![]() |
|
|
#12 | |
|
"Mike"
Aug 2002
822410 Posts |
Quote:
|
|
|
|
|
|
|
#13 |
|
Aug 2002
3×83 Posts |
What exactly would one graph? Everything but the last iterations is essentially just a sequence of bits.
I have suggested in the past a "towers of hanoi" type animation based on the number of iterations - this would allow it to appear to be working quickly even though we all know that LL work takes weeks to months. |
|
|
|
|
|
#14 | |
|
Oct 2002
7 Posts |
Quote:
One way to avoid the "script kiddie" problem would be for the download files to be signed with a PGP key and/or had an MD5 checksum available, much in the way that many of the Linux distributions handle their updates. Now that I think of it, it might be a good idea to do this even if you don't automate the process. |
|
|
|
|
|
|
#15 |
|
Oct 2002
Lost in the hills of Iowa
26·7 Posts |
Auto-updating the executable might be tolerable as a "I just finished all the work in my intermediate files, time to go look for a new upgrade" basis - but losing work due to an executable change in the middle of a LL would be WAY annoying.
|
|
|
|
|
|
#16 | |
|
Sep 2002
23×37 Posts |
Quote:
compare your computers output to thers stuff like that nothing fancy just line graphs or something to make it a little more interactive |
|
|
|
|
|
|
#17 |
|
Dec 2002
Frederick County, MD
2·5·37 Posts |
I like the idea of a "towers of hanoi" screen saver because the solution with the minimum numbers of moves takes 2^n-1 moves, a Mersenne number! n is the number of disks. Anyway, it could be aranged so that a disk is moved after a certian number of iterations, and the number of iterations would be chosen so a disk is moved every 2 to 4 seconds, and the number of disks would be chosen so that the puzzle would be solved at the end of the LL test.
Obviously, the number of iterations between disk moves would not be constant throughout the whole test so that the puzzle would be solved exactally at the same time as the test finishes. This can be applied to TFing also, and P-1 I suppose. Of course, you wouldn't have to use the screen saver if you want the LL test gets as many CPU cycles as posible, but the screen saver should be programed to use as little CPU as posible. |
|
|
|
|
|
#18 |
|
Sep 2002
23×37 Posts |
i dont think a animation just repeating its self would do much better in makeing people more intrested in praticipating.
i know it would be simple(ish) to do. i mean look at folding for example they have plugins that give you a welth of information about how your doing compared to other people i think its this comparison that drives them to be more compedative with eachother and drives them to do better and recruit more people |
|
|
|
|
|
#19 |
|
Dec 2002
Frederick County, MD
2·5·37 Posts |
It actually looks like there was once a GIMPS program with a screen saver. if you look at programs.txt, you'll see that there was a Version WQ3 and WQ4 that apparently had screen savers. Of course, this was way before my time in GIMPS, so who knows what that was all about? (read: I'm curious about this tidbit of history, does anybody know about it?)
|
|
|
|
|
|
#20 | |
|
"Sander"
Oct 2002
52.345322,5.52471
29·41 Posts |
Quote:
|
|
|
|
|
|
|
#21 |
|
Aug 2002
47 Posts |
I'd like to make some simple suggestions for the client.
1. Over a year ago George added a time stamp to some of the lines in the screen output - I'd like to suggest that he add that to *most* of the lines in the screen output (and also the results.txt output, if it isn't already done there). 2. There is an ongoing discussion in Lone Mersenne Hunters section on 'faster factoring' where it would be useful to add a time-stamped line at significant checkpoints in trial factoring and P-1, like when the bit depth changes (and also restart the iteration counter and timer at those points). 3. (Let's get greedy..) How about some trial factoring timings in the 'benchmark' part of the program - a small range of mersenne sizes and bit depths. Roy |
|
|
|
|
|
#22 |
|
Sep 2002
23×37 Posts |
i have a idea as well
when you run benchmark is there anway to have it run the bench then prompt you for some info chipset ram type timings ETC then store that in your results.txt and next time it comunicates with the server to send thouse results to a database or something like that ( they were working on a bench database in a diffretn thread) |
|
|
|
![]() |
Similar Threads
|
||||
| Thread | Thread Starter | Forum | Replies | Last Post |
| Enhancement request | chris2be8 | GMP-ECM | 0 | 2009-12-08 18:19 |
| Suggestion | henryzz | Marin's Mersenne-aries | 1 | 2009-04-14 10:33 |
| Bug? Feature? AND/OR Enhancement Request? | petrw1 | PrimeNet | 2 | 2007-10-16 19:13 |
| Formula entry enhancement? | Xyzzy | GMP-ECM | 5 | 2007-08-22 19:54 |
| A suggestion for the client/stats server/??? | QuintLeo | Software | 2 | 2003-01-28 09:08 |