Is it possible to reduce Spark's memory footprint ?
Most IM software take 5-15MB of memory, but Spark need over 40MB !
Not many people are willing to keep such a bloat software running in the background.
sake,
Almost all of that memory is coming from Java itself. To some degree, that's the price you pay for having a rich cross-platform application. However, there are still several things that can be done to reduce memory usage:
Further optimizations in Spark should shave shave off exra MB. We're working on some of these now.
There are some good improvements to memory usage coming in the next version of Java, which will ship this fall.
New computers always have more and more RAM. ![]()
Derek asked me what I thought the biggest difference will be between having Spark be open source, and having a HOW-TO development guide. Here are my thoughts:
My first reaction is I can "just go do it" when it comes to making a change, I dont have to wait for overworked develpers to accept my change. But it really goes beyond that too. When I contribute code to an open source project, there is a feeling of ownership and pride that goes along with it. I think this helps develop a community around a product too, since it starts breaking down the strict producer-consumer model most software companies have. We have already seen examples of this with Wildfire:
Community members can contribute great new ideas
Community members can submit fixes with bug reports
Community members can get a better understanding of the limitations and strengths of the software
The key word there is "can." Not all will, there will still be many people who "leech" but that really isnt any different than the current producer-consumer model. By allowing the community complete access to the software, and even "free" ownership in it, the community will become more proud to use those products.
I for one am glad to see a product like Spark go open source. I think that means great things will come to it, and the company that provides care and feeding to the project.
It is good news for me. So I can customise according to my corporate needs
Congratulations to everyone at Jive! You've definitely earned it by releasing top quality software.
I watched the memory footprint in the XP Task Manager as I triggered different events in Spark.
1. Startup with autologon: 37,100k (no windows dispayed, tray Icon only.
2. Open the client window by clicking on the tray icon: 39,140k.
3. Close the client window by clicking on the "X": 39,456k
4. Reopen the client window: 39,844k
5. MINIMIZE the client window: 2,936k
6. Restore the client window: 15,548k
I don't know much about the JRE, but I did notice that it appears to clean up nicely when all application windows are minimized, but it does not perform the same cleanup when the client is minimized to the system tray.
J.D. taking less memory when minimized is common for all windows programs. I think this is achieved by moving most of date to swap. Which is slow and takes longer for program to get up. If program is small and fast you dont see difference so it's nice when it takes less memory minimized. But with java, i can even prefer it to take 40Mb all the time and dont use swap ever. I like when my apps are reacting quick after long idle time. So i dont have to wait even a few seconds.
Ignore the "Mem Usage" column in Task Manager. Enable the "Virtual Memory Size" column and use that.
If Spark uses 40mb of RAM, I am presuming that this wouldn't be a good option to have running on a Citrix server. The majority of our users have WinTerms and run all applications off of Citrix. Any opitions I have?
sbaier -- if for some reason Spark doesn't work in your environment currently, you always have the option of using another client that implements the XMPP protocol. There's a large list of clients available at http://www.jabber.org and they generally all work great with Wildfire.
A nice to use IM, easy and very powerful, so guys can u please keep me updated with new techs...?
hi,i got a question about how to confige the asterisk Plugin.
client: spake.
server: wildfire.
XMPPConnection conn = new XMPPConnection("myserver.foo.com", "user", "password");
PhoneClient client = new PhoneClient(conn);
which file should this code be put in?
thanks.
Asia
China
Is it possible to reduce Sparks memory footprint ?
Most IM software take 5-15MB of memory, but Spark need over 40MB !
Not many people are willing to keep such a bloat software running in the background.
I too run a Citrix type network, and 50 users running spark @ 39-45MB RAM, About 2GB's - Can't happen... Any find a Lite client thats like spark just not much a memory pig?
great