[wp34s] New integration program

[emece67]

Find attached my findings when comparing this new program against the built in integration program.


  Tanh-Sinh_comparison.pdf (Size: 21.23 KB / Downloads: 72)

Finally I used a battery of 40 test integrals of various kinds.

On this table:

• "Better" means that the new program is more precise and faster
• "More precise" means that it is more precise but slower
• "Faster" means that it is equally precise but faster
• "Equal" means that it is both equally precise and fast
• "Slower" means that it is equally precise but slower
• "Less precise" means that it is less precise but faster
• "Worse" means that it is less precise and slower
  

When counting the correct digits I was only concerned about the 12 visible digits (if the expected result is 0, the table shows the absolute value of the exponent of the returned result). Time was measured with a handheld chrono, so don't expect high accuracy on such. Times marked as ">xxx" means that I stopped measuring the time when xxx seconds get reached. "est. time" means that I estimated the time required for the built-in integrator by measuring a few iterations on the real machine and using the simulator to get the number of total iterations needed. The machine was always in SLOW and DBLOFF modes.

Regards.

NOTE: updated to show results of last v1.0l version. Some typos fixed.