[Top][All Lists]
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [Openexr-devel] scanline order
From: |
Charles Henrich |
Subject: |
Re: [Openexr-devel] scanline order |
Date: |
Fri, 7 Nov 2003 00:53:56 -0800 |
User-agent: |
Mutt/1.4.1i |
> performance depends, of course, on the particular order in which you are
> accessing an image's scanlines, but I would guess that making the scan line
> blocks smaller would make reading the file slower, because seekg() would be
> called more often. (When you call readpixels(), the probability that the
> scan line you want is in the block that is currently in memory, or in the
> next block in the file, increases with the block size.)
Hrm, I didnt realize the Ilm library was caching internally. So if any
scanline within the 32 line chunk is accessed, then all subsequent lines
should be available? Hmm. I'll have to do a little more digging to find out
whats going on. Does the library keep multiple cache's, or if you access
alternative lines (say something like last, first, last-1, first+1, etc..)
would that basically invalidate the cache and cause continuous decompression
of each 32 line block?
-Crh
Charles Henrich Digital Domain address@hidden
http://www.sigbus.com/~henrich