Report top-5000 primes here

[Siemelink]

Originally Posted by mdettweiler

Okay, glad to hear that it's getting fixed. And, don't feel bad about it--messing up in some way when reporting primes is a very, very common mistake and has probably happened to at least 75% of us here at some point.

Guilty!

[mdettweiler]

Originally Posted by MyDogBuster

I'd like to see you shell script if I could. I did use the -s option but haven't quite figured out how to que up work behind it yet.

Okay, essentially what I'm doing is queuing up stuff like this in a shell script:

#!/bin/sh
./phrot.p3-linux -b=3 -o -s totest.txt
cd ../../mprime
./mprime -d

What this does is run my Phrot executable on the file totest.txt, with the stop-on-PRP option enabled, so that it will exit on three conditions: 1) a Ctrl-C or other such shutdown signal; 2) it finds a PRP; or 3) it reaches the end of the totest.txt file.

After doing this, it executes the "cd ../../mprime" command to change to my mprime directory, and then run the mprime program. (mprime is the Linux version of Prime95, used by GIMPS, another prime search project; since their system is somewhat automated and has a large variety of work types to choose from, it makes a great choice for filler work like this if I don't have anything else hanging around to switch to ).

Note that the commands to start mprime in my example can be replaced with whatever else you'd like to start. Sometimes, if, say, I have an NPLB manual file sitting around, I'll have it switch to my LLR directory and start that instead. Or, sometimes I'll have it start an LLRnet client.

Another thing you can do is queue up multiple different files for Phrot to test, say, one k after another if you're doing multiple k's. This works best if your multiple files are in separate directories, i.e. with completely separated Phrot installations, and also if your computer is running 24/7. (It will still work OK if your computer has to be rebooted, though it can get a little more confusing. However, having the different files set up with different Phrot installations helps a bit with this.)

The shell script for that would work like this:
#!/bin/sh
./phrot.p3-linux -b=3 -o -s riesel45-24.txt
cd ../phrot2
./phrot.p3-linux -b=3 -o -s riesel45-372.txt
cd ../phrot3
./phrot.p3-linux -b=3 -o -s riesel45-1264.txt

...and et cetera, for all the k's you'd like to test on that computer.

Note that the above instructions are designed for Linux; on Windows the process is slightly different. Namely, you'd be using a batch file instead of a shell script, which means that, first of all, it has to have an extension of .bat or .cmd; you don't put the #!/bin/sh line at the top of the file; and you won't need to put "./" things in front of your Phrot commands.

Bonus tip for Linux only: how to string up commands without needing an actual shell script file. Bash, the Linux command line interpreter, has many handy-dandy shell script features--many of which work just as well on the command line as in a pre-written shell script. One of these is to put semicolons between multiple commands on the command line, and they will be run in sequence, just as if they were in a shell script. Like this:

max@max:~/prime/phrot$ ./go ; cd ../mprime ; ./mprime -d
(note: in this case, "go" is a small shell script that I use to start Phrot so that I don't have to keep re-typing all those command line flags )

You can even queue up multiple things on one line like this:
max@max:~/prime/phrot$ ./go ; cd ../llr ; ./llr -d ; cd ../llr2 ; ./llr -d ; cd ../mprime ; ./mprime -d
The above example would first run Phrot until it exits (which, if I've got the -s option set, would be either when it finds a prime, or finishes the file), then switch to one of my LLR directories, run a file there until it finishes, then switch to another LLR directory and do the same; and lastly, it will run mprime when all of those finish. I usually like to always put some sort of mprime command at the end of all of my chains of manual-reservation files; that way, if I'm not at the computer when the file finishes, it will immediately switch to mprime, which always has a continuous supply of work fed to it by PrimeNet (GIMPS's proprietary automated-testing system).

Hope this helps!

Max

[MyDogBuster]

Thanks Max. I think I can get something to work in Windows similiar to that. I keep forgetting that Windows command line will not execute another command until the prior one is complete. DUH And I've used the command-line for 15 years or so.

[kar_bon]

from Liskovets-Gallot-Conjecture: Riesel odd n:

147687*2^843689-1 is prime!

[kar_bon]

29847*6^141526-1 is prime!

[gd_barnes]

Originally Posted by kar_bon

29847*6^141526-1 is prime!

Karsten, congrats on our first top-5000 base 6 prime!

And once again: This knocks out a base 36 k:

29847*36^70763-1 is prime

[kar_bon]

perhaps another base-36-downer!?

and in such close series to the last one:

48950*6^143236-1 is prime!

PS: just looked at the base 36 page: another one down for this base too!

[gd_barnes]

Originally Posted by kar_bon

perhaps another base-36-downer!?

and in such close series to the last one:

48950*6^143236-1 is prime!

PS: just looked at the base 36 page: another one down for this base too!

Somebody bring out the fire extinguisher. Karsten is HOT! That's 2 n>800K primes base 2 and 2 n>140K primes base 6 for CRUS in the last month, all of course top-5000. Congrats!

That's amazing that you continue to find primes on k's that also make a prime on base 36. You are searching many more k's that would not be base 36 and further: if the prime were odd-n, the base 36 k would remain.

For the record on base 36:

48950*36^71618-1 is prime

There are now "only" 93 k's remaining on Riesel base 36 all at n=25K with the exception of the 3 k's that are being searched for base 6.

I like base 6. This will be a very fun team effort...top-5000 range, faster testing with Phrot, not a huge # of k's remaining on either side. I think we'll get a lot of "higher" top-5000 primes out of it before the testing gets so long that the effort is not as interesting anymore for people.

Karsten, could you provide your sieve file for n>150K (assuming you sieved further). We have a sieve file for n=150K-400K for Sierp base 6 sieved to P=6T. I'm in the preliminary stages of thinking how we're going to do a team effort on both bases. Since the files could be sieved together, I'm thinking we may need to sieve deeper for the higher n-range.

Since you started this and are kindly unreserving the effort at n=150K, you'll have first "dibs" on how the Riesel side is processed. Although we'd like to use a PPRnet server, if you'd like some large manual files instead (perhaps we'd split up the k's instead of n-ranges up to n<200K), that would be no problem.


Thanks,
Gary

[kar_bon]

no, i've not sieved a higher p-range than yours (but to n=1M).

perhaps there's no need for a Team Drive, if i find all remaining primes to n=150k

[gd_barnes]

Originally Posted by kar_bon

no, i've not sieved a higher p-range than yours (but to n=1M).

perhaps there's no need for a Team Drive, if i find all remaining primes to n=150k

Yeah, find them all! lol

What is your sieve depth? Since it's not sieved as deep, if you'll forward the file to me now, I'll coordinate a team effort to get it sieved deeper so that it can be combined with the much larger Riesel file for a very efficient combined sieve.


Gary

[mdettweiler]

Another one for Sierp. base 33:

1818*33^79815+1 is prime!

Only one k remaining for Sierp. base 33 (at n=100K and unreserved as previously reported in the bases > 32 status thread)--and this prime makes the top-5000 to boot!