A colleague has a problem :
When he runs a code with Mathematica 9 on i7-2600 cpu (4 cores / 8HT) it takes 1 min.
When he runs the same code with Mathematica 10 on i7-4810 cpu (4 cores / 8HT) it takes 30 min !
Is it normal to have a such computing time difference for the same calculation as the cpu is quite similar ?
If no, how to fix it ?
I thought it could come from a well known bug with multi-core and mathematica 10…
anyway, here is the code in attached file.
it takes 121 secondes to evaluate the entire notebook with mathematica 10.0
and on the SAME computer, it takes 5 secondes to evaluate the entire notebook with mathematica 9.0
the PC is a HP Elitebook 840 notebook / windows 7 SP1 64b / i5-5300U @ 2.3GHz CPU / 16 GB RAM
Though I am not quite sure about why you observe the difference in the speed, you may use the modified notebook to speed up your plot. One options is to use the PerformanceGoal option and the other is to use the ParallelTable to distribute the data generation over several kernels.
I’m not expert in mathematica but if your solutions improve a lot the speed, they do not produce the same result, unfortunately.
if you make a zoom around 0 you will see the that the plots with your methods gives much less details than the original plot…
I changed also mathematica 10.0 to 10.3 and the computation time went from 81s to 67s without changing a line of code…
then, it’s already better… but v9.0 is still better…