I'm usign PAMGuard version 2.02.03. As I'm working with a huge dataset, everytime I scroll the window (both user display and click detector display) PAMGuard runs out of RAM and it crashes. I read from a previous answer (Maximum wav file size for reasonable processing in PAMGuard) that I can change the RAM allocated from the .ini file of the viewer mode, and I did it. But seems it didn't change really the RAM allocated for JAVA to run. I also restarted the computer, so every change should have been applied. What else can I do to increase the RAM used by Java?
-
$\begingroup$ Maybe user213 could answer this. $\endgroup$– sm1Commented Jan 12, 2023 at 14:59
-
1$\begingroup$ What value did you put for the .ini file and what is the RAM for your computer? Also, have you looked at the task manager when PG is running to see how much memory PG is actually using? $\endgroup$– seleneCommented Jan 13, 2023 at 2:14
-
1$\begingroup$ I added at a little example .ini text to the answer you linked to above to compare! bioacoustics.stackexchange.com/a/226/42 $\endgroup$– seleneCommented Jan 13, 2023 at 2:22
-
2$\begingroup$ Thank you, Selene. I already did that passage, and set a higher value than the default 4096. But it didn't change, for some reasons. I then discovered that the problem was with the dataset, it was corrupted. So, it was trying to use an indefinite amount of RAM that was not available to PAMGuard. I still don't know how to change it, but seems it's not necessary. Maybe I should close the question? $\endgroup$– Cristina MarcolinCommented Jan 14, 2023 at 11:42
-
2$\begingroup$ Ah ok so it was a file issue rather than a PG issue? I think it still might be helpful for people with the same issue so I’d suggest you ‘answer’ your own question and just state what you said in the comment above! $\endgroup$– seleneCommented Jan 15, 2023 at 16:12
1 Answer
It turned out that the problem with the RAM was due to an abnormal use of the RAM allocated to PAMGuard. When I used a backup version of the database, the problem disappeared. Hence, I discovered that it was a problem with the dataset, that was corrupted for whatever reason. As corrupting a dataset with PAMGuard can be quite easily, if someone sees an abnormal drainage in the memory, try to use an older version of the dataset. It's probably the reason. Even with large dataset, the used memory should never be above around 1000 MB.