I also read things said in wolfram forums. The issue is totally different. There is no way to change or play with options. In our case the integral is straight forward with no singularities at all! Something like this:
F[t_]:=NIntegrate[e^t ((a+b)-3a^2 )^8, {a,0,b},{b,0,3}]
Plot[F[t],{t,0,4}]
That's it. No singularities.. no cut-off's nothing. Pure and simple double integral. When you try with v6.0 it does it 5 or sometimes 7 times slower than v5.2.
Let's assume there is something we miss in these calculations.. ok?
Let's go back to Wolfram's built in function called Benchmark; to test how fast it handles 15 various problems. On both versions this command exists. You give a try on both.. you still come out with the result that overall Mathematica 6.0 is slower than Mathematica 5.2.
This is the reality of it. Maybe v6.0 while calculating it also thinks how to decorate the result and show it nicely with colors and so on.. But who needs to wait 5 times more for "loox" as you said, right?
I hope Wolfram will take care of this issue. Meanwhile I will stick to v5.2.
Or even 4.1 which is btw faster than 5.2 :)
Try before you buy, or better wait until 6.1 or so.
No comments:
Post a Comment