Loop of + benchmark now in Articles section

[Gene Wright]

New Benchmark

Enjoy!
Gene

[Glenn Shields]

It's interesting that the 35s is faster doing register arithmetic
(STO + Z) than simply +ing on the stack. Is that because dropping
the stack each time uses more steps? regards, Glenn

[Gene Wright]

Maybe it is due to the different types that might be on the stack. Is it a real? Is it a complex number? Each + might require that to be checked, whereas a STO+ operation might not. Guessing.

[Gerson W. Barbosa]

No 49G?

HP 49G

Count: 12,351

Code:  1. << DO 1. + UNTIL 0. END >>



HP 49G 

Count: 11,041

Code: << WHILE 1. REPEAT 1. + END >> 

Gerson.

[Tim Wessman]

I think everyone tries to forget that unit for a variety of reasons. . . :-D

TW

[Paul Dale]

What about in integers?

#1 << DO #1 + UNTIL 0. END >>

My 49 died a while back :-(
So did my 48sx :-(
And no batteries for my 28s and a broken battery door to boot :-(

- Pauli

[John Keith]

For my 50g, I get 19692 (average of 3 runs).

John

[Gerson W. Barbosa]

Quote:

---

What about in integers?

#1 << DO #1 + UNTIL 0. END >>

---

#7164d (WS = 16)

#7130d (WS = 64)

Gerson.

[Mark Storkamp]

16C using increment

HP-16C

Count: 340

With 0 in I

Code: Lbl A, Isz, Gto A

[Gjermund Skailand]

Hp 50G 12,000,000

User-rplfunction 'SUM(I=1, 12000000,1)' interpreted by custom library in HPGCC. The library evaluates functions, sums and numerical integrals from the Equation Writer in double precision for most real functions.

[Frido Bohn]

Hello Gene,



congrats, a real great compilation of that simple benchmark across several generations of calculators!



However, for the great HP41C/CV/CX one benchmark is missing: MCODE

In the book "HP41 MCODE for Beginners" of Ken Emery, published 1985, a 20-step machine code for this 1+ addition algorithm is presented (pages 50-52) which I wont list here.

To make things short:

Count: ~130 000

This is about 100-fold of FOCAL or Synthetic programs!



The 41 was, is and remains amazing