Rolling along... :)

[Brucifer]

Okay... now I'm rolling , and have been reading up on the PSP threads. So as I understand it the low ranges we have reserved now are to increase the sieved scope to bring them up to the depth of the sob stuff to match the requirements for the combined stuff.

Are we as a team going to work on this area, as it needs to be done so that the PSP effort can move along jointly with sob? Myself, I think I'll grind away on the sieving stuff I'm on right now. It should go pretty fast anyway if we get a few people working on it, and maybe that would help Citrix out as it wouldn't disrupt the others that are already crunching away on other stuff????

I've still got some riesel stuff to finish yet too, but I can cut down on some of it to get a few systems devoted to psp project sieving.

Brucifer

[NeoGen]

They got a milestone set at 15000, that says "Merge with SOB when done till here". That we can reach quickly, as currently the unreserved ranges are already close to 14000.

[AMDave]

I agree.

There is a short term achievable goal here that would be a fantastic permanent result for the project.

I am at 52% done on my 500,000,000,000 range in Reisel with 12 days to go to finish it. (approx 7th Jan)
I'll cut the whole farm over, at that point, and assist to eliminate this extra DAT file permanently. (that is - if 72CJ5 doesn't finish it all by then :lol: )

[Brucifer]

heh............ :) I just reserved the 13850-14000 range. Left some for someone else. However if one of you wants to lead the way in the stuff above 15000 and figure it out for us, then go for it. I can reserve the 14000-15000 range and knowck that out in a few days. Then at least that section of the project will be completed for them. I've also got a riesel range to go yet, but for a short term to knock out this other I could handle that.

I did email the results for the 13850-1400 as the "how-to" here said to. On the PSP site, there was a comment about PM-ing them to ld or whatever the name was... slips my mind at the moment. I'm assuming though that all those guys talk with each other.

What I won't do for the PSP project though is join any Yahoo mail group. So if we can do the stuff above 15000 and just stay with the PSP forum, then I'm for that.

Bruce

[Anonymous]

no need to PM
just post in their forum that you finished the range
citrix will update the list on the forum.
ltd will update the db list & stats

[AMDave]

dang it.
that was me.
Forgot I flushed my cookies
Thats twice this month.
I hate anonymity :evil:

...I recon you will be into the 15,000s before I finish my RSeive range.

[NeoGen]

I can reserve the 14000-15000 range and knowck that out in a few days.

I say go for it! :)

I'm still at little over 50% of my combined range, so I'll probably only be done with it by the end of the year.

And yep, we can continue sieving beyond the 15000 milestone, or follow a couple of other directions, it's the users choice really.
Some time ago I asked about that on the PSP forums. Check out the thread about it:
http://www.mersenneforum.org/showthread.php?t=5172

[Brucifer]

Okay, I'll go ahead and reserve the stuff up to 15000, and get that knocked out. Then after that I've got to go back and finish a couple riesel ranges.

[Brucifer]

Okay, I reserved up to 15,000. Now for stuff above that, what .dat file do we use? And is that stuff single, combined, whatever???

[ltd]

Okay, I reserved up to 15,000. Now for stuff above that, what .dat file do we use? And is that stuff single, combined, whatever???
All ranges lower then 85000 use the same dat file as you use now. They are done with n=50Mil by SOB already so there is no need to use the combined file.

The ranges above 85000 marked with "NOT combined" also use this dat file. These ranges are gaps left over from the swichting to the 50Mil dat file. These ranges are my most wanted as it will make reservation clearer when there are only two starting points left.
One for second pass (<85000) and one for combined (>101300)

The ranges from 101300 to 125000 use the combined dat file.

Hope this makes it clearer,

Lars

[NeoGen]

It's pretty simple and straightforward. And no need to mess around with Yahoo groups or SOB. :)

One thing tough... on the range reservations thread you have

95000-98000 Available as firstpass test NOT combined
98000-98700 Available as secondpass test NOT combined
(EDITED... dumb question removed :P)
What's the difference between them? Have to sieve with different .dat files?
And can one request a small chunk out of it (ex:95000-95200), or do we have to request and sieve the whole 3000 at once?

And finally... have we been helping to eliminate many candidates on PRP? Each sieved range results in a bundle of factors, but I don't know how do those relate to the part of easing the work for the PRP'ers.

[ltd]

All unresered ranges lower then 101300 use the "normal" dat file as you use it now.

The difference between the firstpass and second pass ranges is the following.

For second pass test there has been already a sieving run with a dat file containing a maximum n=20Mil. So the chances of finding factors for n<20 Mil are very low. There are sometimes factors due to the fact of either software errors or human errors when reporting results back.
For the firstpass ranges you will find factors within the complete set more or less evenly distributed.
You can request ranges from these set in chuncks you like but i would request to reserve at least 50G if you are not only testing the client.

The best help for the PRP tests is when we do firstpass tests as the chances to find factors that are shortly before beeing PRP tested are higher. We had focused on the second pass sieving to reduce the dat file as much as possible to make combined sieving also intersting for the SOB contributors even for the price of not doing optimal for the PRP side.

Edit:
Some statirstical Info:
In the last 10 days we have found 2984 factors.
107 of the are lower then 20Mil
21 of them are lowerr then 2.2Mil.
Each factor removes at least one PRP test. In most of the cases even two tests as there is no need for a double check PRP test also)
And to be ahead of the next question.
With the second pass sieving in the last 10 days we found 15 factors not reported after the first pass. That are 14% of the factors found in the range lower then 20Mil!!!!

Lars

[ltd]

One more thing.
I hope it is ok for you that i jump in on all the PSP discussions not beeing a member of AMD users.

Lars

[Brucifer]

I'm glad you are jumping in Lars. It's less formal here without having to worry about a bunch of the hard core "math types" jumping on us for our obvious "to them" questions. :D Some good interesting data!!

Bruce

[NeoGen]

I second that :)

[vaughan]

Lars you are most welcome to post here.

Information from the project admins is very useful for several reasons.

1. It acknowledges the voluntary contributions made by our small but active team.

2. It shows us the project is alive and not static

3. It educates the users.

From my own perspective I have been interested in Maths for years but just couldn't get my head around anything more complex than second year university. I branched away from Linear Algebraic Equations, Analytic Geometry and Calculus, and ended up in Biological Sciences. Give me a complex Immunohaematology problem any day! However, there is still an underlying interest in Pure Maths that gets piqued when I read the Mersenneforums.

Feel free to post in our forums, join as a member - hey you don't even have to crunch for us :!: but it would be nice if you did ;) and remember there are never enough GHz for all the projects out there :)

[vaughan]

Hey NeoGen you got your XP2000+ overclocked or something? In the PSP Sieve stats update today you did 300 versus my XP2800+ that did 285.

Nice going team - we moved up a rank. :thumbup:

[NeoGen]

I did? In PSP sieve? :?:

But I haven't turned in my range yet. It's almost done, at 93%, but I haven't even turned in partial results or anything.

If anyone's turning in results on my behalf, then THANK YOU! :)

[ltd]

For PSP sieve you have registered the following results:

on the 13th 158 factors
on the 24th 142 factors

and vaughan has

on the 24th 139 factors
on the 27th 146 factors

Lars

[NeoGen]

And now a few more for me as I finally finished and turned in the combined range :)

[NeoGen]

Ltd, with new factors coming in every few days from sieving, and thus eliminating tests to PRP'ers, do you have to constantly update those txt files in the PRP reservation thread?

[ltd]

Ltd, with new factors coming in every few days from sieving, and thus eliminating tests to PRP'ers, do you have to constantly update those txt files in the PRP reservation thread?

The txt files are updated from time to time. You can find information about the last update time and the real size of the txt files here:
http://www.ldausch.de/candleft.htm

When the new server is set up the files will be updated on a daily base.

Lars