shouldn't we have one now
Printable View
shouldn't we have one now
Glad that this comes up.
We have one. It's a bit hidden, though.
Short story: the new 7k-dat is in the Yahoogroup SierpinskiSieve (For all those who have access) and the new 17k-combined-with-PSP-dat is available here.
Long story: In the yahoogroup, there has been a poll, whether SoB should continue sieving alone or switch to the combined .dat. There was overwhelming evidence (7 out of 7 votes) for switching. Here in the forum, the response was similar, nobody against switching and some favourable. So, the sieving effort is in transition right now. As you point out correctly, the websites need to be adapted, as well the sieving forum as Matt's site, and everybody needs to be informed about it. I don't have the rights to do so, and the ones who do seem to be busy; but anyhow, this is the future, I think.
If you want to do efficient work, it is recommended to sieve the next range you reserve with the combined .dat, as you get more than twice as much factors for 30% more work. (To swich in the middle of a range will perhaps result in problems for later gapchecking).
Hope this made sense, Yours H.
I'm in the middle of exams in uni at the moment, but once that's over I can change the sieving site if required. Since becoming a bit more involved in sieving I'm inclined to think a forum thread may be all that is required - but if people want to have automated sieving thing I'll be happy to adapt what I've done into a new system for combined sieving.
If I remember correctly, your system has already comments implemented, that should be enough. Only thing to change would be some information and links on the first page.
If we are doing a range with the combined dat file, do we need to reserve the range anywhere else, or just on the automatic system (http://mattford.org.uk/sob/)?
It depends on the range. If you got it through Matt's system, a comment "combined dat" is enough.
But you must submit the factors at two different places:
1)At SB's submit page
2)At PSP's submit page
Yours H.
NOTE:
If you only submit once at the PSP site, your factors are not lost, but you won't get credit at the SB site.
If you only submit once at the SB site, and do not send your fact.txt file to factrange, your PSP factors will be lost.
This post has been edited for one minor error and for emphasis.
HHH,
I could have just added my post, but I was afraid that people would read only your post and not go on to the next post. I hope that you do not mind.
Joe
Basically the difference between the two reservation systems is simply the range at this point.
At SoB we didn't really start using the 991-50M dat project wide until about 800T, currently we are around 1250T.
PSP basically started again from scratch with the 991-50M dat and fully joined in the secondpass sieve effort with SoB around 100T (little less but not much).
Quite a few of the heavy hitters started sieving with PSP back around that point and with some pretty great efforts we now have that level around 600T.
The question will remain what is the dirrection?
Currently the best bet is to stick with the combined dat since most of the ranges >1200T will probably be resieved again if they are done with the SoB only dat.
If your more of an SoB person reserve with Matt's site as per usual except use the combined SoB-PSP dat. However please indicate if you have used the combined dat or not. If you do not indicate Joe and I have to guess which could lead to potentiall missed factors for SoB. If your more of a PSP person or want to find the most factors reserve using the mersenneforum for now.
Eventually we will probably switch to a combined reservation system and personally I really like Matt's application and think he is doing a great job.
The only question remains is what to do with the range from 800T to 1200T??? The worst case senario is using seperate dat files for this range. It's not much of a benifit to PSP to use a PSP only dat and SoB could potentially miss factors between 20M<n<50M.
I personally suggest PSP skips the 800T to 1200T range and we fully combine the efforts once PSP reaches 800T. Several reason primally based upon the factor density and combination of the projects.
Comments? Questions?
Second if SoB finds another heavy weighted prime soon such as 55459, the sieve effort will be less effective and our growth in n will increase at a substantial rate.
Thanks for the help. I'll be using Matt's reservation system, as I'm familiar with how it works. I'm sieving with the sob only dat at the moment, but when this range is finished I'll do the combined one.
I've had a look at the PSP site, but there's 2 options: do I want
SOB factors reported by PSP or
SOB factors reported by user?
Also, I've not been sieving for a while. Should I still send the factrange.txt file to factrange@yahoo.com?
Silverfish,
Since those are pretty high ranges now, I would suggest that you reserve at least 500G at a time. I know it will take a while to finish but it makes things easier for us.
After the range is finished please send factors to factrange@yahoo.com as well think of it as a back-up. Generally if Joe or I have any questions those questions are quickly answered when a user has submitted all their fact, factrange, factexcl, files.
Also for the PSP submitting thing.
Personally I submit all of my factors first through the SoB site, then send them to PSP with the factors submitted by user option, followed by e-mailing to factrange@yahoo.com. That order is probably the importance to everyone.
I assume the option "SOB factors reported by user" means that you (the user) will have already submitted your factors to SOB yourself and so this does not need to be done by the PSP admins (which happens with the other option). The order suggested by VJS seems entirely appropriate to me.
That's right.Quote:
Originally Posted by Matt
Thanks for the help. I'll be sieving with the combined dat when I've finished my current range, which is just a short one to get me back in the swing of things. I'll probably do a 500G or a 1000G range next.
so it should sound like SOB factors already reported by user?
That's the jist I got
4. Download the DAT file to the sieve folder
http://www.teamprimerib.com/vjs/Newe...k_20051019.zip
is it newest?
The current combined dat file is at http://www.psp-project.de/test/sievecomb.zip
Thanks Sheep,
Sheeps suggestion and link are correct.
The dat he has linked is the dat we should all be using at the present time.
The DAT I've linked to on the co-ordination page is the 8k SOB-only DAT file. I think there's still ongoing discussion about what we're doing in terms of combined or otherwise isn't there? Or should I change it to combined DAT and explain about the dual factor submission process?
Please change it to the combined DAT and explain about the dual factor submission process.Quote:
Originally Posted by Matt
and i think there should be 7-k dat file )))))
Death,
We will certainly release a 7k dat later but probably only for P-1 purposes. Currently the best application of sieve is using the combined dat.
SOB siever who start with a combined range and have no userid at PSP so far please send me a private forum message and tell which username you want to use. I will create an userentry for you.
To see if your username is already entered go to
http://www.psp-project.de/sieveimport.php
and look in the username box if you find yourself.
Cheers,
Lars
I mean combined with 7k ))Quote:
Originally Posted by vjs
Or in combined sieve it should contain 8k from SoB?
ps /me goes to combined sieve again (100G), bye WCG )))
Death the SoB only dat is the one that contains only 7k.
The combined dat is the dat with both the sob k's ( 7 k's ) plus those of PSP.
The combine dat is the one we should all be using at this point.
but right now afaik combined dat contain 8k from SOB not 7k
Only if you are using an old one.Quote:
Originally Posted by Death
The current one is:
18K = 7K from SOB + 11K from PSP
13221KB May 6, 2007
Uploaded to the SierpinskiSieve Yahoo Group May 8, 2007.
Also obtainable here.
Pointed to from the Sieve coordination (991<n<50M) thread,
and the PSP Sieve: Pick a Range thread,
and the second and sixteenth posts of this very thread that we are in now.
uhm..
Can we have a repacked .dat now? With all dublicates removed.
Thats a strange reason maybe. But my top speed a home is 299k. If I only got +1% it goes over 300k )))
please....
Death try overclocking your PC by 1mhz on the FSB this will probably give you 300 kps and would certainly be alot easier.
how hard is it to overclock the CPU? How much can it be done without any risk of damage on a athalon 64 3400+?
Here is a nice overclocking guideQuote:
Originally Posted by Keroberts1
http://www.xtremesystems.org/forums/...nitor+software
The suggesting was for only a very minor overclock of 1 Mhz on the FSB. As such does not require the work needed for a larger overclock. There is lots of info on the web on overclocking. I just picked one that I had booked marked. This is a new dat thread so maybe this can get moved at some point.
oh man. my bios only allows me to set freq by 33/40/50 etc, not by +1 increase...
Might want to try 40 it's a considerable increase but... might work.
To come back to the original topic.
A new dat file is available for combined sieving.
http://www.psp-project.de/test/sievecomb.zip
can I just replace old one in a middle of sieving?
PRS52 v0.102b for processor type SSE2
Found SoBStatus.dat with unfinished work.
Continuing from last save point.
Setting priority to Process 0 Thread 3
Running in Sierpinski mode
Using CMOV optimizations
Using SSE optimizations
Using SSE2 optimizations
Expected 17 k values.
Found 17 k values.
k 10223: mod 96: 5 29 41 77 89 :
k 21181: mod 264: 20 68 92 116 188 212 236 260:
k 22699: mod 360: 118 190 262 334 :
k 24737: mod 120: 7 31 103 :
somewthing is missed
k 55459: mod 180: 46 58 94 118 130 154 166 :
k 67607: mod 360: 27 131 171 251 :
k 79309: mod 360: 38 182 254 326 :
k 79817: mod 120: 23 39 63 71 111 119 :
k 90527: mod 24: 23 :
k 152267: mod 360: 3 123 147 219 267 339 :
k 156511: mod 360: 48 120 264 336 :
k 168451: mod 180: 0 36 108 :
k 222113: mod 120: 21 53 77 93 101 :
k 225931: mod 360: 32 56 128 176 248 272 :
k 237019: mod 180: 70 106 142 178 :
k 258317: mod 120: 15 39 71 87 95 111 119 :
k 265711: mod 120: 0 24 48 :
kpowerlcm: 72
kmodlcm: 15840
grange: 15840
Setup took 6.672096 seconds.
Yes.Quote:
Originally Posted by Death
I'm not sure if you are trying to be funny, or just don't realize that a prime has been found for 33661.Quote:
Originally Posted by Death