Re: multispectral file access

Hi again Bob,

On 1 September 2014 20:09, Bob Friesenhahn <[email protected]> wrote:
> Testing shows that for a 5000x5000 TIFF file, most of the time spent
> in the vips benchmark is spent reading/writing the TIFF file (both
> reading/writing are similarly slow), with hardly any time being spent
> in the crop and convolution algorithms.

I added a ppm version of the test, but it seems to only make a small
difference on my laptop:

ppm-gm, 1.13, 245.7890625
gm, 1.17, 246.78515625

That's gm-1.3.18 with the MMAP env var off. It seems very slightly
slower with MMAP.

It's not a very rounded test of vips either: 70% of CPU time
(according to cachegrind) is spent on a single line of code in the
bilinear interpolator.

John

------------------------------------------------------------------------------
Slashdot TV.  
Video for Nerds.  Stuff that matters.
http://tv.slashdot.org/