|
|
In article <Xns### [at] 204213191226> , che### [at] sympaticoca
(Coridon Henshaw) wrote:
> Additionally, the slow memory leak caused by successful renders (reported
> in http://news.povray.org/povray.beta-test/21761/) is still present and is
> awaiting (another) confirmation.
While I do not want to claim there cannot be one small memory leak or two, I
should point out that the Windows task manager display (or a similar display
in any other OS) is by no means a tool to track or determine there is such a
leak. To the contrary, it is completely misleading!
A simple way to explain what you see there is the message window:
How could the added message text there be stored _without_ allocating memory?
Thorsten
____________________________________________________
Thorsten Froehlich
e-mail: mac### [at] povrayorg
I am a member of the POV-Ray Team.
Visit POV-Ray on the web: http://mac.povray.org
Post a reply to this message
|
|
|
|
"Thorsten Froehlich" <tho### [at] trfde> wrote in
news:3c6653c6@news.povray.org:
> While I do not want to claim there cannot be one small memory leak or
> two, I should point out that the Windows task manager display (or a
> similar display in any other OS) is by no means a tool to track or
> determine there is such a leak. To the contrary, it is completely
> misleading!
The usefulness of Task Manger for this purpose tends to depend on which
data column is used to determine an application's memory footprint...
In any case, I can confirm the same results with Sysinternals' Process
Explorer and the other diagnostic tools at my disposal.
> A simple way to explain what you see there is the message window:
I already took that into account by clearing the message window before
noting pvengine's memory usage. After discounting the message window,
pvengine leaks around 8kb after each render. Granted, this is peanuts
compared to the partial render leak that can eat tens of MB, but 8kb can
add up over a long animation.
Post a reply to this message
|
|