Stability and Memory Pressure in 8.2

[Gary C Martin]

On 10 Sep 2008, at 08:19, Marco Pesenti Gritti wrote:

>> 711
>> [olpc at xo-05-24-02 ~]$ pmap 1676
>> 1676:   python /usr/bin/sugar-shell
>> --- snip ---
>> 0804a000  13888K rw---    [ anon ]
>> --- snip ---
>> total    54988K
>>
>> 759
>> [olpc at xo-05-24-02 ~]$ pmap 1316
>> 1316:   python /usr/bin/sugar-shell
>> --- snip ---
>> 09865000  18996K rw---    [ anon ]
>> --- snip ---
>> total    62188K
>>
>> On cursory glance through the other lib's shown by pmap linked in,  
>> I don't
>> see any obvious monster deltas in 759 vs 711. Sure, most of the  
>> libs seem to
>> have grown a little in size, but the single, unfortunately opaque,  
>> [anon]
>> block seems to hold the majority of whatever the mem delta between  
>> 711 and
>> 759 is (as reported by pmap).
>
> Can you try to apply the following patch and see if affects the anon  
> block?
>
> diff --git a/src/sugar/graphics/style.py b/src/sugar/graphics/style.py
> index 1b510d1..90d8565 100644
> --- a/src/sugar/graphics/style.py
> +++ b/src/sugar/graphics/style.py
> @@ -124,8 +124,8 @@ XLARGE_ICON_SIZE = zoom(55 * 2.75)
> FONT_SIZE = zoom(7 * _XO_DPI / _get_screen_dpi())
> FONT_NORMAL = Font('Bitstream Vera Sans %d' % FONT_SIZE)
> FONT_BOLD = Font('Bitstream Vera Sans bold %d' % FONT_SIZE)
> -FONT_NORMAL_H = _compute_font_height(FONT_NORMAL)
> -FONT_BOLD_H = _compute_font_height(FONT_BOLD)
> +FONT_NORMAL_H = 10 #_compute_font_height(FONT_NORMAL)
> +FONT_BOLD_H = 10 #_compute_font_height(FONT_BOLD)
>
> TOOLBOX_SEPARATOR_HEIGHT = zoom(9)
> TOOLBOX_HORIZONTAL_PADDING = zoom(75)

Nothing conclusive I'm afraid after testing the variation several  
times. It didn't make much impact to the [anon] block I mentioned  
(shrank by ~100-200Kb) though sugar-shell resident did seem to drop by  
~1Mb according to top (see below). But overall (with the 5 activities  
launched), the changes actually seemed to increase the total buffers/ 
cache value (as shown by free) by ~3-6Mb (up to ~174M).

/usr/bin/sugar-shell
before change -> 15.3% (RES=35m, SHR=14m, DATA=20m)
after change -> 15.1% (RES=34m, SHR=13m, DATA=20m)

Memory voodo...

--Gary