IronBits - Riesel Sieve

[em99010pepe]

Just simple the best. (http://stats.free-dc.org/new/teamstats.php?proj=rsieve&team=Free-DC)

:allhail:

Keep going...

Carlos

[IronBits]

:blush: :umm:
:bigtrain: I think I can, I think I can, I think I can... :|party|:

[LAURENU2]

:blush: :umm:
:bigtrain: I think I can, I think I can, I think I can... :|party|:
Hay thats my line:swear:

[Bok]

28M in 7 days!!! :banana::banana::banana:

Bok

[em99010pepe]

Ironbits,

It took me 2 years to make 60M and you in 8 days.... :lmao: :smoking:

Carlos

[IronBits]

Now you know how I felt when the BIG IRON came after my #1 spot in DPAD a few months ago... :rotfl:

:beep::cheers:

[em99010pepe]

Congratulations on 2nd place within Free-DC team.

[tnerual]

he will not stay a long time at this place :

Chuck 1 88,703,701 9,345,579 0.5 day


congrats for 1st place !:cheers:

[IronBits]

Well my Winblows box HDDs took a dump, thanks to Winblows Vista Beta 2 :(
So I've restarted over on a range that was almost completed.
Good news is I'm running SUSE 10.1 with Wine and getting the same out put, so I'm a happy camper :D
I had to move resielator over to another Windblows box so it found 2800 new Factors that were re-submitted, so I hope the Server can work it out properly.

[tnerual]

never trust windows ...

and never use the last version.

actually, i use win2000 pro and it's very stable and less unsecure than win xp-2003-vista.

the only problem is at the install you have to download all the patches and services pack ...

my problem with linux is compatibility ...
i have a hp server (pentium pro 200 :thumbs: ) and when i run the suse install, i have all the install in graphical mode, and when the computer restart after finishing the install, i have a xserver problem: unable to start graphical mode ...
and it's still not very "install friendly" (yes yes yes next next yes i agree yes yes next)


will stay on windows 2000 for a long time

laurent

[em99010pepe]

I had to move resielator over to another Windblows box so it found 2800 new Factors that were re-submitted, so I hope the Server can work it out properly.

No problem about that. The server will take off the old factors.
Next time before re-running the range ask Bryan about the last submitted factor because you can check where you stopped sieving.

Ex:

Right now I am sieving the 167845-168095 range, last factor found was 167975738897591 | 31859*2^17796108-1, so 167975738897591 is the last position I should restart the range if something happens.

Carlos

[IronBits]

The ranges in question are these
181125-181325 proc1
181325-181525 proc2

[em99010pepe]

The ranges in question are these
181125-181325 proc1
181325-181525 proc2

Ok. I will ask Lee to check the last factors submitted by you.

Carlos

[IronBits]

Any news yet?

[em99010pepe]

Sorry IronBits. Didn't see this message.

Nothing yet. I will re-send the email.
Forget that range, sieve others.

And congratulations for the 140M points. Good work.

Carlos

[IronBits]

Good idea, range abandoned.
It's a partial set from a larger range I reserved.
I got thru about 1/2 of it before the computer crashed.

range abandoned
pmax=180725000000000
pmin=180325000000000

[em99010pepe]

Lee Replied. Only Bryan can help us out and he is returning today.

[IronBits]

I sent you an email to forward off to the keepers of the Sieve ;)

[bryan[RS]]

Jeff / IronBits -

Thanks for being patient, as I was out in the middle of the Caribbean and far away from all things RS.

Last factor I show in that range: 180720934209961 (so 180720).

However, there's a break from 180418266543377 to 180446331483739, but beyond that the factors seem in order (so maybe you need to run 180418-180446? Is that the part that failed?

180325-180825 is showing as abandoned. Should I split up the done pieces and re-issue as an abandoned range (180720-180825), or would you like to finish? I can easily reassign.

From 181125-181325: 181211847677381 is the highest factor.

From 181325-181525: 181412903327233 is the highest factor.

Bryan
Stats Administrator
Riesel Sieve Project

[IronBits]

Yes, please split up the done pieces and re-issue as an abandoned range.
I've got all the computers on other ranges lasting upwards of 16 days now ;)
Might be completed faster if it was put back in the queue. :)

Hope you had a great time! :drool: :D

[TeeJay]

Bryan, if we use RieselSieve to submitt our factors along the way,
Do we need to re-submitt all of them again when done ?

Also, when reserving ranges, do they overlap?
Meaning reserve x-y, then y-z, does y get done twice?

Thanks,
>>TeeJay


']Jeff / IronBits -

Thanks for being patient, as I was out in the middle of the Caribbean and far away from all things RS.

Last factor I show in that range: 180720934209961 (so 180720).

However, there's a break from 180418266543377 to 180446331483739, but beyond that the factors seem in order (so maybe you need to run 180418-180446? Is that the part that failed?

180325-180825 is showing as abandoned. Should I split up the done pieces and re-issue as an abandoned range (180720-180825), or would you like to finish? I can easily reassign.

From 181125-181325: 181211847677381 is the highest factor.

From 181325-181525: 181412903327233 is the highest factor.

Bryan
Stats Administrator
Riesel Sieve Project

[bryan[RS]]

TeeJay:

1) No, you only have to submit them once. The page automatically calculates the # of factors in a range only when the user selects "finished" for the status. You can submit all the factors at once or one by one - the script doesn't care.

2) No, y does not get done twice. How the sieve program runs, it's really x to (y-1) and then y to (z-1). The user inputs x-y and y-z, however.

Bryan

[TeeJay]

Ahhh... thanks for the input.
It was a pain to round up all them pesky fact.txt files and copy/paste...
I will just use RieselSieve to upload the factors, and when done mark
the range as complete on the reservation page.

Thanks for the help Bryan ! :banana:
>>TeeJay


']TeeJay:

1) No, you only have to submit them once. The page automatically calculates the # of factors in a range only when the user selects "finished" for the status. You can submit all the factors at once or one by one - the script doesn't care.

2) No, y does not get done twice. How the sieve program runs, it's really x to (y-1) and then y to (z-1). The user inputs x-y and y-z, however.

Bryan

[TeeJay]

Bryan, any idea why such a disparity between BOK's stats and your stats?

I see BOK has some higher and some lower than yours...

I ASSume yours only counts the credits after final submission...

Thanks,
>>TeeJay

[bryan[RS]]

Hi TeeJay,

The main reason for the (fairly slight) discrepancies is that Bok's tats are being pulled from our old stats system (I know, how do the old stats still get updated and used.....), while the pages at stats.rieselsieve.com use the new system. The main reason that I haven't switched over is there is some discrepancy in the scores, especially for sieveing, when multiple updates in a day are done. The LLR data is pretty good - the new pages are the more correct stats for those, as some duplicate scores are in the old DB that shouldn't count.

Also, team changes are handled differently in the old/new stats. This also leads to some changes in scores.

I know there's a number of reasons, but I really am working to get them all corrected. :)

Bryan

[TeeJay]

Bryan, what brought this to my attention is the fairly slight discrepency in
IronBits' scores... (btw, I crunch for my brother,IB) :thumbs:
It is now over 100m difference...

Thanks,
>>TeeJay


']Hi TeeJay,

The main reason for the (fairly slight) discrepancies is that Bok's tats are being pulled from our old stats system (I know, how do the old stats still get updated and used.....), while the pages at stats.rieselsieve.com use the new system. The main reason that I haven't switched over is there is some discrepancy in the scores, especially for sieveing, when multiple updates in a day are done. The LLR data is pretty good - the new pages are the more correct stats for those, as some duplicate scores are in the old DB that shouldn't count.

Also, team changes are handled differently in the old/new stats. This also leads to some changes in scores.

I know there's a number of reasons, but I really am working to get them all corrected. :)

Bryan

[tnerual]

the difference is here:

bryan stats show only the first dump of the day ...

bok stats take all dump of the day ...

look at the free-dc mercenaries account: 1 or 2 factor a day ... 20000 points on bryan stats, but 400000-500000 points on bok stats.

correcting bryan stats should be very easy once he point out where is the small code line to correct ... and finding it is very diffficult

Laurent

[TeeJay]

Thanks for the Heads Up... 'preciate it !

>>TeeJay


the difference is here:

bryan stats show only the first dump of the day ...

bok stats take all dump of the day ...

look at the free-dc mercenaries account: 1 or 2 factor a day ... 20000 points on bryan stats, but 400000-500000 points on bok stats.

correcting bryan stats should be very easy once he point out where is the small code line to correct ... and finding it is very diffficult

Laurent

[em99010pepe]

I am uploading my factors under Free-DC Mercenaries account and I can garantee that I upload more than two factors a day.
For the last two days I had problems with my work machine that's why my low output. I think everything is back to normal.

Carlos

[bryan[RS]]

Good news everyone - I had several consecutive hours I could devote to this problem, and think I finally have it solved. I'm crossing my fingers for the 4:15 (EDT) update from Bok's stats, but I think we should be very close.

What it broke down to was a MySQL error, which I knew. What I didn't know was where the problem was. It wasn't in the GROUP BY statements, like I though, it was actually in the beginning, where I delete the old data. For the sieve stats, the user table wasn't being cleared for today's date - meaning that only the first stats run was being counted. It was too obvious of a problem for me to catch. What pointed it out was when I decided to run the stats as 4 parts, rather than 1 (Misc, LLR, Sieve, P-1). By breaking up the statements to make the new procedures, I found the problems.

I've re-run all data from Jan 1, 2006 until now. There are some users who actually were too high (Footmaster especially, I'm still not sure what that problem was), and they will see a reduction, while most users will see a bump up.

Once I'm sure this is working right, I will re run the last two months of 2005, as it seems that's when the problem started appearing. This will only affect 17 users, and not large swings, but there will still be some change. This means nearly all users should see a match between what was reported in Bok's stats (and the old RS site stats) and the current RS site stats.

Thanks for all your patience - since the stats are basically a one-man operation, things take a bit of time to correct & sort out.

Bryan

[IronBits]

:clap: Good work! :cool: