RPL speed HP-48G vs HP-50g

[enkidu]

I've recently picked up an HP-50g and decided to run some of my simple RPL programs on it. From what I'd read it is claimed to be faster than my HP-48G, but when I run the identical code on it I find that it is soundly beaten by the 48.

Is this normal or is there something I've done wrong?

[Joe Horn]

(04-30-2024 01:16 AM)enkidu Wrote:  I've recently picked up an HP-50g and decided to run some of my simple RPL programs on it. From what I'd read it is claimed to be faster than my HP-48G, but when I run the identical code on it I find that it is soundly beaten by the 48.

Is this normal or is there something I've done wrong?

This can happen for several reasons, one of which is the use of "integer type objects" in the 50g program instead of the floating-point real numbers used in all 48G programs. For a fair comparison, make sure that all the literal numbers in the 50g program have a decimal point at the end, e.g. "153." instead of "153". The 48G doesn't display the trailing decimal points on real integers, but the 50g does.

If the program is longer than a few lines, just edit it in Approx Mode (MODE, CAS, Approx) to automatically convert all integers from integer type objects to floating-point reals.

[David Hayden]

(04-30-2024 12:28 PM)Joe Horn Wrote:  If the program is longer than a few lines, just edit it in Approx Mode (MODE, CAS, Approx) to automatically convert all integers from integer type objects to floating-point reals.

You can also toggle approximate node by holding down [Right Shift] and pressing [Enter].

[carey]

(04-30-2024 12:28 PM)Joe Horn Wrote:  

(04-30-2024 01:16 AM)enkidu Wrote:  I've recently picked up an HP-50g and decided to run some of my simple RPL programs on it. From what I'd read it is claimed to be faster than my HP-48G, but when I run the identical code on it I find that it is soundly beaten by the 48.

Is this normal or is there something I've done wrong?

This can happen for several reasons, one of which is the use of "integer type objects" in the 50g program instead of the floating-point real numbers used in all 48G programs. For a fair comparison, make sure that all the literal numbers in the 50g program have a decimal point at the end, e.g. "153." instead of "153". The 48G doesn't display the trailing decimal points on real integers, but the 50g does.

If the program is longer than a few lines, just edit it in Approx Mode (MODE, CAS, Approx) to automatically convert all integers from integer type objects to floating-point reals.

enkidu,

If you don’t mind, after applying Joe’s suggestion, please report your results for the benefit of readers curious about the speed difference after comparing apples to apples. Thanks!

[enkidu]

Thanks, this indeed made a significant difference. The 50g now runs the program somewhat faster than the 48G!