Fri Apr 24, 2015 12:49 am
I notice two things.
First, a bug. In about, it says the version is 10.0.0.0. But in system information, it says 10.0.2.0 (the recent Feb 3, 2015 one), which is also what the variables say. Presumably the about box should say 10.0.2.0 as well. At any rate, I appear to have the fixed version.
Second, I figured out from system information how to run the benchmark thing. It runs a sequence of 15 tests. I notice while most of the tests use only 25% CPU (i.e. 1 core), some use 50% (i.e. 2 cores, and some use 100% (i.e. all 4 cores). This leads me to believe that some operations are more suited to multicore computation than others.
How can I find out which operations run on multiple cores? Is there any way to influence this? Image processing should be well amenable to multicore computation, yet that uses only 25% for me, which is disappointing.