| Author |
Message |
|---|
 Sysadm@Nbg Volunteer moderator
Volunteer tester
Project scientist
Send message
Joined: 5 Feb 08
Posts: 1224
ID: 18646
Credit: 877,929,236
RAC: 321,810
                      
|
27_104T.zip is the current sieve file
your output shows you pointing to a file 27_110T
perhaps here is the failure ???
____________
Sysadm@Nbg
my current lucky number: 113856050^65536 + 1
PSA-PRPNet-Stats-URL: http://u-g-f.de/PRPNet/
 |
|
|
vkojlfVolunteer tester
Send message
Joined: 10 Oct 08
Posts: 337
ID: 30361
Credit: 41,827,004
RAC: 0
             
|
|
There is no 27_104T.zip in the sievefile directory, only 27_110T.zip. The 27_110T.txt is in the same directory as the sr2sieve executable. |
|
|
|
|
27_104T.zip is the current sieve file
your output shows you pointing to a file 27_110T
perhaps here is the failure ???
That is a typo :)
The filename is 27_110.zip
It was sieved to 104T when i started. I have sieved it up to 110T.
The file you shall use is 27_110T.zip
/Lennart |
|
|
vkojlfVolunteer tester
Send message
Joined: 10 Oct 08
Posts: 337
ID: 30361
Credit: 41,827,004
RAC: 0
|
|
Lennart,
Any suggestions to my problem? I have everything setup the same as I have in PPSE or Primorial sieves, but this one just doesn't like to cooperate.
If I leave out the -i option, sr2sieve errors out with
"ERROR: Failed to open input file `sr2data.txt'."
But with above command line it gives
"ERROR: Invalid or unrecognised file `27_110T.txt'."
so maybe there's something wrong with the sieve file? |
|
|
VatoVolunteer tester
Send message
Joined: 2 Feb 08
Posts: 851
ID: 18447
Credit: 713,903,832
RAC: 1,641,370
                         
|
|
I see the same as vkojlf.
Looks like a normal newpgen format file.
Shouldn't we be running this with sr1sieve at this stage?
____________
 |
|
|
|
|
|
make a file "sr1sieve-command-line.txt"
Download sr1sieve-1.4.1-bin.zip http://pgllr.mine.nu/software/sr1sieve/
sr1sieve -p110e12 -P111e12 -i27_110T.txt -ff27_110T-111T.txt -t4 -q
Use this on a quad Linux
-t4 is use all 4 core
sr1sieve -p110e12 -P111e12 -i27_110T.txt -ff27_110T-111T.txt -q
Use this on Windows
This example shows sieve 110T to 111T
!!!!! There is no checkpoint !!!!!!
If you stop it you have to check where you stoped it and change
-p110e12 to where you stoped it.
Example: if you stop at 110541000122345 Change -p to -p110541e9
/Lennart |
|
|
HonzaVolunteer moderator
Volunteer tester
Project scientist
Send message
Joined: 15 Aug 05
Posts: 1957
ID: 352
Credit: 6,139,676,503
RAC: 2,272,243
                                   
|
|
I'm using latest sr2sieve.exe to do 12121.
On E5430 (2,66GHz with 12MB cache) it is doing about 5,45Mp/sec each core, Win x64.
Needs different format for input sieve file but actually *does* checkpoint.
(one of my quads restarted unexpected but was able to continue after 35 hours of sitting idle from where it stopped).
____________
My stats |
|
|
warddrVolunteer tester
Send message
Joined: 7 Feb 08
Posts: 254
ID: 18735
Credit: 24,054,820
RAC: 0
      
|
|
As you said I can see a small speed increase with the new sievefile, my new reservation (just started) will be ready one hour earlyer
more information:
XPx64
Intel core2quad 2,5 Ghz (1 core)
ETA 2 days
____________
 |
|
|
VatoVolunteer tester
Send message
Joined: 2 Feb 08
Posts: 851
ID: 18447
Credit: 713,903,832
RAC: 1,641,370
|
As you said I can see a small speed increase with the new sievefile, my new reservation (just started) will be ready one hour earlyer
Bah! I started my reservation about 3 hours prior to this being posted!
I'll let it run to completion with 27_110T rather than fiddle about with it.
____________
|
|
|
warddrVolunteer tester
Send message
Joined: 7 Feb 08
Posts: 254
ID: 18735
Credit: 24,054,820
RAC: 0
|
|
You can stop the sieve, copy the first line from the old sieve file, delete the first line form the new sievefile and place the first line from the other sieve file there, then restart the sieve and everything will work well
____________
|
|
|
|
|
|
I am just starting to dabble in this. I have very carefully followed the instructions on how to work around the non-checkpoint issue. I read in another post that if you stop the run, you would need check the log and insert the last value in the header of the sieve. I set up a little test run, and when I stop the sr1 app, it seems that the header for the sieve file is automatically updated with where it left off. Isnt that accomplishing the same thing as a checkpoint? This is my command line:
sr1sieve-x86-windows.exe -P200e12 -i27_140Te0.txt -o27_140Te0.txt -ff27_199T-200T.txt -v
Like I said, I am fairly new at this, so if I am missing something obvious, let me know. If I can get this working like I want, I am anxious to reserve a range and get crunching.
____________
|
|
|
JohnHonorary cruncher
Send message
Joined: 21 Feb 06
Posts: 2875
ID: 2449
Credit: 2,681,934
RAC: 0
               
|
I am just starting to dabble in this. I have very carefully followed the instructions on how to work around the non-checkpoint issue. I read in another post that if you stop the run, you would need check the log and insert the last value in the header of the sieve. I set up a little test run, and when I stop the sr1 app, it seems that the header for the sieve file is automatically updated with where it left off. Isnt that accomplishing the same thing as a checkpoint? This is my command line:
sr1sieve-x86-windows.exe -P200e12 -i27_140Te0.txt -o27_140Te0.txt -ff27_199T-200T.txt -v
Like I said, I am fairly new at this, so if I am missing something obvious, let me know. If I can get this working like I want, I am anxious to reserve a range and get crunching.
You are correct, the work around accomplishes the same thing but with less work. :) Therefore, once you stop sr1sieve, you can start it back up without editing...no need to "check the log and insert the last value in the header". You're good till the end of that range.
However, when you start a new range again, you'll need to edit the sieve file to the start of the new range. And then you're good again till the end of the range.
____________
|
|
|
|
|
|
Not a very early moment to ask this ;)
Why 27 and 121? I had several "higher maths" classes at university, but for engineers, number theory and primes were not considered very important and basically left out. So I am wondering how 27 and 121 were chosen.
Regards, Peter
____________
There are only 10 kinds of people - those who understand binary and those who don't
 |
|
|
|
|
Not a very early moment to ask this ;)
Why 27 and 121? I had several "higher maths" classes at university, but for engineers, number theory and primes were not considered very important and basically left out. So I am wondering how 27 and 121 were chosen.
Regards, Peter
There wasn't much mathematical rationale behind the choice. www.15k.org was running a coordinated search for all K < 300 in an attempt to get them all searched up to n=1Million. I reserved 121 because it had high weight at the time, and was a palindromic perfect square (looked cool, did not mean too much in terms of primes) to coordinate a distributed effort to 1M.
When we got there, the group decided that they wanted to take it further, and has since found a number of top 50 primes (project: 12121search)
As 121 tests became large and more computationally intensive some of the members wanted faster tests, so we began looking for another K to test. We decided on 27 because it is a low K (low = low FFT = faster LLR tests for equivalent n) and because it is a perfect cube. Because it is a perfect cube, arithmetic factors can be removed (all n divisivble by 3). So it looked like a good place to continue.
As computation power increases, I surmise that all relatively low K ( < 300 and possibly higher) will be searched deeply, it just so happens that 1, 3, 27 and 121 are currently much deeper than most of the others, but you can expect other K to catch up quickly within the next decade. |
|
|
|
|
|
OK, thanks for the explanation.
As primes are notoriously unpredictable, I guess "looked cool" is just as good a reason for picking a number as anything else ;)
____________
There are only 10 kinds of people - those who understand binary and those who don't
|
|
|
|
|
|
Does anybody have experience with sr2sieve on i7 CPUs? Our latest toy is a dual-i7, but -t16 is not available. Does HT give a speedup on sr2sieve?
____________
There are only 10 kinds of people - those who understand binary and those who don't
|
|
|
