To calculate the parallelization efficiency, it’s crucial that you use a mathematical equation called Amdahl’s Law.
Donald Kinghornto help us get established in the scientific computing market.
What is Amdahl’s Law?
to make it use this equation, you first need to determine the parallelization efficiency of your program.
Luckily, you don’t need to change out your CPU a bunch of times to do this.
Instead, it’s possible for you to simply set the program’s affinity through Task Manager in Windows.
Note that setting the affinity only lasts until the program is closed.
The next time you spin up the program, you have to re-set the affinity again.
In our example, for two cores the speedup is 645.4/328.3 which equals 1.97 .
For example, for 4 cores the equation would be:
This equals 2.5.
In our case, the actual fraction was .97 (97%) which is pretty decent.
As an example, lets use a Xeon E5-2667 V3 and a Xeon E5-2690 V3.
Anything different (even within the same program) may have drastically different results.
