Progress update
 

K4 S593 at 700K

R711 tested to n=300k (100-300k)

1 prime found, 15 remain

Results emailed - Base released

[QUOTE=rebirther;618186]R711 tested to n=300k (100-300k)

1 prime found, 15 remain

Results emailed - Base released[/QUOTE]

1 out of 16 for a trebling of n-range seems notably low. Low enough to make sure there wasn't a mistake somewhere?

[QUOTE=VBCurtis;618212]1 out of 16 for a trebling of n-range seems notably low. Low enough to make sure there wasn't a mistake somewhere?[/QUOTE]


It is as it is, bad luck, I have only one base left with the old llr 3.8.21, all other are running now with llr2 and a Gerbicz check which detects bad hardware and rerunning from the last check point. We have already planned some double checks, one is running at the moment.

Stand-alone, these last two bases for n=100k-300k are not extremely unusual. Taken as a whole for the year of 2022, what has happened is extraordinary. We had ~20-25 bases with 2 k's remaining, all of which had no primes for either n=300k-500k or 400k-500k followed up by R88 with 0 out of 19 for n=250k-500k.

After the above, there were some bases with a normal # of primes, and then we had these two for n=100k-300k:

R711:
Sieve depth: 1e15
# of tests: 71198
expected # of primes: 3.7
actual # of primes: 1

S954:
Sieve depth: 1e15
# of tests: 60181
expected # of primes: 3.0
actual # of primes: 2 (k=119 had both primes)

By themselves, nothing to be alarmed about. It's just become a recurring theme this year. What sticks out at me the most about R711 is that there were 8 primes for n=50k-100k but only 1 prime for n=100k-300k. This happens often on the project due to sheer randomness but it's just been strangely frequent recently.

S964 tested to n=300k (100-300k)

2 primes found, 14 remain

Results emailed - Base released

Progress update
 

K4 S593 at 714K extending to 800K

R612 Completed to 25k


488 Primes found, 969 k remain


Primes attached

R639 tested to n=300k (100-300k)

4 primes found, 13 remain

Results emailed - Base released

S985 tested to n=500k (400-500k)

nothing found, 3 remain

Results emailed - Base released

Reserving S536 to n=500k

Using LLR2 with Gerbicz error checking