next up previous
Next: Results in Distributed Memory Up: Simple Communications Performance Measurement Previous: Message Startup Latencies

Results in Shared Memory Mode (Figure 5)

The mpiJava curve mirrors that of C with an almost constant offset up to 8K, thereafter the curves converge meeting at 256K. Under MPICH, the curves for C and mpiJava mirror each other in a similar fashion to those under WMPI, again there is a constant offset and convergence at around 256K.

Figure 5: PingPong Results in Shared Memory (SM) mode
\begin{figure}\centerline{\epsfig{figure=fig5.eps,width=135mm}}\end{figure}

Under WMPI the peak bandwidth of C is around 65 MBytes/s and mpiJava is 54 MBytes/s. The peaks occur at around 64K. Under MPICH the bandwidth is flattening out, but still increasing for C and mpiJava, at the 1M. The actual rate measured at this point is about 50 MBytes/s.

Clearly the WMPI C code perform best of those tested. The performance of mpiJava in SM under WMPI is good - it exhibits a fairly constant overhead of 95$\mu$s up to 2K, thereafter it converges with the C curve. The performance the C code under MPICH is slightly surprising as the NT and Solaris platforms used for these tests had similar specifications. It is assumed that the performance reflects the usage of MPICH rather than a native version of MPI for Solaris. Even so, the MPICH results for mpiJava show that it exhibits reasonable performance.


next up previous
Next: Results in Distributed Memory Up: Simple Communications Performance Measurement Previous: Message Startup Latencies
Bryan Carpenter 2002-07-11