[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index] [List Home]
{Spam?} Re: [aspectj-users] AJDT and memory

Hi,

I've tried the new dev build with these options. These are the results:

On Friday 06 August 2010 00:56:18 Andy Clement wrote:
> Under bug 278496 I've gradually been working on this, the idea is to
> reduce the memory used for compile-time weaving in AJDT, using
> something similar to the strategy I used to reduce memory for Loadtime
> weaving.
> 
> Then work as normal.  I'm interested in whether:
> - you see a noticeable reduction in heap usage (maybe turn on the
> eclipse heap status monitor under Window>Preferences>General)

Memory usage is down a bit (I'd say about 20%, compared to just 
minimalModel=true). The largest AjState object went down from 250M to just 
under 200M. Eclipse now takes just over 700M after a full GC, compared to 850M 
without type demotion.

> - you see a change in performance.  With the more eager eviction model
> these options employ, there may be more messing around to compile
> stuff  (probably only noticeable on full builds)

The build seems a bit slower, but not much. It's still fast enough.

> - it breaks !

Yes, it breaks a lot unfortunately. I had to click away about 70 errors. They 
all seem to be caused by a stack overflow:
at org.aspectj.weaver.ReferenceType.isAssignableFrom(ReferenceType.java:430)
at org.aspectj.weaver.ReferenceType.isAssignableFrom(ReferenceType.java:399)
at org.aspectj.weaver.ReferenceType.isAssignableFrom(ReferenceType.java:430)
at org.aspectj.weaver.ReferenceType.isAssignableFrom(ReferenceType.java:399)
etc.

Unfortunately, I can't dive into this deeper now and won't have time to test 
this the next couple of weeks, but I'll check back on this in september.

Good luck with this, it seems you're on the right path. With these options 
disabled, memory usage goes up by 500M!

Best regards,
Emond Papegaaij