HP Forum Archive 16

[ Return to Index | Top of Index ]

New Bench ? :-) Message #1 Posted by gileno on 12 Feb 2007, 4:05 p.m. 5 RADIANS 10 A=1 20 FOR I=1 TO 2499 30 A=TAN(ATN(EXP(LOG(SQR(A*A)))))+1 40 NEXT I 50 PRINT A

Re: New Bench ? :-) Message #2 Posted by Xerxes on 12 Feb 2007, 5:34 p.m., in response to message #1 by gileno Hi gileno, have you missed this question ? ;-) A good idea. After having a benchmark for a non numerical problem, I also thought about a simple but convincing numerical benchmark. I think the basic problem is to find an usable program for testing. Probably a mix of simple float calculations and some mathematical functions are not bad. An example for a simple float test: 0->A LBL0 A*A/A+A-A+1->A GOTO0 Edited: 12 Feb 2007, 7:30 p.m.

Re: New Bench ? :-) Message #3 Posted by John Keith on 12 Feb 2007, 8:16 p.m., in response to message #1 by gileno This is a well-known benchmark usually known as the "Savage" benchmark. I have run it on several HP models in the past, and I will post results if I can find them in my old (deeply buried!) papers. I would definitely enjoy seeing results from other machines and look forward to many submissions! Thanks, John

Re: New Bench ? :-) Message #4 Posted by Gerson W. Barbosa on 12 Feb 2007, 9:22 p.m., in response to message #3 by John Keith Yes, this has been discussed here already: http://www.hpmuseum.org/cgi-sys/cgiwrap/hpmuseum/archv016.cgi?read=103356 The original benchmark program can be found here (Page 12, Column 1): http://www.amigau.com/68K/dg/dg25.htm Regards, Gerson.

Re: New Bench ? :-) Message #5 Posted by Tim Wessman on 12 Feb 2007, 9:58 p.m., in response to message #3 by John Keith Well, 50G w/HPGCC @12MHz takes .52s and returns 2500 Not sure exactly how many digits are being used internally. Anyone know? TW

Re: New Bench ? :-) Message #6 Posted by Xerxes on 13 Feb 2007, 11:01 a.m., in response to message #5 by Tim Wessman With the SDK compiler, the FX-9860GSD takes 0.42 seconds. Why 12 MHz on the HP-50G? Is there another reason except for less power consumption?

Re: New Bench ? :-) Message #7 Posted by Tim Wessman on 14 Feb 2007, 2:16 a.m., in response to message #6 by Xerxes Its just the default speed when compiling. You have to add in a sys_slowOff(); at the beginning, or modify a loader file to not mess with system speed. Keep in mind that there is always a .05-.06s penalty to START the program in saturn land. That is how long an empty program takes to run. TW

Re: New Bench ? :-) Message #8 Posted by John Keith on 14 Feb 2007, 11:05 p.m., in response to message #3 by John Keith A couple of timings: HP-71B: 344.82s HP-49g+: 63.48s (using BINTs for loop variables) Both return 2499.99948647 like all Saturn-based HP's. John Edited: 14 Feb 2007, 11:06 p.m.

[ Return to Index | Top of Index ]

Go back to the main exhibit hall