Sir Isaac Newton was certainly wrong when he (or some one before him) postulated that time flows at a pre-determined, uniform pace…. And though contrary of this statement/postulate has been my experience on more than handful of occasions, it has rarely been bitter than in the past three weeks. Time during this period has flown past in a jiffy and what makes it bitter is that I have almost nothing constructive progress to show/account for Barring about four days - when I had to change apartments during an arctic breeze while not feeling so very well (thanks to all those who helped) - I think I have spent rest of the time in trying to make SMEAGOL work. After some hacking and commenting certain chunks of source code in a certain some file, I did manage to get the program to compile properly. Not only that, I also managed to run all the test calculations on single and multpile processors. But my happiness and success stops there - none of the calculations/simulations that we actually want to work on, can be run successfully.

Developers and some fellow users, in the course of many correspondences, suggested that I try different compilation options as well as different MPI implementations. But none of them - LAM MPI, MPICH1 MPI, MPICH2 MPI, OpenMPI and/or the ones that came installed as part of the Rocks Cluster Software - work I have, on the other hand, managed to get some other program to work successfully with MPICH2 MPI. But as far as SMEAGOL (and SIESTA 1.3 to a certain extent) is concerned, luck/success seems to be hiding somewhere. Adding to my misery, I cannot even get LAM MPI, MPICH1 MPI and OpenMPI to install properly on our cluster. Developers of the code suggest that we try MPICH1 MPI (or something else) while developers of MPICH suggest that we ask code-developers to make the code compliant with MPICH2! This is kind of a never ending circus.

After typing and re-typing the same set of commands over and over again, my fingers - without my own knowledge - automatically start typing make clean and make as soon as they come in contact with a keyboard. Our cluster configuration is as follows: Pentium 4 front end, running Red Hat Enterprise Linux 4.x on top of Rocks Cluster Software 4.0. If any of you have any idea about this problem, help me and please put me out of my misery… please……….