Romberg Integration for 33s, 35s

[Matt Agajanian]

Hello all.

I looked through the Arrticle Archives and the 34s port is a little bit hard to understand and it seems that the register exchange instructions are cryptic.

So, has anyone ported the WP-34S Romberg Integration program to the 33s or 35s? If so, could you let me know and if it's suitable, could you list the program and instructions, please? Thanks

Edited: 25 Mar 2012, 1:45 a.m.

[Paul Dale]

The code in the articles section is straight port of the 41C PPC integration routine. The PPC manual has extensive documentation on all its routines and this one isn't an exception.

Don't have the PPC manual? Buy the museum DVD set. It is on there.

- Pauli

[Matt Agajanian]

Got the DVD a coupla week, ago. Which item should I be looking for?

[Gerson W. Barbosa]

You should look for HP-41C PPC ROM Users Manual in Calculator Software Manuals.

[Matt Agajanian]

Sounds like a plan

[Nick_S]

Quote:

---

The code in the articles section is straight port of the 41C PPC integration routine.

---

Is this the same routine as the modified Romberg integration method used in the 15c?

Nick

[Paul Dale]

The 15c has a machine code implementation not a keystroke program, so they won't be the same. However the basic algorithm will be similar.

- Pauli

[Nick_S]

I gather that the 15c uses non-uniformly spaced samples to avoid problems with periodic integrands and that it usually does not evaluate the function at either limit of integration to reduce problems when the function is not defined at the one of the limits. Do you know if these features are also implemented in the PPC integration routine?

Nick

Edited: 26 Mar 2012, 7:10 a.m.

[Paul Dale]

Non-uniform samples is a standard transform.

Not evaluating at the end points is part of the algorithm I believe.

- Pauli

[Nick_S]

Thanks for this information. I also came across the forum article by Les Wright which gives the wp34s version of the code here:

http://www.hpmuseum.org/cgi-sys/cgiwrap/hpmuseum/articles.cgi?read=1126

he says that "There are first two transformations of variable--the first a linear one to standardize the interval of integration to [-1,1], the second a nonlinear one so that points are sampled at uneven intervals and problems with some periodic functions are avoided."

I look forward to playing with this code on other platforms.

Nick

Edited: 26 Mar 2012, 10:01 a.m.