5th RPS Drive: 14 Ks < 300 divisible by 3

[amphoria]

688-689 complete, no primes.

Taking 691-692.

[arminius]

687to688 complete, no primes

[amphoria]

689-690 coplete, no primes.

Taking 692-693.

[Kosmaj]

Battlemaxx did 678-679, and I did 684-685, no primes.

I'm taking 693-694.

[Kosmaj]

B'maxx is now working on 561-562 and hopefully we'll close that gap shortly!

[gd_barnes]

Originally Posted by Kosmaj

B'maxx is now working on 561-562 and hopefully we'll close that gap shortly!

Sorry, I beat him to the punch. Fortunately in this case, there were no primes. Attached is the results file for all 14 k's for the n=561K to 562K range.

It's odd that I started on this 6 days ago and the gap has been outstanding for a very long time and you posted this just hours before I finished.

Sorry if I offended anyone here but this one was just too long and the search now appears up to nearly n=667K on all 14 k's for this drive.

Battlemax, I saved you a primeless 5-6 day CPU search. (3.2 Ghz processor)


Gary

[Kosmaj]

Gary

Stepping the toes of the others is not the way we handle affairs here.

Our first objective is search for primes but we are also trying to build a friendly atmosphere where we respect equally those who can complete a file in two days like those who need two months.

Please try to understand the reservation system first, and aftert that, if you want to take part in the drive, reserve your file in advance like everybody else.

Thanks.

[gd_barnes]

Originally Posted by Kosmaj

Gary

Stepping the toes of the others is not the way we handle affairs here.

Our first objective is search for primes but we are also trying to build a friendly atmosphere where we respect equally those who can complete a file in two days like those who need two months.

Please try to understand the reservation system first, and aftert that, if you want to take part in the drive, reserve your file in advance like everybody else.

Thanks.

Point taken.

[amphoria]

691-692 complete, no primes.

Taking 694-695.

[Kosmaj]

Originally Posted by gd_barnes

Point taken.

I'm glad to hear that

BTW, note that you processed 849 candidates while our test file sieved to 7.4T had 735.

I took into account your results and we are now done to 668k.

[gd_barnes]

Originally Posted by Kosmaj

I'm glad to hear that

BTW, note that you processed 849 candidates while our test file sieved to 7.4T had 735.

I took into account your results and we are now done to 668k.

OK, glad to hear that. Yeah, I knew I probably processed more candidates than anyone in the 5th drive would have but I wasn't going to sieve to 7.4T! I looked back at my sieve file and see that I sieved until 686G. I generally sieve until the rate being removed in Srsieve (or NewPGen for a single k) for just that one small range of 1000 n was about the same as the average time it took to process one candidate...in this case, about 10 mins. each. As a general rule, I read somewhere here that sieving should be about 5-10% of the total search time. That seems to hold true if you run a quick test before sieving at both the low end of your range and the top end to see how long each runs and then average the two when determining the rate to have NewPgen or Srsieve sieve to. I tend to like to sieve until a little more than the average, but technically doing so takes a little extra total time. I think the sieve ran about 10 hours or so, which is 5-10% of the 5-1/2 days it took for it to LLR process.


Gary