stephenbrooks.orgForumMuon1GeneralStats by magnav0x are now available!
Username: Password:
Search site:
Subscribe to thread via RSS
magnav0x
2003-11-02 20:08:25
Ok I put up the finishing touches on my Muon statistics.  The link is:

http://stats.dbestern.net/muon1

The stats run occurs daily at 2am GMT (10PM EST)

As of right now there is only about 48 hours of data in the database, so your weekly numbers will not represent your actual weekly output for another 4 or 5 days.

Let me know what you think!  And let me know if you see any errors!  Cool
Pollock[Romulus2]
2003-11-03 22:37:38
It is not hard to tell that the page is a rework of climateprediction with the 'model year' column headers.  It is a nice page though.
magnav0x
2003-11-03 22:51:45
Yes it is a rework of climatprediction.....seeing how i wrote the climateprediction stats at http://stats.dbestern.net/climateprediction and the chessbrain stats at http://mag.chessbrainstats.com

My engine....my layout Big Grin

No reason to redo the whole stats engine (which I just about had to do because of the way Stephan's rawstats dump is organized) and layout for each project.  But thanks for pointing out that I forgot to change a column name.
[OCAU] badger
2003-11-04 15:21:50
looks pretty good, "I like the laurels particularly"

I like the potential conquests, although a little more averaging might be helpful, I dumped a couple of weeks worth of points the other day.  The stats have just picked it up, and the potential conquests says I will be overtaking nearly everyone (ie including slipstream!) which seems unlikely (especially as one of my boxes died last night).

www.BadgerMotorsport.tk Proudly sponsored by GRX-Computers
magnav0x
2003-11-04 18:06:42
Potential conquests is based on weekly outputs, not daily, so it's pretty accurate.  I wouldn't dare base it on monthly rates.
[OCAU] badger
2003-11-04 22:04:42
woohoo, look out slipstream in 380 days...

seriously I am pretty sure I haven't dumped 714860.9 mpts in the last week.
my total score is only 3.68 million, and I have been doing this for nearly a year (started 18 nov 2002) although I have a bunch more power than when I started (with just a p2 - 400)

www.BadgerMotorsport.tk Proudly sponsored by GRX-Computers
magnav0x
2003-11-04 23:10:19
You are right, I need to change that, I was under the impression that Mpts was what they were called.  I assume now that Mpts stands for Million Particles?  lol
Diabolos.de
2003-11-05 02:23:24
First: Its look very good.
Second: You say its only about 48 hours.  Is that the problem why i cant find my name?  Today i have look at every Page, 600-700 Pages, but i cant find my name Diabolos.de.  Should i look later?


To Stephen: your Website stop below the histogram from muon1viewresults.exe

Signature: No Thx iam not ill.
magnav0x
2003-11-05 07:07:19
Diabolos, sorry you will be in the stats this evening when the stats run occurs.  The last stats update was 2 days ago, because the official stats were down yesterday.
magnav0x
2003-11-05 07:44:18
Well diablo, if you take a look at all user stats, you are ranked 126....for some reason it's just not there!  Eek In your team's stats you show up, but if you goto the drill down for you, your team rank and project rank don't show up Roll Eyes

I'll have to look at this closer this evenening and see if I can't find the problem.
magnav0x
2003-11-05 13:20:58
Ok, Diabolos.de, I found the issue that was causing you to not show in the stats and resolved it.  Happy stats tracking!  Wink
[OCAU] badger
2003-11-05 14:19:25
quote:
Originally posted by magnav0x:
You are right, I need to change that, I was under the impression that Mpts was what they were called.  I assume now that Mpts stands for Million Particles?  lol


mpts = mega particle time steps (ie one mpts = one million times that a single particle has been simulated through one timestep (= 0.01ns))
the mpts is the base unit of output measurement (ie what is recorded in the stats)

thus with my total score of 3.68 million mpts (note this is only since v4.3 started, I forgot that before this there was a different scoring system), I have processed 3.68e12 particle timesteps, ie equivalent to following 1 particle through 36.8 seconds of simulation (doesn't seem like much does it) or 3.68 billion particles through 10ns of simulation or 3.68 million particles through 10 micro seconds (thers never a mu when you need one).

one thing I have noticed now that the mpts are shown in the commandline, is that the output varies.  I have autosave on, so it should be processing the same number of mpts for each save interval.  I have the machine idle except for muon, and it does about 100mpts between each save, however it varies from 97 to 105 mpts.  I am not sure if this is dependant on the properties of the simulation (ie a high yield or low yield one) or just random.

edit: I just worked out that 714 860 mpts is my total output since I moved to "solenoids to 15cm", being 690 862 for v432x and the rest the one machine still on v431 (the k6-400).  been more than a week I think...

edit2: my actual output is about 25 000 mpts per day, or about 175 000 per week, if all machines are running, since one just died, it'll drop back to 16 000 or so.

edit 3: had a look at some other peoples stats, it seems that your "last week" column is actually 'since v432 started' as a bunch of the values in the "solenoids to 15cm" section of Stephen's Stats are the same as the last week column.


www.BadgerMotorsport.tk Proudly sponsored by GRX-Computers

[This message was edited by [OCAU] badger on 2003-Nov-05 at 23:02.]

[This message was edited by [OCAU] badger on 2003-Nov-06 at 0:22.]

[This message was edited by [OCAU] badger on 2003-Nov-06 at 0:41.]
Sir Avanor
2003-11-05 17:15:14
Nice site.  If I can make a suggestion, it can be a bit hard to read with light grey text on a white background.  If it were up to me, I'd have a black background and light colors for the text.

He who liketh long words.
magnav0x
2003-11-05 19:10:13
Ok hrm, Mpts sounds about what I thought it was right.  I don't know why your daily info would be incorrect.  I base all numbers on ($todays_total - $yesterdays_total).  Seems to all be ok according to my teams output.  Today was only a 12 hour stats run, due to the outage of stats the last two days.  BTW, the stats include 4.3 and 4.32 Mpts.  I assume the stats are indeed suppose to include solenoidsonly AND solenoids15cm.  Correct me if I'm wrong.  If I'm right then it's Stephan's stats that are completely fubar.  Here is what I noticed.  I'll use a random example that shows exactly what the problem is:

User 'testboy' started crunching Muon at version 4.3 and has...let's say 719,000 Mpts at 4.3. User 'testboy' takes a break and runs a different project for a bit.  When he returns he notices he's now 2nd place in his team and decides to start crunching muon again.  *GASP* what's this?  A new version?  He downloads 4.32 He starts crunching and finishes his first set (100 results) and they upload.  Next stats run, he looks at his team's stats and he only has 3,200 Mpts.

This is the case with all three members on my team that have done solenoidsonly AND solenoidsto15cm results.  The mpts in the section are recorded for 'testboy' when stephan does the stats run.  His stats right now are 719,000 in the database....the stats gen keeps going and gets to the section of the stats.  Oh looky here, 'testboy' is in the results and it doesn't check to see if he had previous results in the rawstats.txt it just plops 3,200 Mpts into the database and his previous 719,000 completely disappear.

Please tell me I'm wrong and isn't suppose to be included in the 4.3 statistics, because if it is.....the official stats are FUBAR.
magnav0x
2003-11-05 19:29:42
After looking at it some more, I'm very confident it is completely fubar.  Though team totals aren't effected by the problem, just the users.

quote:

13. Free-DC.org 0 17338 797070.1 (0.171%) 13.784161
13-1. IronBits[Free-DC.org] 0 4657 186371.8 (23.4%) 2.446771 4
13-2. magnav0x[Free-DC.org] 0 2083 85503.5 (10.7%) 3.696214 11
13-3. Beyond[Free-DC.org] 0 2009 79318.9 (10.0%) 1.905858 0
13-4. excaliber[Free-DC.org] 0 555 58823.8 (7.4%) 9.119736 689
13-5. Supp[Free-DC.org] 0 1304 57650.4 (7.2%) 13.784161 3314
13-6. the-mk[Free-dc.org] 0 707 27985.7 (3.5%) 1.357845 0
13-7. wirthi[Free-dc.org] 0 96 2287.1 (0.3%) 0.241724 45
13-8. wirthi [Free-DC.org] 0 45 1456.8 (0.2%) 0.28099 16



Some how the numbers just don't add up Roll Eyes

Beyond[Free-DC.org] (182967 in rawstats.txt) (79318.9 official stats)
wirthi [Free-DC.org] (363482 in rawstats.txt) (1456.8 official stats)
Diabolos.de
2003-11-06 02:06:38
THX

Signature: No Thx iam not ill.
Herb[Romulus2]
2003-11-06 03:23:05
Database Last Updated At: 11/05/2003 26:00 GMT LOL

-------------------------------
I'd say more, but I can't reach the keyboard from the floor.
magnav0x
2003-11-06 08:46:54
Yeah, don't you like the extra two hours of crunch time?  Big Grin
[OCAU] badger
2003-11-06 14:56:44
1. I am pretty sure you are correct when it comes to Stephen's team stats (they are doing 4.3 unless you have 4.32 results), however I don't really look at them.  I am pretty sure you wouldn't need them, as you can do it with the individual user stats.
2. Your daily scores seem to be ok, it is the weekly ones which are cactus.  I am not sure if this is because you don't have a database yet (only been running a few days?).
3. I don't have any problems reading it, I like the colour coding, although it might be better to do weekly output than daily, as daily will fluctuate a lot...
4. The stats should include both 4.3 and 4.32, although if you just want to do only 4.32 I wouldn't mind as I am doing much better in 4.32(#7 instead of #30).
5. perhaps the best transfer should be the 4.32 best, as the 4.3 transfer is less meaningful...

www.BadgerMotorsport.tk Proudly sponsored by GRX-Computers
magnav0x
2003-11-06 16:11:07
Right now weekly info will take a few more days probably to fill out.  Currently my stats engine parses straight from the rawstats.txt, and includes both optimizations
best muon is determined by the users best muon whether it be SolenoidsOnly or SolenoidsTo15cm.

As far as I know...I'll keep color coding based on daily output, as that's how I have always done it.  When I have more time (probably a week or two from now) I will add fields for each user showing their total results, etc for both 4.3 and 4.32.
[TA] D347h
2003-12-20 01:59:14
Nice stats, magnav0x !!

Join [TA] !

TA Forums | TA WebPage
: contact : - - -
E-mail: sbstrudel characterstephenbrooks.orgTwitter: stephenjbrooksMastodon: strudel charactersjbstrudel charactermstdn.io RSS feed

Site has had 25162920 accesses.