Perfect numbers: Difference between revisions

m
→‎optimized using digital root: add/changed comments and whitespace.
m (→‎optimized using only even numbers: changed indentations for the ISPERFECT function.)
m (→‎optimized using digital root: add/changed comments and whitespace.)
Line 1,760:
===optimized using digital root===
This REXX version makes use of the fact that all ''known'' perfect numbers > 6 have a ''digital root'' of   '''1'''.
<lang rexx>/*REXX program tests if a number (or a range of numbers) is/are perfect. */
parse arg low high . /*obtain the specified number(s). */
if high=='' & low=='' then high=34000000 /*if no argsarguments, then use a range. */
if low=='' then low=1 /*if no LOW, then assume unity. */
if high=='' then high=low /*if no HIGH, then assume LOW. */
w=length(high) /*use W for formatting the output. */
numeric digits max(9,w+2) /*ensure enough digits to handle# number*/
 
do i=low to high /*process the single #number or a range. */
if isPerfect(i) then say right(i,w) 'is a perfect number.'
end /*i*/
exit /*stick a fork in it, we're all done. */
/*──────────────────────────────────────────────────────────────────────────────────────*/
/*──────────────────────────────────ISPERFECT subroutine────────────────*/
isPerfect: procedure; parse arg x 1 y /*getobtain the number to be tested. */
if x==6 then return 1 /*handle the special case of six. */
/*[↓] perfect #number's digitalRoot = 1.*/
do until y<10 /*find the digital root of Y. */
parse var y r 2; do k=2 for length(y)-1; r=r+substr(y,k,1); end end/*k*/
y=r /*find digital root of digthe digit root. */
end /*DO until*/ /*wash, rinse, repeat ··· */
 
if r\==1 then return 0 /*Digital root ¬ 1? Then ¬ perfect. */
s=1 /*the first factor of X. _ ___*/
do j=2 while j*j<=x /*starting at 2, find the factors ≤√X≤√ X */
if x//j\==0 then iterate /*J isn't a factor of X, so skip it. */
s = s + j + x%j /*··· add it and the other factor. */
if s>x then return 0 /*SumIs the sum too big? It ain't perfect.*/
end /*j*/ /*(above) is marginally faster. */
return s==x /*if the sum matches X, it's perfect! */</lang>
'''output''' &nbsp; is the same as the traditional version &nbsp; and is about '''5.3''' times faster &nbsp; (testing 34,000,000 numbers).