Announce: HP-15C Simulator / Release 3.3.00, Build 5403

[Torsten]

I am pleased to announce a new version of my HP-15C Simulator for Windows, Linux and Mac OS X: Release 3.3.00, Build 5403.



It can be downloaded from the simulator home page.

[Namir]

Thank you for the new version. Excellent work!

Namir

[Jeff_Kearns]

Torsten,

Great simulator indeed. I do have a question regarding internal accuracy of calculations. When I run the example in the TVM program, I get the following:

Torsten HP-15C Simulator: 333,783.2505
Official HP-15C Emulator: 333,783.3508
Real HP-15C and HP-15 LE: 333,783.3508
Swiss Micro DM-15: 333,783.3508
Correct answer: 333,783.349948 (HP-30b, HP-19BII and HP-17BII)

What do you believe causes this discrepancy? Can it be identified and addressed in a subsequent release? I actually prefer the interface of your simulator over the official emulator (even if it is a bit slower).

Regards,

Jeff

[Jeff_Kearns]

Mike,

My point is that the simulator is 'incorrect' by $0.10.

Jeff

[Gerson W. Barbosa]

Great tool and nice listings! Thank you very much!

Code:

# --------------------------------------------
# HEWLETT·PACKARD 15C Simulator program
# Created with version 3.3.00
# --------------------------------------------
# --------------------------------------------

   000 {             } 
   001 {    42 21 11 } f LBL A
   002 {       43 11 } g x²
   003 {       43 36 } g LSTx
   004 {          40 } +
   005 {           2 } 2
   006 {          10 } ÷
   007 {       43 32 } g RTN

# --------------------------------------------

[Torsten]

Jeff,

I can not confirm your result of 333.783,2505 for my simulator.

I have tested it on Windows 7 64bit and XP 32bit. In both cases, I got the "correct" result of 333.783,3499484285. The simulator uses the full 17 digits available in IEEE format.
If "Strict HP-15C behaviour" is checked in the "Preferences", the result is the same as with the hardware models: 333.783,3508. In "strict mode", the precision is reduced to the 10 digits of the emulators using a copy of the original ROM.

Torsten

[Torsten]

(01-20-2014 09:15 PM)Gerson W. Barbosa Wrote:  Great tool and nice listings! Thank you very much!

Did you see the HTML export? It's even nicer ;-)

Torsten

[Gerson W. Barbosa]

(01-20-2014 09:24 PM)Torsten Wrote:  

(01-20-2014 09:15 PM)Gerson W. Barbosa Wrote:  Great tool and nice listings! Thank you very much!

Did you see the HTML export? It's even nicer ;-)

Torsten

Indeed!



And that's only the listing! Thanks again!

Gerson.

[Jeff_Kearns]

(01-20-2014 09:19 PM)Torsten Wrote:  I can not confirm your result of 333.783,2505 for my simulator.

Torsten:

Thanks for your reply. I re-checked the program listing and it was accurate. I checked the "Strict HP-15C behaviour" and still it gave the same result: 333.783,2505. I was (or so I thought) meticulous in conducting the test the first time.

Then I verified the storage registers and 0.089633601 instead of '0' was found in register 3. I saved 0 in register 3 and it then gave the correct result. I will assume that I erred and I apologize for raising any concerns about your excellent simulator. It is a very impressive piece of work indeed. Not only that, it is even more accurate (in that it more closely matches the results of the financial calculators) when the "Strict HP-15C behaviour" is left unchecked.

Regards,

Jeff

[RMollov]

(01-20-2014 09:24 PM)Torsten Wrote:  Did you see the HTML export? It's even nicer ;-)

Torsten

Is it me, or RollUp icon in HTML has rectangle after 'R' rather than an up arrow?
Great job otherwise, cheers!

[Torsten]

Quote:Is it me, or RollUp icon in HTML has rectangle after 'R' rather than an up arrow?
Great job otherwise, cheers!

This is a font issue. You do not have a font on your system, that has the character. I recommend to install the DejaVu Fonts. There are more details in the simulator help files under "Usage | Preferences | Font Settings".

Torsten

[RMollov]

(01-21-2014 10:42 AM)Torsten Wrote:  This is a font issue. You do not have a font on your system, that has the character. I recommend to install the DejaVu Fonts.

Torsten

Thanks a lot, it all works perfectly now.
Cheers,

[Ed Wright]

Torsten,

I finally got a chance to spend some time using the simulator. Great work! Thank you.

[RMollov]

Solver problem.
Corrected after Walter's findings.

EQN:

R*( TAN (ACOS (R/(R+h))) - ACOS(R/(R+h)) ) - dL/2 = 0

R= 6,371,000m (stored in register .0)
dL=1m (stored in register 0)

h=?

Code:


   000 {             } 
   001 {    42 21 11 } f LBL A
   002 { 45 40 48  0 } RCL ✛ .0
   003 { 45 10 48  0 } RCL ÷ .0
   004 {          15 } 1/x
   005 {       43 24 } g COS-¹
   006 {          25 } TAN
   007 {       43 36 } g LSTx
   008 {          30 } −
   009 { 45 20 48  0 } RCL ✕ .0
   010 {    45 30  0 } RCL − 0
   011 {       43 32 } g RTN

RAD mode
Initial Guesses: 0, 1

Solve 'A' ->

Returns solver error unlike all HP calculators I have.
I realize this simulator uses different algorithm but never the less.
WP-34s also struggles at this.
I'm wondering if I'm doing something wrong.

I found interesting differences in answers given by: HP48G, HP35s and HP15c LE - real calculators and simulators: iPhone's 42S and Free15c. They all solve the case.

Regards,

[walter b]

(01-24-2014 05:35 AM)RMollov Wrote:  Solver problem.

EQN:

R(TAN (ACOS (R/(R+h)) - ACOS(R/(R+h))) - dL/2 = 0

R= 6,371,000m
dL=1m
h=?

...

RAD mode
Initial Guesses: 0, 1

Returns solver error unlike all HP calculators I have.
I realize this simulator uses different algorithm but never the less.
BTW WP-34s also struggles at this.
I'm wandering if I'm doing something wrong.

I'm wondering if your EQN is wrong. I can simplify it to R(0) - dL/2 = 0 which doesn't include h anymore.

[RMollov]

(01-24-2014 08:22 AM)walter b Wrote:  I'm wondering if your EQN is wrong. I can simplify it to R(0) - dL/2 = 0 which doesn't include h anymore.

You're always correct, thank you for correcting my typos.

It should be:

R*( TAN(ACOS (R/(R+h))) - ACOS(R/(R+h)) ) - dL/2 = 0

Did I count the brackets properly this time?
I'll correct my original post accordingly.

The HP15c listing above is right though.
Shall I post the WP34s version, which also doesn't seem to work?

Regards,

[walter b]

(01-24-2014 09:50 AM)RMollov Wrote:  Shall I post the WP34s version, which also doesn't seem to work?

Feel free to do so.

d:-)

[Dwight Sturrock]

Very nice. Runs over 50 times faster on my Linux based computer than the HP-15C LE.

[peacecalc]

Hello RMollov,

which kind of problem do you describe?

Quote:R*( TAN (ACOS (R/(R+h))) - ACOS(R/(R+h)) ) - dL/2 = 0

R= 6,371,000m (stored in register .0)
dL=1m (stored in register 0)

h=?

The value of R looks like the earth radius for me. And as a not very well scaled problem.

Greetings
peacecalc

[Thomas Klemm]

Quote:R*( TAN (ACOS (R/(R+h))) - ACOS(R/(R+h)) ) - dL/2 = 0

R= 6,371,000m (stored in register .0)
dL=1m (stored in register 0)

h=?

In register 0 the value dL/2 = 0.5 should be stored.

Quote:The value of R looks like the earth radius for me. And as a not very well scaled problem.

I guess it's the solution to a rope around the earth problem:
Take 3 of Sweet & Short Math Challenges #15: April 1st Spring Special

Cheers
Thomas