[argyllcms] Re: colprof Not enough memory to process in chunks
- From: Graeme Gill <graeme@xxxxxxxxxxxxx>
- To: argyllcms@xxxxxxxxxxxxx
- Date: Sat, 16 Aug 2008 00:50:50 +1000
Fabrizio Levati wrote:
I was just wondering if you can explain why I get so different results
jumping from PC
No, I've not seen such behavior on my machines, as it seems to
be data set specific.
Is it the A2B table creation that is taking all the time,
or the B2A table creation ?
to MAC
About to optimise input curves and matrix again
............................................................................
............................................................................
........................colprof: Warning - xfit_fit: Conjgrad failed to
converge
Note that this is strange, the A2B table creation proceeds quite differently,
taking a lot more iterations, and ultimately failing to converge. Normally
one would expect similar behavior between MSWindows and OS X, although
there may be slight differences due to different roundings due to compiler
differences, but note that the final self fit errors are identical.
Some quick checking using MingW to compile the code indicates that this
problem is correlated with using the GCC compiler, so it's possibly a compiler
quirk or bug. I'll see if I can figure out what GCC is doing that messes things
up.
Graeme Gill.
Other related posts:
