Re: X performance suffers under emacs 21.1.1

[Daniel Ortmann]

address@hidden (Gerd Moellmann) writes:

> Richard Stallman <address@hidden> writes:

>>     Or maybe one could implement another idea: record where the image
>>     was displayed, and remove images from the cache when the buffer
>>     they were displayed in is killed

>> That seems like a good idea.

> Here's another idea that's more precise than just acting when buffers
> are killed: one could remove images when their Lisp specs are
> unreferenced, like in a weak hash table.  That, maybe combined with a
> delay of some sort for the remote case (time and/or # of redisplays,
> or whatever) should work pretty well, I think.

> Code compiles and is untested; but I'm pretty sure it works.  Daniel,
> could you please give it a try?  It's a diff against the current CVS
> version, but there's a good chance it can be applied to other versions
> as well.

The patch applied, but does not seem to affect the memory cache.

Note X before:
  PID USERNAME PRI NICE  SIZE    RES STATE    TIME   WCPU    CPU COMMAND
18766 ortmann   35   0 16376K 13088K RUN      0:53 14.06% 14.06% emacs
  264 root       2   0 14448K  8776K select  14:29  0.24%  0.24% XF86_S3

And X after 35 minuts downloading over a slow link:
  PID USERNAME PRI NICE  SIZE    RES STATE    TIME   WCPU    CPU COMMAND
18766 ortmann    2   0 19728K 14068K select  11:07 31.05% 31.05% emacs
  264 root       2   0 37420K 21556K select  18:25 19.09% 19.09% XF86_S3

-- 
Daniel Ortmann, 2414 30 Av NW, #D, Rochester, MN 55901
address@hidden (h)   / 507.288.7732 (h)
address@hidden (w) / 507.529.3887 (w)