A lot of this has to do with the JVM preallocating memory. You should
play with the memory switches.
On Sat, 30 Dec 2000, Adam Langley wrote:
> On Sat, Dec 30, 2000 at 05:18:58AM -0600, Mark J. Roberts wrote:
> > AFAIK top understands all this.
>
> Yep, that's ok. Some tool don't however and sum all the values.
>
> From my machine (running all of 20 seconds):
> VmSize: 105312 kB
> VmLck: 0 kB
> VmRSS: 15532 kB
> VmData: 100252 kB
> VmStk: 260 kB
> VmExe: 16 kB
> VmLib: 4568 kB
>
> It's the VmData that is worrying here - that suggests actual
> malloc'ed memory is 100MB (shit). Maybe someone could single-step
> fred and watch the memory go up?
>
> (and Whiterose - sorry couldn't resist ;)
> VmSize: 3364 kB
> VmLck: 0 kB
> VmRSS: 1348 kB
> VmData: 1064 kB
> VmStk: 8 kB
> VmExe: 424 kB
> VmLib: 1628 kB
>
> (5 minutes later)
> It might not be fred's fault at all:
>
>
> class MemTest {
> public static void main (String[] args) {
> System.out.println ("Sleeping");
> try {
> java.lang.Thread.sleep (9000000);
> } catch (java.lang.InterruptedException e) {
> }
> }
> }
>
> gives:
>
> VmSize: 88732 kB
> VmLck: 0 kB
> VmRSS: 6152 kB
> VmData: 83812 kB
> VmStk: 260 kB
> VmExe: 16 kB
> VmLib: 4444 kB
>
> 83MB data without doing anything! (and 5 threads I might add). Shit,
> that's a bloated JVM:
>
> java version "1.3.0"
> Java(TM) 2 Runtime Environment, Standard Edition (build 1.3.0)
> Classic VM (build 1.3.0, J2RE 1.3.0 IBM build cx130-20000815 (JIT enabled: jitc))
>
> AGL
>
> --
> There is no grief which time does not lessen and soften.
>
_______________________________________________
Freenet-dev mailing list
[EMAIL PROTECTED]
http://lists.sourceforge.net/mailman/listinfo/freenet-dev
