Differences

This shows you the differences between two versions of the page.

--- scaling:start [2018/12/19 15:05]
ccrosby
+++ scaling:start [2021/12/09 16:42] (current)
@@ Line 61: / Line 61: @@
 These results are best interpreted in the form of a graph:
+{{:scaling:scalegraph1.png|}}
+This graph looks OK (the run time comes down as the number of nodes goes up), but is actually quite difficult to interpret, in terms of finding the optimum number of nodes to use.  It is much more useful to plot the reciprocal of the runtime, such as the number of runs per hour:
+{{:scaling:scalegraph2.png|}}
+This clearly shows that the application is scaling linearly.  Adding more nodes will almost certainly make it run faster.
+The following two graphs illustrate a case which does **not** scale well.  The runtime graph does not look as good as the first example, but it is not immediately clear where the scaling starts worsening.  The "Runs per hour" graph provides much better insight.
+{{:scaling:scalegraph3.png|}}
+{{:scaling:scalegraph4.png|}}
+In this case there is definitely no point in trying to use more than 8 nodes, and the scaling is already showing diminishing returns after 2 nodes.  In this case it may be more efficient to use fewer nodes per run, but submit more simultaneous runs.

wiki.chpc.ac.za