Palindromic gapful numbers: Difference between revisions
m
→Ludicrously fast to 1,000,000,000,000,000th: *10,000 on 64 bit
SqrtNegInf (talk | contribs) (Added Perl) |
m (→Ludicrously fast to 1,000,000,000,000,000th: *10,000 on 64 bit) |
||
Line 2,738:
</pre>
=== Ludicrously fast to
Astonishingly this is all done with standard precision numbers, < 2<sup><small>53</small></sup>. You realise this is like ten thousand times the ''square'' of the previous limits, and still far faster.<br>
I will credit [[Self_numbers#AppleScript]] and the comment by Nigel Galloway on the talk page for ideas that inspired me.
<lang Phix>-- demo/rosetta/Palindromic_gapful_numbers.exw
Line 2,918:
{100_000_000_000,1,9},{1000_000_000_000,1,9},
{10_000_000_000_000,1,9},{100_000_000_000_000,1,9},
{1000_000_000_000_000,1,9}
{10_000_000_000_000_000_000,1,9}} -- 64 bit only
-- (any further and you'd need mpfr just to hold counts)
atom t0 = time(), count, keep, start
for i=1 to length(tests)-(machine_bits()!=64) do
{count, keep, start} = tests[i]
atom from = count-keep+1
Line 3,043 ⟶ 3,044:
Completed in 0.7s
</pre>
On 64bit you'll also get
<pre>
10,000,000,000,000,000,000th palindromic gapful number ending with:
9: 968787878787878787639936787878787878787869
</pre>
I would agree that the last entry does not feel very convincing. Depending on how much
| |||