The Museum of HP Calculators

HP Forum Archive 21

[ Return to Index | Top of Index ]

HP-15C LE torture tests
Message #1 Posted by ckelley2 on 24 July 2012, 4:24 p.m.

Hey guys, I just bought a new limited edition 15c and was going through the definite integral torture tests as an exercise in learning to use my new calc. I noticed that the new LE edition is giving more correct answers on the torture test. Is there any interest in updating the torture results page with the new limited edition 15c?

Cheers

      
Re: HP-15C LE torture tests
Message #2 Posted by Mark Scheuern on 24 July 2012, 4:43 p.m.,
in response to message #1 by ckelley2

I thought the firmware was exactly the same as the 15C, with the emulation layer added. No?

      
HP-15C LE torture tests
Message #3 Posted by Paulo MO on 24 July 2012, 6:49 p.m.,
in response to message #1 by ckelley2

Hi

Please note that the display mode (as in FIX4, or FIX9) will affect the integracion accuracy/precision.

Best

Paulo

            
Re: HP-15C LE torture tests
Message #4 Posted by ckelley2 on 24 July 2012, 9:27 p.m.,
in response to message #3 by Paulo MO

Specifically Integration Test 3.3: integrate(x*exp(-x), 0, inf)

I made the program exactly as typed (in FIX 9) for the formula to integrate: CHS, exp, times, RTN.

I then entered 0 and 228 into the stack for the bounds and after about 10 seconds it gave 1.000000000 as the answer which is exact to 9 digits.

The chart says the original HP-15C gave the wrong answer of 0.0147 when 228 was used for infinity. This is fascinating to me because I thought the algorithms were identical with the limited edition as Mark said.

Edited: 24 July 2012, 9:28 p.m.

                  
Re: HP-15C LE torture tests
Message #5 Posted by ckelley2 on 24 July 2012, 9:58 p.m.,
in response to message #4 by ckelley2

Also integration test 3.6 now converges in about 8 seconds to 1.772453849 (good to 8 decimals) whereas before it was still running after 5 minutes.

                  
Re: HP-15C LE torture tests
Message #6 Posted by Gerson W. Barbosa on 24 July 2012, 10:04 p.m.,
in response to message #4 by ckelley2

Quote:
The chart says the original HP-15C gave the wrong answer of 0.0147 when 228 was used for infinity.

The displayed result 0.0147 implies FIX 4. That's what I get on the HP-15C after 12 or 13 seconds, which matches what is reported at http://www.voidware.com/calcs/torture_integration.htm. This is also what I get almost instantly on the HP-15C LE and on the newly arrived DM-15CC. 

01 f LBL A
02   CHS
03   e^x
04   *
05 g RTN  

f FIX 4 
0 ENTER 228 
f Integrate A    -->   0.0147

When the calculator is set to FIX 9 both the HP-15C LE and DM-15CC will give 1.000000000 (HP-15C LE: 10 seconds; DM-15CC: 50 seconds).

                        
Re: HP-15C LE torture tests
Message #7 Posted by ckelley2 on 24 July 2012, 10:10 p.m.,
in response to message #6 by Gerson W. Barbosa

Thanks for the reply Gerson! Do you know why they chose FIX4 then for the comparisons to other calculators? I guess it seems a little unfair, or too much HP fanboi over here, haha.

                              
Re: HP-15C LE torture tests
Message #8 Posted by Gerson W. Barbosa on 24 July 2012, 10:15 p.m.,
in response to message #7 by ckelley2

Perhaps this is a hint:

"since most machines have 10 digits, five correct digits is required to score."

FIX 4 on the HP-15C will generally grant 5 or 6 correct digits.

P.S.: This HP-15C was my first HP calculator, back in '83. I've just used it for this test :-)

Edited: 24 July 2012, 10:19 p.m.

                                    
Re: HP-15C LE torture tests
Message #9 Posted by ckelley2 on 24 July 2012, 10:20 p.m.,
in response to message #8 by Gerson W. Barbosa

Yeah, that's probably what they did. One last thing though! integration 3.6 in FIX4 really does give a solution in a few seconds on HP-15C LE, over 5 minutes on C?

'83, glad to hear its still running. oh and I was born in '85, haha.

Edited: 24 July 2012, 10:21 p.m.

                                          
Re: HP-15C LE torture tests
Message #10 Posted by Gerson W. Barbosa on 24 July 2012, 10:37 p.m.,
in response to message #9 by ckelley2

It would be in better shape if my younger brother (born in '67) hadn't used it all through his college years and for 15 more years at work (he's a civil engineer). Previously I had used it during at least one and a half year before replacing it with an HP-28S.

Yes, the new 15C LE is much faster. Namir Shammas has carried out extensive speed tests on it:

http://h20331.www2.hp.com/hpsub/downloads/HP_Calculator_eNL_09_September_2011.pdf

Regards,

Gerson.

                                                
Re: HP-15C LE torture tests
Message #11 Posted by Olivier De Smet on 25 July 2012, 1:25 a.m.,
in response to message #10 by Gerson W. Barbosa

go15c for android on a galaxy tab gives the perfect 1.000000000 result in about 5 secondes too :-) (this could be reduce to 3 secondes if needed)

And with a nice alpha programing mode.

Btw how can dm-15c be sold with the rom, did hp allowed this ?

If so, I could also put the rom included in the app (anyway, with google the rom is not too difficult to find...)

Olivier

Edited: 25 July 2012, 1:39 a.m.

                                                      
Re: HP-15C LE torture tests
Message #12 Posted by Gerson W. Barbosa on 25 July 2012, 9:36 a.m.,
in response to message #11 by Olivier De Smet

Quote:
Btw how can dm-15c be sold with the rom, did hp allowed this ?

At one time the Nonpareil Simulator included the classic voyagers except the HP-10C, but they have been removed due to licensing issues. Apparently there shouldn't be any:

http://nonpareil.brouhaha.com/microcode_copyright_status/

Not sure about that though.

Gerson.

[ Return to Index | Top of Index ]

Go back to the main exhibit hall