Dear Molpro users: I was having the same weird problems with Linux and molpro in a cluster.
I wonder if you could send me the "/usr/src/linux/.config" file that fixed your parallel problems. I would like to check if there is something in the kernel configuration that can be useful to fix this parallel bottleneck. I want to test it first before replace or boot with one of those kernels. In my case this could be dangerous because the system is SuSE 7.3 distribution (kernel 2.4.19). Thanks in advance Regards, Rodolfo Briones Jerez Universidad de Chile On Mon, 19 May 2003 21:51:59 +0300 Gershom Martin wrote: > FOLLOW-UP: > > this appears to have been a Linux kernel issue. The problem disappeared > when we > upgraded the Linux kernels from 2.4.18-14 (the version released with > RedHat 8.0) > to 2.4.18-27.8 (the last RedHat erratum). Then we still had instances > where > a MOLPRO job would get killed in PBSPro but keep running, and require > manual > kill -9. This too disappeared after we upgraded kernels yet again, to > 2.4.20-13.8 (which was released for RedHat 8 on May 13). > > I wonder if the earlier problems Charlie Bauschlicher was having could > have been kernel issues as well? (RH 7.3 has an early 2.4.18 kernel). > > Regards > Jan Martin > > On Tuesday, May 6, 2003, at 09:30 Asia/Jerusalem, Gershom (Jan M.L.) > Martin wrote: > > > Greetings: > > > > we have a cluster of dual Xeon machines running (among others) MOLPRO > > 2002.6 built > > for mpp with the Global Arrays toolkit. > > As long as we were running RedHat Linux 7.2, all was peachy. Recently > > we have > > upgraded to RedHat 8.0, and MOLPRO jobs started hanging. (When it > > takes half a day to > > do an SCF with 180 basis functions, you get a bit suspicious.) > > > > More than that, the MOLPRO processes will not even die when "kill > > -9"ed: the machine > > needs to be rebooted. I tried rebuilding MOLPRO from scratch with the > > latest version > > of Global Arrays (using ifc 7.0 and Intel MKL): to no avail at all. > > The *scalar* > > version seems to behave, which suggests the problem is some > > incompatibility between > > Global Arrays and a component of RedHat 8 (possibly the 2.4.18 Linux > > kernel?). > > > > Did anyone else run into this? ANy suggested workarounds? > > > > Many thanks in advance > > > > Bets regards > > Jan Martin > > ----------------------------------------------------------------------- > > -- > > Prof. Gershom (Jan M.L.) Martin Computational Quantum > > Chemistry > > Department of Organic Chemistry / Sieff Building, Room 239 > > Weizmann Institute of Science / IL-76100 Rechovot / ISRAEL > > Email: [EMAIL PROTECTED] Phone: > > +972(8)9342533 > > Web site: http://theochem.weizmann.ac.il/ FAX: > > +972(8)9344142 > > ----------------------------------------------------------------------- > > -- > > *** Scientific research, no matter how 'pure' and useless it may seem, > > ** > > *** has an annoying habit of paying for itself many times, in the long > > ** > > *** run, in the form of greatly increased productivity. > > ** > > *** (Robert A. Heinlein, "Beyond this horizon") > > ** > > ----------------------------------------------------------------------- > > -- > > > > > ------------------------------------------------------------------ Jumping through hoops to get E-mail on the road? You've got two choices: Join the circus, or use Molly Mail. Molly Mail -- http://www.mollymail.com ------------------------------------------------------------------ Having trouble sending email from different locations ? Need a single outgoing mail server that will work from anywhere ? Set it to smtp.com and never have to change it again ! http://www.smtp.com ------------------------------------------------------------------
