[QUOTE=paleseptember;408139]Curses and phooey. Yes, S576. Thankfully it was churning away for <3 hours. I can't edit my original post, so let's say that I'm taking S576 from new with the NewBases script. AFAICS it's the lowest CK unstarted.[/QUOTE]

There are 4 bases with lower CK unstarted but none under base 800. They are: S808, S918, S843, and S927. See [URL="http://www.mersenneforum.org/showpost.php?p=201754&postcount=1"]here[/URL].

S737 and S803
 

Reserving S737 and S803 up to 400K

[QUOTE=gd_barnes;408143]... See [URL="http://www.mersenneforum.org/showpost.php?p=201754&postcount=1"]here[/URL].[/QUOTE]

Reserving R675 from 2.5K to 25K using file from linked post above.

I have run srsieve for R675 and S576 to 1e9, and am now sieving using sr2sieve to extend that further. Factors at being removed at >1/sec for both bases, so it's got a little way to go.

What is the perceived wisdom for swapping from sieving to primality testing? When the time to find a factor is half the length of a primality test? 20%? 10%? (Edit: I've done some searching on the forum, and have seen some varying suggestions. 70% comes up a few times.)

[QUOTE=paleseptember;408266]
What is the perceived wisdom for swapping from sieving to primality testing? When the time to find a factor is half the length of a primality test? 20%? 10%? (Edit: I've done some searching on the forum, and have seen some varying suggestions. 70% comes up a few times.)[/QUOTE]

I came up with the 70% number (actually 1/sqrt2). That is my estimate for how to find the sample candidate from the sieve to time a primality test for, and then sieve until testing that sample candidate is faster than the sieve removal rate. Ex: If my sieve file goes from 10k to 20k, I test a candidate with exponent 17k (70% of the way from 10k to 20k), and then sieve until my factor rate is equal to the time for that test.

However, that number assumes you're going to test every candidate in the file for primality, which is not the case with CRUS. So, our "average" test is a ways lower in the file than 70%. Halfway is a reasonable estimate, lower if your base produces many primes or your file covers a wide range (2.5k to 25k is pretty wide, a factor of 10).

I would sieve until factor rate is equal to a primality test at about 8k, then test the file from 2.5k to 10k, remove primed k's and exponents below 10k, and then continue to sieve the rest until factor rate is equal to a primality test at around 18k.
Hope this helps!

S808
 

Reserving S808 to n=25K

EDIT: I am testing this using the posted file to n=2500.

S997
 

Reserving S997 to n=25K

S592
 

S592 tested n=2500-25K

400 primes found - 370 remain

Results emailed - Base released

Reserving R792 to n=100k (50-100k) for BOINC

Reserving S533 to n=200k (100-200k) for BOINC

Reserving S1018 as new base with srbsieve, KEP will give me the srbsieve.ini with the right phases. Running up to n=10k, I will create a sievefile for n=10-25k after the run for the future.

S533 tested n=200k (100-200k)

nothing found

Results emailed - Base released