stephenbrooks.orgForumMuon1GeneralTime tag in the sample files
Username: Password:
Search site:
Subscribe to thread via RSS
Herb[Romulus2]
2003-12-13 05:17:59
Apart from ignoring all these results in Michals MuonManipulation tool, can you shed some light into their appearance?
Uhh, the additional semicolon in line2 killed my exell-import functions, but that is peanuts Wink

Thanks

-------------------------------
I'd say more, but I can't reach the keyboard from the floor.
Stephen Brooks
2003-12-13 13:49:41
The time tag is something that (as of last week) my results-receiver program adds itself.  Everything after the four types of brackets on the 2nd line is a 'comment' so really anything could be written there.  The units are 10^-4 years, so the beginning of Jan 1st 2004 is 40000 (i.e. formula (year-2000)*10000).  I added these so that I could see in what order various new things emerged and maybe do other analysis later.

HB Pencils, also sold as "Moron's Choice" Graphite Cigars.
Michal Hajicek
2003-12-14 03:41:48
I'm still using v4.32d, so there was no problem at my side.  I will try to fix this ASAP.
Michal
[AMD Users] Michal Hajicek
2003-12-14 10:22:36
Tool updated, everything should work.
Herb[Romulus2]
2003-12-15 05:34:20
Thanks you both Smile

Another question for the new gen=x parameter in v4.33, all but gen=0 are at the end of the parameter set, the latter starts the string.  That is very uncomfortable in data handling, any chance to equalize this?

While we are at it, any brief method description for the different gen values available?
So far for me gen=0 made only crap, while gen=1 and 3 produced high results, 2 below average.

-------------------------------
I'd say more, but I can't reach the keyboard from the floor.
[AMD Users] Michal Hajicek
2003-12-15 13:23:56
This is question for Stephen, but I think #gen is method used for genome creation (Random,Interpolate,Mutate,Crossover)-short client observing should show you the order.  And imho param order is not important for stephen's reading procedure - also sfinalf is sometimes at the beginning,... (this and #gen - problem for me too (but solvable - additional code)) - so you have my vote in your question Wink.
Stephen Brooks
2003-12-16 03:35:27
Yes, the #gen field is for logging what sort of optimisation step was used to generate the result.  #gen=0 is a random design, and the reason why it appears at the _beginning_ of the genome is that before instantiated by a lattice file, a "random" design is actually empty - the parser automatically fills in missing parameters with random values.  Anyway, that was probably more detail than you wanted to know Big Grin. More usefully:
#gen=1 is a mutation
#gen=2 is a linear interpolation
#gen=3 is a genetic crossover.
My results handler treats the parameters as an unordered set (that's why I pair the parameter name with the value), so assuming stuff about the order is dangerous.  You could think of a script to sort the fields in alphabetical order if you wanted - actually maybe I'll write a utility for that.

HB Pencils, also sold as "Moron's Choice" Graphite Cigars.
Herb[Romulus2]
2003-12-16 04:08:55
Thanks for the explanation Smile

The sortorder was ok so far except for the gen=0 now.  As for gen=0 can generate very good results by accident only, it's no real problem as in most cases it will be filtered away anyway.
The sfinalf problem was only related to very low values time ago, which I never tracked in a spreadsheet anyway.

My script knowledge is limited to ancient DOS-commands, a little tool would be really helpfull.
If my memory serves right, them Rechenkraft guys had a tool in v4.2 which extraxted directly to csv or excel.

-------------------------------
I'd say more, but I can't reach the keyboard from the floor.
Stephen Brooks
2003-12-16 07:16:29
OK, there is a utility here and up on the main page.  It sorts the genome into a particular order and outputs a CSV file from an input file.  Just works non-interactively at the commandline for now.

HB Pencils, also sold as "Moron's Choice" Graphite Cigars.
Herb[Romulus2]
2003-12-16 07:59:15
Simply COOL Cool

-------------------------------
I'd say more, but I can't reach the keyboard from the floor.
: contact : - - -
E-mail: sbstrudel characterstephenbrooks.orgTwitter: stephenjbrooksMastodon: strudel charactersjbstrudel charactermstdn.io RSS feed

Site has had 25163796 accesses.