Profile: Gipsel
log in
Your personal background.
I like to squeeze an algorithm until it runs as fast as it possibly can. Ten years ago assembler was the way to go to have full control. But in the meantime I got older and now I'm restricting myself to C most of the time ;)

I got project developer here at Collatz because Slicker asked for some performance tricks for the ATI/CAL application. But now I'm also involved in improving the CPU app. I hope we beat the performance of the old 3x+1 code by almost two orders of magnitude even on CPUs!
Your opinions about Collatz Conjecture
Even as no mathematician was able to prove it yet, the Collatz conjecture is most likely true. That means I'm quite sure we won't find a number disproving it. But one never knows.

The sequence for all numbers tested so far ends in 4, 2, 1 (repeats 4, 2, 1 in a loop if not stopped). It has been already proven that no sequences with a loop consisting of fewer than 35,400 steps exists (besides the 4, 2, 1). With some time, we may be able to push our numbers into a region (a lot bigger than now) that such step counts become common (in the moment we are at ~2000 steps for the longest runs). That may increase the probability of finding anything. And with a faster code and GPUs helping we may achieve that a hundred times faster than the old 3x+1 project ;)
Your feedback on this profile
Recommend this profile for User of the Day:I like this profile
Alert administrators to an offensive profile:I do not like this profile
Account dataView
TeamPlanet 3DNow!
Message boards279 posts
Friends (8)

Profile
Crunch3r

Profile
Bayha Family Power

Profile
L@MiR

Profile
Kalessin

[P3D] Crashtest

Profile
GalaxyIce

verstapp*

Profile
MrBad



Main page · Your account · Message boards


Copyright © 2018 Jon Sonntag; All rights reserved.