Curiouser and curiouser, said Alice...

Check it out: N 1 to 1000 -

HP-75 exponentiating (N^2) - 1m2s. N*N - 13.65s

HP-71b exponentiating (N^2) - 1m2s. N*N - 26.36s

TI-74 exponentiating (N^2) - 15.8s(!) N*N - 14.45s(!!)

I'm guessing the differences may be down to processor architectures, speeds and chosen algorithms. It's interesting to me that the 71 and the 75 are so close exponentiating, but a major difference in multiplying by itself.

The 74 is just pretty quick either way! Which seems really weird, given it's abysmal performance in exponentiating to a fractional power.

That's all relative. Virtual II running at the //e's "actual" speed (~1MHz) came in at 50.4s for exponentiation and 4.8s for N*N.

The handhelds all run on AAA batteries. I think the 75 runs at 700kHz, but don't quote me on that.

Still, it's interesting that in Applesoft (Microsoft's 8-bit BASIC with Apple modifications), exponentiation is 10x slower than just multiplying N*N; a much larger difference than the ~2 to 4x difference on the HPs, and a 10% difference on the TI.

Nobody's mining any crypto on these bad boys!

Originally posted at Nice Marmot 16:06 Saturday, 2 September 2023