WARNING! some R5000 indys have fp problems

[Gary Sheldon Thompson]

Hi 
	I thought you all might like to know about an incompatability between
some R5000 based Indys and xplor. We recently had problems with running
YASAP a nmr structure calculation protocol by micael nilges on only one
of our R5000 based Indys, the protocol failed with a large number of
shake errors at random points in the calculation. We have now tracked
this down to a hardware (floating point) problem with some R5000 based
processor (/ processor daughter boards), as a replacement of our
revision B processor board with a revision F processor board cured the
problem..... 

I hope this helps some people, as it took us several week and a spare
R5000 Indy that worked to track down this problem 

gary


-- 
*********************************************************************
                                       
 Dr. Gary S. Thompson           Address: 
 Wellcome Postdoctoral Fellow   Biochemistry and Molecular Biology, 
 Sheena Radford's Laboratory    Woodhouse Lane,    
                                University of Leeds,    
 Studying the structure and     Leeds, LS2 9JT.    
 folding of proteins by nmr.    England. 
                         
                                e-mail: garyt at bmb.leeds.ac.uk.
                                tel:    0113 233 3134
                                fax:    0113 233 3167
                              
*********************************************************************