Performance information for the MASS libraries for Blue Gene/L and Blue Gene/P

Content

The following table provides approximate performance data for the MASS libraries running on Blue Gene/P. (The results are also representative of Blue Gene/L MASS.) The data was obtained by timing many repetitions of a length-1000 loop (for scalar functions) or calls for length-1000 vectors (for vector functions) over 1000 random vectors and includes all overheads. Timing in this way brings the input and output vectors into the on-chip cache (because the loop is short enough for the vectors to fit in the cache). Performance may deteriorate significantly when the input and output vectors are not in the cache. Performance may also deteriorate for arguments at or near the end-points of the valid argument ranges. The system library measurements were made with the versions of the library available on the test system. They may vary from the versions timed for previous versions of MASS. Users may experience performance that differs from that found in this table.

The column labelled libm give the results from using the functions in the MASS C/C++ source code library libmassv.c to call the functions in the system math library. The C source code was compiled with the IBM Visual Age C/C++ compiler using the -O option. The columns labelled libmass and libmassv obtain the results using the libmass.a and libmassv.a library respectively. The columns labelled libmass speedup and libmassv speedup give the ratio of the corresponding entries from libm and libmass, libmassv respectively. Vector function names are used to indicate the corresponding scalar functions in libm and libmass. A blank entry in any column indicates that the corresponding function is not available.

The range key at the end of this section describes the ranges from which random arguments were uniformly generated.