[Opera-Linux] OK, what can I do about Opera's never-ending thirst for memory?!

[Eirik Byrkjeflot Anonsen]

Daniel Eckl <daniel.eckl at gmx.de> writes:

> Hi Kenny!
>
> I think I am puzzled by all these values... I just read 16,5MB and
> 8,4MB? Where do I fail in interpreting the writings?
>

Yes but that is memory used by the X server process, not by the opera
process.  So it is just indicative of there being a difference.  What
the side effects would be on the opera side of things is unknown.  (In
principle, there should be no difference, but who knows...)

eirik

>
> 2009/6/18 Kenneth Crudup <kenny at panix.com>:
>>
>>> Also, I'm running, on the machine only running XVnc headless, a clone of
>>> the very ~/.opera dir that's on my laptop, and will be checking for memory
>>> usage on that box as well as I surf on it.
>>
>> Get this- the VM footprint of this setup has *decreased* from 1.6GB to
>> just over 1GB; the actual resident is staying consistent at ~800MB.
>>
>> The NVidia-running version has gone from ~900MB to now at 1.7GB VM and
>> resident of ~980MB.
>>
>> In either case the test isn't as thorough as normal yet 'cause I'm
>> working, which isn't consistent with surfing sites that have, er,
>> lots of pix in 'em. :)
>>
>> More later, but its really starting to look like the NVidia Binary blob
>> is if not the reason, a conduit to this problem. Later on I'll look at
>> some "xrestop" snapshots (running that constantly slows down X). Here's
>> a snapshot of that, though, which implies it's not using as much mem
>> as before:
>>
>> ----
>> res-base Wins  GCs Fnts Pxms Misc   Pxm mem  Other   Total   PID Identifier
>> 4400000  1519   29    1 1603  696     8448K     53K   8501K 11044 Java apps (JDownloader) and nvidia 180.* AND 185.* pro
>> ----
>>
>> compare to before:
>> ----
>> res-base Wins  GCs Fnts Pxms Misc   Pxm mem  Other   Total   PID Identifier
>> 4400000  1486   22    1 1371  652    16403K     51K  16455K 11044 Index of /unix/10-unite/x86_64-linux - Opera
>> ----