Any tips on dealing with the garbage collector that slows everything down?
could this 'garbage collector' be what is causing my massive pauses every time I make an edit to a very large MM? Every time I make an edit, or just even select a differnt node, Freemind hangs/pauses for about 25 seconds, consuming 100% of processor power for that duration, then becomes usable for a few moments...then during edits or right after edits, hangs again. This started as a 305 second delay, but has been increasing and is now making freemind virtually unusable at 20-20 seconds delay per edit!! This doesn't happen if I make anew MM and keep it small, but I have a very large MM now (3.08MB) and I think that this is somehow the cause of my problem.........
You have failed to provide any information in order to solve your problem such as Java version, RAM, operating system, computer speed, ....... Basically you left us nothing to go on.
If you are using Windows XP, that could be your problem. Otherwise I do not know. I see slow performance on XP, but if flies on my linux box.
I am having pause problems on large (<1Mb) MMs.
I am using a Dell 2GHz with 1.5 MB of RAM.
I am using JRE 1.5.0_08.
I am working with several very large text-only mindmaps. Sometimes I will have several maps open in a single instance of Freemind, and other times I will have the maps open in their own instances.
I have found Freemind to start slowing (pausing as I move using the cursor) down at around 1M of file size.
Additionally, I have found that when I leave Freemind open overnight, that it takes several minutes to wakeup and sometimes I have to kill the process and restart it.
Is there anything I can do to get around this?
BTW: I really love this tool. I joined SourceForge because of it and have given it to over 60 people. Please keep up the good work.
i can confirm this is a problem. either it is unavoidable because a large amount of nodes (in MM size >1MB) really take up a huge amount of memory + java overhead = crazy windows swapping, or the code is not very good.
my configuration is 1.7 P-M, 1.5 GB RAM, XP SP2, JDK 1.6. i do run plenty of applications 24h concurrently, including firefox with about 4-5 windows and 30 tabs total, thunderbird, sunbird, acrobat reader, skype with min. 15 open chats, msn live msgr, pidgin and few other apps. is 1.5 GB really not enough? anyone care to share experiences with RAM configurations up to 4 GB? is that enough for running significant java apps (freemind, netbeans) alongside other software without significant swapping?
I have a Windows 2000 Pro. SP4
with 1570MB ram, 2.4GHz processor,
JRE Version 1.5.0 (build 1.5.0_06-b05)
I assume you either have an installation problem or the problem is running Windows 2000. I have seen problems with XP, but under Fedora Core 4, freemind works great, even on large mindmaps.
alos, splitting up the massive mindmap into individual MMs eliminates these delays: this correlation between total file size and delays/hang/pauses implies a bug along these lines...perhaps that garbage collectors thing mentioned above? I don't know but I do know about the pauses//delays I've written about...I will be happy to work on troubleshooting this or IDing the cause(s) if one of y'all is interested..I can follow instructions reasonably well...
yeah. so i signed up for an account here @ sourceforge so I can be a more useful part of this place(which has already helped me enormously..)
zirkonyx is my username
i am the one who couldn't get the borwser to work, and the one posting here about the strange delays which corrlate to MM size. looking forward to working with you guys on such things...
(I got the browser to work; that was just an implementation problem) but these delays are puzzling me and frustating me quite a bit.
I have seen this problem even with moderately sized memory maps on my laptop, back when it had only 512M of memory. I've upgraded to 1.5G, and now most things fly, including FreeMind. But I still notice slowdowns if I leave the machine on overnight. I strongly suspect that what happens is all the FreeMind code and data get paged out to disk so the dozens of threads that Windows runs, whether needed or not, can use the resources. When you try to start up again in the morning, you have to wait for Windows to swap all the ohter stuff back out and reload the FreeMind code and data. This is also probably why it works better on Linux. Linux doesn't have quite so many extra processes, and I suspect it had better memory management.
Log in to post a comment.
Sign up for the SourceForge newsletter:
You seem to have CSS turned off.
Please don't fill out this field.