0000217: World Gen consumes massive amounts of memory

(0000378)
DoctorZuber (reporter)
2010-04-02 19:53

(0000459)
bradb (reporter)
2010-04-03 02:14

Similar to 0000216 but this report states that the game crashes immediately. It might be that other reporter is not reaching 100% usage.

(0000747)
Someone_r_c_tm (reporter)
2010-04-03 19:21

This issue seems to be related to having graphics on for me. If I set [GRAPHICS:YES] in init.txt generating a world causes memory usage to rise to around 1.5GB. If graphics is set to NO memory usage only rises to about 110MB or so for me.

(0000775)
Holy Mittens (reporter)
2010-04-03 21:01

This seems to have magically disappeared for me. Even though I am still using graphics, I generated a world today and it didn't have a sudden jump in memory usage prior to world gen actually starting (at which point it began to rise, but not excessively).

(0000959)
Holy Mittens (reporter)
2010-04-04 17:08

Nope, it's back now. Turning off graphics does correct the problem though.

(0005701)
Footkerchief (manager)
2010-04-29 13:49

Is this problem still around in 31.03?

(0007398)
Footkerchief (manager)
2010-05-27 13:59

This seems to be the same problem as 0000046. Please continue any discussions there.

(0007406)
Logical2u (manager)
2010-05-27 18:47

If I may, I think this "bug" is more about there's too much memory usage in general in World Gen, leading to a possible crash. But 0000046 is specifically a memory leak after multiple world generations.

If anything, in my eyes 0000046 is the child of this bug.

(0007407)
Footkerchief (manager)
2010-05-27 19:14

I didn't see anything about multiple world generations in 0000046, and people in 0000046 were talking about high memory usage (even mentioning the same 1.5 GB figure) as part of the issue. 0000046 doesn't mention a crash, but there's no corroboration of a crash in this report, either.

Basically, these reports seem mostly identical but the discussion in 0000046 seems more productive (more memory usage stats etc).

(0007408)
Logical2u (manager)
2010-05-27 19:22

Well 0000046 is about the memory leak - they generated a world then quit the game and the process persisted, hogging the same amount of memory. (or too much). [I quote, "The process is taking 1.4gb of ram even after running another instance of DF. Required a "end process" to recover."]

This is just about they generated a world that used too much memory (and there were a few reports about hitting the 2gb limit and crashing, but you're right, neither of them are mentioned here, which I expected).

But yes, 0000046's discussion is better, I'm just seeing it as a child of "high memory usage" because 0000046 reads as "high memory usage persists after qutting" where as this is just "high memory usage". I'm not sure the distinction is super important, but it's just how I see it anyways.

Issue History
Date Modified	Username	Field	Change
2010-04-02 19:49	Holy Mittens	New Issue
2010-04-02 19:53	DoctorZuber	Note Added: 0000378
2010-04-03 02:14	bradb	Note Added: 0000459
2010-04-03 11:18	Footkerchief	Relationship added	child of 0000046
2010-04-03 19:21	Someone_r_c_tm	Note Added: 0000747
2010-04-03 20:37	Someone_r_c_tm	Issue Monitored: Someone_r_c_tm
2010-04-03 21:01	Holy Mittens	Note Added: 0000775
2010-04-04 17:08	Holy Mittens	Note Added: 0000959
2010-04-14 13:04	lord_chaos22000	Issue Monitored: lord_chaos22000
2010-04-29 13:49	Footkerchief	Category	General => Technical
2010-04-29 13:49	Footkerchief	Note Added: 0005701
2010-05-27 13:57	Footkerchief	Note Added: 0007397
2010-05-27 13:57	Footkerchief	Tag Attached: AWAITING UPDATE
2010-05-27 13:59	Footkerchief	Tag Detached: AWAITING UPDATE
2010-05-27 13:59	Footkerchief	Note Deleted: 0007397
2010-05-27 13:59	Footkerchief	Note Added: 0007398
2010-05-27 13:59	Footkerchief	Relationship replaced	duplicate of 0000046
2010-05-27 13:59	Footkerchief	Status	new => resolved
2010-05-27 13:59	Footkerchief	Resolution	open => duplicate
2010-05-27 13:59	Footkerchief	Assigned To	=> Footkerchief
2010-05-27 18:44	Logical2u	Relationship added	has duplicate 0001889
2010-05-27 18:47	Logical2u	Note Added: 0007406
2010-05-27 19:14	Footkerchief	Note Added: 0007407
2010-05-27 19:22	Logical2u	Note Added: 0007408
2010-06-09 06:45	Toady One	Status	resolved => closed
2010-06-29 07:38	Footkerchief	Category	Technical => Technical -- General

Notes
(0000378) DoctorZuber (reporter) 2010-04-02 19:53	0000216

(0000459) bradb (reporter) 2010-04-03 02:14	Similar to 0000216 but this report states that the game crashes immediately. It might be that other reporter is not reaching 100% usage.

(0000747) Someone_r_c_tm (reporter) 2010-04-03 19:21	This issue seems to be related to having graphics on for me. If I set [GRAPHICS:YES] in init.txt generating a world causes memory usage to rise to around 1.5GB. If graphics is set to NO memory usage only rises to about 110MB or so for me.

(0000775) Holy Mittens (reporter) 2010-04-03 21:01	This seems to have magically disappeared for me. Even though I am still using graphics, I generated a world today and it didn't have a sudden jump in memory usage prior to world gen actually starting (at which point it began to rise, but not excessively).

(0000959) Holy Mittens (reporter) 2010-04-04 17:08	Nope, it's back now. Turning off graphics does correct the problem though.

(0005701) Footkerchief (manager) 2010-04-29 13:49	Is this problem still around in 31.03?

(0007398) Footkerchief (manager) 2010-05-27 13:59	This seems to be the same problem as 0000046. Please continue any discussions there.

(0007406) Logical2u (manager) 2010-05-27 18:47	If I may, I think this "bug" is more about there's too much memory usage in general in World Gen, leading to a possible crash. But 0000046 is specifically a memory leak after multiple world generations. If anything, in my eyes 0000046 is the child of this bug.

(0007407) Footkerchief (manager) 2010-05-27 19:14	I didn't see anything about multiple world generations in 0000046, and people in 0000046 were talking about high memory usage (even mentioning the same 1.5 GB figure) as part of the issue. 0000046 doesn't mention a crash, but there's no corroboration of a crash in this report, either. Basically, these reports seem mostly identical but the discussion in 0000046 seems more productive (more memory usage stats etc).

(0007408) Logical2u (manager) 2010-05-27 19:22	Well 0000046 is about the memory leak - they generated a world then quit the game and the process persisted, hogging the same amount of memory. (or too much). [I quote, "The process is taking 1.4gb of ram even after running another instance of DF. Required a "end process" to recover."] This is just about they generated a world that used too much memory (and there were a few reports about hitting the 2gb limit and crashing, but you're right, neither of them are mentioned here, which I expected). But yes, 0000046's discussion is better, I'm just seeing it as a child of "high memory usage" because 0000046 reads as "high memory usage persists after qutting" where as this is just "high memory usage". I'm not sure the distinction is super important, but it's just how I see it anyways.