Re: [ccp4bb]: NFS Performance

[Anastassis Perrakis <perrakis@nki.nl>]

***  For details on how to be removed from this list visit the  ***
***          CCP4 home page http://www.ccp4.ac.uk         ***

> I believe that neither arp/Warp (I asked Victor about it) nor SOLVE (setting
> SOLVETMPDIR to a local directory does _not_ help) honour CCP4_SCR -  if they
> would the problem would probably not surface.

I completely agree with Kay that this is a g77 problem.
The current version of arp_warp, 5.1, is compiled with g77 that has this problem
in Linux.
I know that for sure, since if I compile with pg77 and run over nfs the problem
dissappears.

Being far from a g77/Linux guru I first hope that in preparation for release 5.2
the new g77 will NOT have this problem.
I will test that tomorrow, I need to get our Linux boxes running first - yes we
are new to Linux ... shame on us.
I also seem to recall that there exists a directive to overcome this (buffering)
problem. I cant remember it though (help?).

And yes, Solve and Arp/wARP do not honour the CCP4_SCR (for good reasons) and
they are not open source so you can correct it locally - but I will do my best
to solve this issu e- I am pretty sure its just compilation problem and not a
general ngs performace error.

I guess I will catch hell for that comment, but its my impression that DENZO has
the same problem, but you dont see it a lot since its only reading not writing
as well ...

    Tassos