Talk:Ulam numbers: Difference between revisions
Content deleted Content added
m added another talk section about Ulam timings (using REXX). |
m →timing for the 100,000th Ulam number: added a "ya can say that again" comment. |
||
Line 18:
:It certainly demonstrates the importance of using a decent algorithm.
::::: Yuppers, ya can say that again. -- [[User:Gerard Schildberger|Gerard Schildberger]] ([[User talk:Gerard Schildberger|talk]]) 21:39, 4 December 2020 (UTC)
:I was quite pleased with my first version but it's been trumped twice now by the Phix and XPL0 algorithms! --[[User:PureFox|PureFox]] ([[User talk:PureFox|talk]]) 21:36, 4 December 2020 (UTC)
|
Revision as of 21:40, 4 December 2020
timings for the Nth Ulam number
Using the (new and current) REXX program, the times
It (the REXX program) is an O(2) polynomial
0.0000005168509818 N^2 - 0.0004990440614098 N + 0.5707466809128310 Rsquared = 0.9999999315551220
(The above, as measured and timed on Paul Kislanko's 10-core PC. -- Gerard Schildberger (talk) 21:38, 4 December 2020 (UTC)
timing for the 100,000th Ulam number
The total time for the 100,000th Ulam number using the original REXX program would've taken a little over three years, and that is on Paul Kislanko's PC. On my old slow PC, it would've taken about a decade or thereabouts. -- Gerard Schildberger (talk) 21:22, 4 December 2020 (UTC)
- It certainly demonstrates the importance of using a decent algorithm.
- Yuppers, ya can say that again. -- Gerard Schildberger (talk) 21:39, 4 December 2020 (UTC)