[sharp-discuss] solvent flattening error in p61 and p6522

[Clemens Vonrhein]

Dear Stefan,

it seems the CCP4 distributed sftools binary for IRIX that is buggy. I
have no problem reading it on

  1) Linux - using either the CCP4 distributed binary or my own
     compiled ones

  2) True64/OSF1 - using either the CCP4 distributed binary or my own
     compiled ones

  3) IRIX - using my own compiled ones (f77 or f90)


The only one where I get the error

  !!! Error, too large indices (> 512) occur !!!
  !!! Fatal error while reading reflections !!!
            0 reflections read successfully
  Trying to clean up the mess
  Deleting new columns
  Deleting new reflections

is the CCP4 distributed IRIX binary. So I recommend to re-compile CCP4
5.0.2 for SGI/Irix yourself: good opportunity to also apply all the
patches from the problems page. You can make it easy for yourself if
you use my ccp4_cv.sh script.

I guess this particular MTZ file (spacegroup? Number of reflections?
sort order?) triggers this bug when others don't. Could be a simple
memory problem (tricky to hunt down though ...).

Cheers

Clemens


On Wed, Jul 06, 2005 at 03:31:48PM +0200, Stefan T. Arold wrote:
> me again...
> 
> I think this might be important:
> sharp solvent flattening seems to run fine for another dataset in SG 
> C2221! The problem really seems specific to p61. Could it be something 
> with the convention for the sort order being different for p61 in 
> mtzdump - sftools ?
> 

-- 

***************************************************************
* Clemens Vonrhein, Ph.D.     vonrhein AT GlobalPhasing DOT com
*
*  Global Phasing Ltd.
*  Sheraton House, Castle Park 
*  Cambridge CB3 0AX, UK
*--------------------------------------------------------------
* BUSTER Development Group      (http://www.globalphasing.com)
***************************************************************