Compiling NAMD

Hi folks,

preliminary tests have shown that FFTW runs very fast with PGI. Since this is a basic computing part of NAMD, I would expect a good performance for the whole code.
But alas, until now I have not managed to compile the code for a smooth run. Yet a couple of different compilers from other vendors had no problem with it.

Without further options I get “C++ runtime abort: internal error: static object marked for destruction more than once”. I have seen this message on the fixed-problem list for version 6.2.5 but even this version didn’t solve the problem.

We are running SLES 10 on AMD Opteron, PGI 6.2.5

Has anybody a workaround or a fix for this problem?

Thank You
Ulf

Hi Ulf,

Sorry for my late reply. I was only able to briefly look at this due to time constraints but have not completed an install. If you don’t mind, please indicate which version of NAMD you’re using and the steps you took to configure, compile, and run NAMD. Also, access to a data set would be useful.

Although we run the CPU2006 version on NAMD (version 2.5) daily, the error could be from a newer version or from the use of FFTW. I hope to have more time to investigate this later this week.

Thanks,
Mat