mersenneforum.org (https://www.mersenneforum.org/index.php)
-   Data (https://www.mersenneforum.org/forumdisplay.php?f=21)
-   -   Interim iterations in status.txt are not random (https://www.mersenneforum.org/showthread.php?t=1334)

 GP2 2003-10-28 21:56

Interim iterations in status.txt are not random

The status.txt file shows how many iterations have been done for each exponent that is being LL-tested (or blank if LL-testing hasn't started yet).

The number of iterations is shown as a decimal number. But if converted to hexadecimal, an interesting pattern emerges: many of the iteration count end in 3F, 7F, BF, FF or 40, 80, C0, 00 or 01, 81.

For instance:
[font=courier new][size=1]
10000363,D ,65,5181823,66.6,6.4,60.4,28-Oct-03 02:57,23-Aug-03 01:05,Team_Prime_Rib,Reboot_It201
[/size][/font]

Here 5181823 = 4F117F (hex)

This only applies to LL-testing iteration counts. For factoring, the "iteration count" is an integer from 1 to 16 (or blank).

Here are the frequency counts for the top 16 values for the last two hexadecimal digits of the LL-testing iteration count in status.txt (Oct 28 2003 16:00 UTC):

[code]
1822 00
1704 ff
1107 40
1074 c0
1058 80
1050 7f
1037 3f
1022 bf
615 01
386 81
33 36
32 7d
29 f4
29 a6
28 95
28 8e
28 70
28 54
28 22
28 16
28 13
[/code]

 GP2 2003-10-28 21:59

If examining the final four hexadecimal digits, there is a preference for 0000, ffff, and to a lesser extent 0001:

[code]
721 ffff
712 0000
232 0001
21 0040
20 003f
14 0c3f
13 e580
13 0081
[/code]

 garo 2003-10-28 22:51

That is primarily because Prime95 performs a number of checks - including whether to communicate with the server - every 65536 iterations.

 All times are UTC. The time now is 08:34.