|
|
On 2011-11-16 16:30, Le_Forgeron wrote:
> Le 16/11/2011 10:31, Roland Mösl a écrit :
>> On 2011-11-16 10:05, D103 wrote:
>>> =?ISO-8859-1?Q?Roland_M=F6sl?=<fou### [at] pegeorg> wrote:
>>>> Just wanted to render 1050 pictures for a scene.
>>>>
>>>> One notebook from 0 to 1049
>>>> The other from 1049 to 0
>>>>
>>>> Hope in 12 hours, they have finished somewhere in the middle
>>>>
>>>> But seems POV does not like this idea.
>
> Povray is not megapov.
>
>
>> I used in the past weeks up to 5 computers to render
>> different scenes.
>>
>> In this case, it's not so easy, because the first
>> 300 pictures contain time consuming reflexion and transparency effects
>>
>> One notebook is a 1,3 GHZ ULV, the other a 2,3 GHZ
>>
>> So it was difficult to quess, where would be the half.
>> So I had the idea, that they could met somewhere in the middle,
>> whereever the middle is
>>
>
> Instead of meeting in the middle, you could use +STPn
> (Same as Frame_Step=n) to split the load.
>
>
> On first machine, +KFI0 +KFF1049 +SF0 +STP5
> On second machine +KFI0 +KFF1049 +SF1 +STP5
> On third ... +KFI0 +KFF1049 +SF3 +STP5
> ...
> On fifth ... +KFI0 +KFF1049 +SF5 +STP5
>
>
> For your two systems' case, it obviously translate to
> On first machine, +KFI0 +KFF1049 +SF0 +STP2
> On second machine +KFI0 +KFF1049 +SF1 +STP2
>
>> http://wiki.povray.org/content/Documentation:Reference_Section_1#Animation_Options
>
> (+STP apply to 3.7RC3 (or later) version of povray, not the 3.6)
Can I assign more or less CPU-cernels for a task?
I did not change to 3.7, because it's somethimes better,
that only one cernel calculates with POV,
and the rest of the notebook is free for other tasks.
--
Roland Mösl
Post a reply to this message
|
|