The Bandwidth



next up previous
Next: The Latency Up: The CM-5 Previous: Comparison of the

The Bandwidth

Figures gif(a) and gif(b) show the bandwidth obtained between two nodes of the CM-5 for

 
Figure: Bandwidth on the CM-5: PVM3 - CMMD  

The first point to notice in Figure gif(a) is that, as expected, the native message-passing library is the most efficient, with an asymptotic bandwidth of 8.06 Mbytes/sec. Nevertheless, the pvm_psend() - pvm_precv() bandwidth is fairly close to the performance of the native system. In fact, pvm_psend() is built on top of CMMD_send_async(), a configuration that explains the good performance (see section gif).

We also see that pvm_send()-pvm_recv() with PvmDataInPlace is much less efficient than pvm_psend() - pvm_precv(). Two factors explain this inefficiency:

Of course, pvm_send()-pvm_recv() with PvmDataRaw is even less efficient. It is built on top of CMMD_send_async but involves an actual data packing-unpacking.



Jack Dongarra
Thu Jul 20 07:22:58 EDT 1995