I've personally tried to sieve and send-to-BOINC a variety of exponent ranges, such that whatever one's interest for prime size there would be R3 work available.
Folks can individually test n=50-100k, or ~250k in 6-10G range (by asking for a chunk from me or later via BOINC), or ~1M range in k<2G; finally, one could sieve for oneself or for BOINC 500k-? in k=2G-6G.
I like offering variety; Gary likes a single n-value for the entire base. I try not to clutter up his pages with too much variety!
That reminds me, a status update: k=60-61G is at 72k, while k=61-63G is at 56k. Each is testing via LLR on a single core; when 60-61G reaches 100k that core will help with 61-63G, if I don't split the job sooner.
|