Thanks for your input. I posted a follow-up to my first message that
sets the background and quantifies the results.


If you would like to see details (or try it out for yourself) here is
the code in question...


In this first function, the call
GCD2(i, N)
is to my own function, pasted further down. Change this to
GCD(i, N)
to get ATP's GCD (assuming appropriate references set, of course).

Public Function Totient(ByVal N As Long) As Long
' Returns Euler's Totient (Phi) function of N such that
' Totient = the number of numbers less than N that are relatively prime
' to N.

Dim i As Long
Dim Result As Long

Result = 1
i = 2
Do While i < N
If GCD2(i, N) = 1 Then Result = Result + 1
i = i + 1
Loop
Totient = Result
End Function


This is my hand-rolled GCD function. It's nothing special, but it blows
away ATP's version of GCD by a factor of (almost) 10...

Public Function GCD2(ByVal a As Long, ByVal b As Long) As Long
' greatest common divisor
' as structured, it is advantageous to pass the lower of the two numbers
' as a
Dim i As Long
Dim Found As Boolean
i = a + 1
Do Until Found
i = i - 1
If a / i = a \ i And b / i = b \ i Then Found = True
Loop
GCD2 = i
End Function


To produce the test results I have a simple Sub that calls Totient(N)
for 1 <= N <= whatever, with timings captured.

Thanks again!



Peter T wrote:
Those debugs are normal though I'm sure a [sic] design flaw. They must
degrade performance but I wouldn't think to the extent of being much slower
than an equivalent VBA function. Why not post your VBA function, your code
that calls the ATP equivalent, and an example of the data.

Regards,
Peter T