stephenbrooks.orgForumMuon1Generalcomputed vs baseline
Username: Password:
Search site:
Subscribe to thread via RSS
Maniacken [US-Distributed]
2004-05-31 12:01:17
The baseline design is 74.3x better than the computed one that was for phaserotation

any way to get one for chicanelinac

and how was the computed one calculated.

kenny
Stephen Brooks
2004-05-31 14:43:02
It should have been 74.3% not 74.3x (now fixed).  The 'baseline' was the original human-designed setup, which our 'computed' designs are trying to beat.  So we haven't yet.  Though at the beginning of this weekend it was 6.59x better than what we'd done.

There'll be one for ChicaneLinacB once I get it simulated at work.
Maniacken [US-Distributed]
2004-05-31 15:13:08
thanks for the quick reply.

how is the human design calculated?
[TA]JonB
2004-05-31 18:26:41
I calculate the PhaserotB baseline value to beat at something like 1.5155.

We have crossed the 1.0... threshold and still climbing, so the curve suggests we could beat the baseline in 3 to 4 days.
hades
2004-05-31 19:22:48
I was the first one to post 1 1.0xxxx result YAY for me Big Grin
Maniacken [US-Distributed]
2004-05-31 19:48:51
nice job hades, keep it going.

wonder how long till we get past 1.0 for chicanelinac.
Stephen Brooks
2004-06-01 04:37:35
quote:
Originally posted by Maniacken [US-Distributed]:
how is the human design calculated? 
So far, I have been given a baseline design calculated by someone else to optimise around.  It is usually found by working with some formulae and a bit of trial and error.
Stephen Brooks
2004-06-01 08:32:24
I've now simulated both baseline designs with v4.41b, so both comparative measures are shown.  SolenoidsTo15cm didn't really have a 'hand-made' design to compare with.

Note that our optimisation is done independently of the baseline designs - i.e. it doesn't "know" about them beforehand, and this is why we're starting off below and will hopefully end up above them.  If for some reason we don't manage to get above the baseline transmission, I could start another optimisation with that design deliberately added as a seed.
[TA]z
2004-06-01 08:53:39
mmm... more tasty stats Big Grin
Maniacken [US-Distributed]
2004-06-01 10:11:16
sweet.

now we got something to aim for.

wonder who will get to 1.0 first.  a test result or a computer trial and error result.
Stephen Brooks
2004-06-01 11:51:09
Hmm well they would be good stats if the stats were working

In the upload that last went, I saw [SG]Sobi uploading about 4 6MB files, most of which appeared to be repeats (i.e. they could have been 1 real upload and 3 mistaken ones of the same thing).  I wonder if that's done in the stats because those files took quite a while to come across.

Otherwise, I did actually compile a trivial modification into the stats just before leaving, so I wonder if I also added an equally trivial bug?  Tune in tomorrow for next week's exciting episode...
Stephen Brooks
2004-06-02 01:22:06
It appears to have stopped due to a silly bug where it decided to prompt me for something and I wasn't there.  If that's all that was causing it, they should be back up soon.
Herb[Romulus2]
2004-06-02 09:59:52
"So far, I have been given a baseline design calculated by someone else to optimise around.  It is usually found by working with some formulae and a bit of trial and error.  "

You, there's some smartass around, ditching our hundred thousands of precious CPU-hours with just some wild guesses Eek Big Grin

Put that design on the table, it's faster just to come down to the current crap values Wink

Just kidding Smile

More tweaking to come up Smile
Pollock[Romulus2]
2004-06-03 08:05:50
quote:
You, there's some smartass around, ditching our hundred thousands of precious CPU-hours with just some wild guesses


ROFLM*AO!
That's a classic, Herb.  Big Grin
[TA]z
2004-06-03 09:58:39
Heavenly Hams are the best hams.
Maniacken [US-Distributed]
2004-07-17 13:19:02
just was wondering if we could get a baseline vs computed for the solenoid to 15 cm to see how it compares to the other two simulations.
K`Tetch
2004-07-26 09:19:24
sorry its late, but did you not read what was posted jun1st - namely "SolenoidsTo15cm didn't really have a 'hand-made' design to compare with."
Maniacken [US-Distributed]
2004-07-26 21:58:44
Sorry i did not see that.  Ill read closer next time.
: contact : - - -
E-mail: sbstrudel characterstephenbrooks.orgTwitter: stephenjbrooksMastodon: strudel charactersjbstrudel charactermstdn.io

Site has had 17087212 accesses.