Intel MPI Benchmark woes on Itanium2

Intel MPI Benchmark woes on Itanium2

Firstly apologies if there's a thread already on this (I couldn't find anything with the limited search facilities available)

The crux: I've downloaded the Intel MPI Benchmarks (IMB) and compiled on Itanium-2 chips using mpich2/icc. However I have had several problems, see below, so was hoping to share experiences with others - let me know what you've found!

The problems so far:

1) MPI-IO benchmark suite hangs just after:

Benchmarking C_IRead_Indv
# #processes = 2
#-----------------------------------------------------------------------------
#
# MODE: AGGREGATE

2) -time N option either doesn't seem to have an effect (no more iterations allowed even if N is increased) or gives nonsense (eg if set N very large then o/p claims it's doing very large negative iterations or gives a sensible number of iterations but completes them in '-0.00' seconds). Whilst this is reproducable, for a given N, it doesn't happen on every instance which is perhaps more worrying.

thanks, Michael

--- @mkbane_mcr
1 post / novo 0
Para obter mais informações sobre otimizações de compiladores, consulte Aviso sobre otimizações.