Testing new Ranges for Sierpinski/Riesel

[pepi37]

I found bug in srbsieve.
On first post you say

You can specify as many of these as you want, although there is a practical limit due to newpgnew limits with sieving

I made 60, and srbsieve just exited.. So I reduce number until 50 ( and with 50) srbsieve works OK. But if it is 51.th in srbsieve.ini: it exits and do no job.

[rogue]

Originally Posted by pepi37

I made 60, and srbsieve just exited.. So I reduce number until 50 ( and with 50) srbsieve works OK. But if it is 51.th in srbsieve.ini: it exits and do no job.

There is a limit of 50 in the code. I could change the program to exit nicely, but I see little reason that anyone would sieve to such a high n with newpgen when using this program.

[pepi37]

Ok , I go off the scale :)

Please can you clarify this

phase= is used to sieve in steps. The first value is the upper n for the phase followed by the number of concurrent k to sieve and finally by the maxp of the sieving.

What doest it mean, and how to calculate that number to be optimal in all phase stages?

[wombatman]

It will sieve a certain # of k at a time. So if you have, say, 10,000 k to go through, you can choose to do 1000 at a time for a given phase. Maybe it's a little faster? I haven't played with it much.

[pepi37]

Originally Posted by wombatman

It will sieve a certain # of k at a time. So if you have, say, 10,000 k to go through, you can choose to do 1000 at a time for a given phase. Maybe it's a little faster? I haven't played with it much.

Great, since there is only 90 left( in my case), and I set to 1000 it is then OK
Thanks for explanation!

[unconnected]

uint32_t for p_max in source code seems not enough, should be uint64_t.

...
Status (02:50:50): Completed phase 4: 28 remaining
ERROR: 64 bit sieve is limited to primes below 2^51.

srbsieve.ini:

...
phase=120,50000,500000
phase=330,30000,3000000
phase=1200,20000,30000000
phase=4400,10000,300000000
phase=10000,5000,3000000000

[unconnected]

#ifdef WIN32
   sprintf(buffer, "del %s sr_b.pfgw pfgw.ini", inFileName);
   system("pfgw64 -Cquiet -f0 sr_b.pfgw > nul");
   system(buffer);
#else
   sprintf(buffer, "rm %s sr_b.pfgw pfgw.ini", inFileName);
   system("./pfgw64 -Cquiet -f0 sr_b.pfgw > nul");
   system(buffer);
#endif

Another small addition - nul should be /dev/null in second case.

[pepi37]

Yes, got same error four times ( on four cores)

Status (09:25:33): Completed phase 5: 13 remaining
ERROR: 64 bit sieve is limited to primes below 2^51.

my ini file

phase=1000,200,10000000
phase=2000,100,30000000
phase=4000,100,100000000
phase=8000,100,300000000
phase=12000,100,1000000000
phase=15000,100,3000000000
phase=20000,100,4000000000
phase=25000,100,5000000000

[rogue]

Okay. I think I've got everyone's suggestions implemented.

[rogue]

One mistake, I have /dev/nul instead of /dev/null. I'll update if another bug is found.

[wombatman]

My house had a short power outage. Is there a good way to try and pick back up with srbsieve, at least in terms of having it remove all the k's that have a prime found, or, even better, near the last n it was searching?