Code Porting
Compiled below is a list of facts you need to be aware of when trying to compile, link, and run programs on the Linux Cluster:
- The Cluster consists of 52 1.3GHz Intel Pentium III processors.
If you have been running jobs
on a personal computer with a more recent and faster processor, these same
jobs, when run on a single processor on the Cluster, will likely
run slower. However, you could gain in throughput (relative to
your personal computer) if your code can take
advantage of the parallel processing available on the Cluster. How much
speed will you gain depends on a number of factors, including the clock speed
of your current machine, the parallel efficiency of your code, and how
many processors you use. If you are not familiar with parallel programming
and would like help in using our systems, please contact Kadin Tseng
(kadin@bu.edu) or Doug Sondak
(sondak@bu.edu) to discuss your
computing goals and resource needs.
- In addition, which of the three SCV systems may be more appropriate for you is another factor to consider. As a reference, we have a comparison of the performance of the
three SCV systems.
- Runtime limit is 48 hours.(More details ...)
- You can request up to 24 nodes per job.(More details ...)
- MPI is the primary means of
multiprocessing. The languages supported for MPI programming are Fortran,
C, and C++.
- Technically speaking, OpenMP is
supported on the Cluster. However, because each dual core PIII contains
only two processors, shared memory parallel processing using OpenMP alone is
not practical.
- The distributed-memory parallel mathematical library ScaLAPACK is available.
- If your program is written in C++, besides the headerfile
mpi.h, you will also need to include mpi++.h.
- All parallel jobs must run through the batch system.
- 64-bit addressing is NOT available on the Intel PIII.
- /scratch is available.
|
