HP-12C undocumented registers

[iMatt04bit]

Hello,

I am writing a Java HP-12C calculator sim, mostly because I wanted a fun exercise for learning Java, and because I really like HP calculators.

I have read the 12c manual and implemented the equivalent of registers in code, and these are working well. But I have the feeling that there may be some registers used during financial computations that are not described in the manual.

I am working on the TVM functions which are producing good precision as tested against other sims and by linear spreadsheet analysis.

If anyone has further insight they could offer on the inner workings it would be much appreciated, thanks,

Matt

[Paul Dale]

Getting the TVM functions correct in edge cases is quite difficult.

There was an article about the 12C which described some of these cases that Prof W Kahan figured out how to solve. Sorry, can't locate the reference at the moment

The 34S TVM library routine is pretty good and would be a good start for making these accurate.


- Pauli

[walter b]

(08-28-2014 07:54 AM)Paul Dale Wrote:  The 34S TVM library routine is pretty good and would be a good start for making these accurate.

Matt, you find that routine here.

d:-)

[Bill Zimmerly]

(08-28-2014 07:54 AM)Paul Dale Wrote:  There was an article about the 12C which described some of these cases that Prof W Kahan figured out how to solve. Sorry, can't locate the reference at the moment

Is this the article you're referencing, Paul?
- Bill

[Paul Dale]

(08-28-2014 09:37 AM)Bill Zimmerly Wrote:  Is this the article you're referencing, Paul?

That wasn't the one, but it contains some good examples nonetheless.


- Pauli

[Didier Lachieze]

(08-28-2014 10:13 AM)Paul Dale Wrote:  

(08-28-2014 09:37 AM)Bill Zimmerly Wrote:  Is this the article you're referencing, Paul?

That wasn't the one, but it contains some good examples nonetheless.


- Pauli

Maybe this one: Printing Financial Calculator Sets New Standards for Accuracy and Capability, HP Journal, October 1977, pp. 22–28;

Quote:The algorithms are based primarily on work done by Professor W. Kahan of the University of California at Berkeley.

[iMatt04bit]

Many thanks for the info, its all great material - I am reading through it all.
I like the analysis by Professor W Kahan - its enlightening.

Matt