Accuracy and the power function

[Dieter]

(02-16-2014 03:11 PM)Joe Horn Wrote:  The middle example contains an extraneous digit near the end.

Right. Has been corrected now.

Quote:On the HP 50g, the errors on these calculations are 1 ULP, -1 ULP, and -4 ULP, respectively.

Same here on a 35s and on the 39G emulator. Hm, 4 ULP in a simple standard function. I admit I would not have expected an error that large. Not in an HP.

Dieter