HPC performance comparisons are always tricky


ARCHER vs ARCHER KNL: node to node comparison from Adrian Jackson blog
13 Apr 2017 Edinburg - Performance comparisons are always tricky to get exactly right, says Adrian Jackson in his EPCC blog "Apple vs oranges: performance comparisons". Some careful thought is needed when doing anything other than comparing two different versions of the same application on the same hardware with the same test case, he says. Even then, it's worth thinking about what performance you are really interested in. Usually it is actually time to solution rather than scaling or some other metric like that. He uses some detailed illustrations on the ARCHER supercomputer.
Ad Emmen