and Java today—moments ago, actually. Problem is exactly the same in both vanilla and OptiFine. JVM arguments are -Xmx3G and -Xmn3G—the same I've been running since 1.8—but it doesn't matter how much or little RAM I give Minecraft, the problems is the same.
Running a GTX 970, Minecraft runs consistently at a 32-chunk render distance at the max frame rate I allow: 140 FPS. However, I also reduced my render distance to just 8 for testing. Same problem.
So, as I said, everything was a fresh install. I pulled my massive, two-year-old SSP world over from backup and started playing, eager to get back into Minecrafting. That's when I encountered the problem. Messed with settings. A few cold boots. The problem lasted all day. Couldn't figure it out, so I posted here.
Just now, I started a new world and the problem did not carry over to it. Scared that there was something wrong with my old world, I tried playing it again . . . No problems.
As far as I can tell the solution to the problem was to start a new world . . .
Resolved, I guess.
Rollback Post to RevisionRollBack
My short story-like journals; quick-and-easy reads:
JVM arguments are -Xmx3G and -Xmn3G—the same I've been running since 1.8—but it doesn't matter how much or little RAM I give Minecraft, the problems is the same.
Are you sure about that? You should never, ever set Xmn to the same as Xmx or even higher than the default - it is just asking for lag spikes (specifically, it allocates memory for the "young generation", containing short-lived objects, of which recent Minecraft versions, since 1.8, are notorious for generating; the default of 128 MB was set by Mojang for a reason and is actually lower than what Java would otherwise allocate with the default 1 GB of RAM (typically, 1/3 of Xmx, so 128 MB would go with just 384 MB of RAM). I don't even use the default arguments myself; I replaced them with -Xmx768M -Xms786M -Xss1024K (the last particular to 32 bit systems to avoid stack overflow crashes).
Also, as for why a new world would seem to fix the problem, the issue could be the way the game handles structure data; specifically, for mineshafts, which are absurdly common even since 1.7 (one every 250 chunks; in 1.6.4 it was one every 100 chunks; in both cases they are well over an order of magnitude more common than all other structures combined, as well as the largest and most complex) and due to the way the game stores the data in memory it uses over a hundred of times more memory than the file size of "Mineshaft.dat" would indicate:
In fact, I even modded the game to prevent it from saving this data; my 3+ year old, 328 MB Survival world runs just as well as a brand-new world (maybe slightly worse around my main base due to 400+ tile entities, but the game does not lag). For vanilla you can periodically delete the file (there is no point in saving mineshaft data other than to ensure they generate correctly when world generation changes, and that would only work if they made them less common, as in 1.6.4-1.7.2, where the 40% of mineshafts that still exist are in the same locations). The only issue if if you explore a lot in a single session, as it automatically recreates the file as chunks are loaded (this is why they said to load structures you wanted to save in 1.6.4 before going to 1.7).
(I would do eveything I could to fix my world before I'd make a new world, the only world I've played on long-term or kept after I stopped playing it, and Minecraft would never be the same to me if it came to that)
I've had no issues with the Xmn being set to the same as the maximum, but I'll remove that argument. Thanks.
Like I say, I made a new, unrelated world for testing, and after making that new world, the old world, which has nothing to do with the new one, now runs just fine. To me, it makes no sense at all, but I'm not worried about it.
Whatever the case, I now have no issues.
Rollback Post to RevisionRollBack
My short story-like journals; quick-and-easy reads:
Been away for a while. When playing the game, I can either move around or stand in one place with the F3 debug screen up and watch as my memory use rises from about 25% to 100%. Of course, at that time, the game "glitches" or pauses. As one can imagine, it's very annoying.
Clean install of Minecraft, OptiFine, nVidia graphics drivers, and Java today—moments ago, actually. Problem is exactly the same in both vanilla and OptiFine. JVM arguments are -Xmx3G and -Xmn3G—the same I've been running since 1.8—but it doesn't matter how much or little RAM I give Minecraft, the problems is the same.
Running a GTX 970, Minecraft runs consistently at a 32-chunk render distance at the max frame rate I allow: 140 FPS. However, I also reduced my render distance to just 8 for testing. Same problem.
DxDiag Log: http://pastebin.com/qEKQu3UR
Additional information available upon request.
Thanks for reading.
My short story-like journals; quick-and-easy reads:
My Quest for Elytra Complete! (Pic Intense, End-Game Spoilers)
[Journal & Pics] After a Year and a Half, I Finally Found a Jungle
FrozenCore: Hardcore Death; 3/20/15 to 5/3/15; Eight Weeks on a Frozen World in Pictures
-
View User Profile
-
View Posts
-
Send Message
Curse PremiumPost the logs that you got this message from please. Preferably in a spoiler.
Former everything.
How to get dxdiag
Apologies, but, "this message?" What message?
My short story-like journals; quick-and-easy reads:
My Quest for Elytra Complete! (Pic Intense, End-Game Spoilers)
[Journal & Pics] After a Year and a Half, I Finally Found a Jungle
FrozenCore: Hardcore Death; 3/20/15 to 5/3/15; Eight Weeks on a Frozen World in Pictures
So, as I said, everything was a fresh install. I pulled my massive, two-year-old SSP world over from backup and started playing, eager to get back into Minecrafting. That's when I encountered the problem. Messed with settings. A few cold boots. The problem lasted all day. Couldn't figure it out, so I posted here.
Just now, I started a new world and the problem did not carry over to it. Scared that there was something wrong with my old world, I tried playing it again . . . No problems.
As far as I can tell the solution to the problem was to start a new world . . .
Resolved, I guess.
My short story-like journals; quick-and-easy reads:
My Quest for Elytra Complete! (Pic Intense, End-Game Spoilers)
[Journal & Pics] After a Year and a Half, I Finally Found a Jungle
FrozenCore: Hardcore Death; 3/20/15 to 5/3/15; Eight Weeks on a Frozen World in Pictures
Are you sure about that? You should never, ever set Xmn to the same as Xmx or even higher than the default - it is just asking for lag spikes (specifically, it allocates memory for the "young generation", containing short-lived objects, of which recent Minecraft versions, since 1.8, are notorious for generating; the default of 128 MB was set by Mojang for a reason and is actually lower than what Java would otherwise allocate with the default 1 GB of RAM (typically, 1/3 of Xmx, so 128 MB would go with just 384 MB of RAM). I don't even use the default arguments myself; I replaced them with -Xmx768M -Xms786M -Xss1024K (the last particular to 32 bit systems to avoid stack overflow crashes).
Also, as for why a new world would seem to fix the problem, the issue could be the way the game handles structure data; specifically, for mineshafts, which are absurdly common even since 1.7 (one every 250 chunks; in 1.6.4 it was one every 100 chunks; in both cases they are well over an order of magnitude more common than all other structures combined, as well as the largest and most complex) and due to the way the game stores the data in memory it uses over a hundred of times more memory than the file size of "Mineshaft.dat" would indicate:
MC-33134 Mineshaft.dat uses too much CPU
In fact, I even modded the game to prevent it from saving this data; my 3+ year old, 328 MB Survival world runs just as well as a brand-new world (maybe slightly worse around my main base due to 400+ tile entities, but the game does not lag). For vanilla you can periodically delete the file (there is no point in saving mineshaft data other than to ensure they generate correctly when world generation changes, and that would only work if they made them less common, as in 1.6.4-1.7.2, where the 40% of mineshafts that still exist are in the same locations). The only issue if if you explore a lot in a single session, as it automatically recreates the file as chunks are loaded (this is why they said to load structures you wanted to save in 1.6.4 before going to 1.7).
(I would do eveything I could to fix my world before I'd make a new world, the only world I've played on long-term or kept after I stopped playing it, and Minecraft would never be the same to me if it came to that)
TheMasterCaver's First World - possibly the most caved-out world in Minecraft history - includes world download.
TheMasterCaver's World - my own version of Minecraft largely based on my views of how the game should have evolved since 1.6.4.
Why do I still play in 1.6.4?
I've had no issues with the Xmn being set to the same as the maximum, but I'll remove that argument. Thanks.
Like I say, I made a new, unrelated world for testing, and after making that new world, the old world, which has nothing to do with the new one, now runs just fine. To me, it makes no sense at all, but I'm not worried about it.
Whatever the case, I now have no issues.
My short story-like journals; quick-and-easy reads:
My Quest for Elytra Complete! (Pic Intense, End-Game Spoilers)
[Journal & Pics] After a Year and a Half, I Finally Found a Jungle
FrozenCore: Hardcore Death; 3/20/15 to 5/3/15; Eight Weeks on a Frozen World in Pictures