Howdy all, one more rendering using the technique I've described previously
here:
https://news.povray.org/povray.binaries.images/thread/%3Cweb.6156782017e7a53dd31149be3dfee7c%40news.povray.org%3E/

This was a test to push the limits a bit further. The fractal is known as a 4D
Xenodreambuie.The .stl file when converted to a mesh2 .inc is 1.9G in
size. I don't know how many polys as it was late and went to bed. I estimate it
took just over an hour to load the data.

I observed using htop process viewer that Povray, at least in v3.7, uses only a
single core to read and write files. In a 6 core i5 processor only core 6 was at
100%.

Cheers.

Post a reply to this message

Attachments:
Download 'zeno_inside.png' (341 KB)

Preview of image 'zeno_inside.png'

> I observed using htop process viewer that Povray, at least in v3.7, uses only a
> single core to read and write files. In a 6 core i5 processor only core 6 was at
> 100%.
> 
> Cheers.
> 

The reason is very simple : Reading or writing a file can't benefit from 
the use of multiple threads. It's a purely iterative process that can't 
get split in several threads.
The same hold true for parsing. Purely single thread.

Post a reply to this message

Alain Martel <kua### [at] videotronca> wrote:
>
>
> > I observed using htop process viewer that Povray, at least in v3.7, uses only a
> > single core to read and write files. In a 6 core i5 processor only core 6 was at
> > 100%.
> >
> > Cheers.
> >
>
> The reason is very simple : Reading or writing a file can't benefit from
> the use of multiple threads. It's a purely iterative process that can't
> get split in several threads.
> The same hold true for parsing. Purely single thread.

Yes and no. At least in reading SSD drives, it's possible to load buffers in a
multi-threaded "pre-fetch" mode. The rationale being that the main thread can
then switch between buffers and block load them to memory instead of single
thread process doing both reading and writing. Or, as the buffers occupy known
addresses and have known sizes, can be chained sequentially for reading. The
programming to accomplish it isn't trivial but it is doable. Cheers.

Post a reply to this message