Sieving for CRUS

[lalera]

hi,
here is a sieve-file for R102
n= 100k to 200k
sieved to 7T

[lalera]

hi,
the following range is done
S106 n= 25k to 100k
sieved to 7T
results emailed

[lalera]

hi,
did you look at the conjecture of R15 ?
http://www.noprimeleftbehind.net/crus/tab/CRUS_tab.htm
... around 36T k´s ...
maybe i should do a range from k=10M to ?M, n=to 25k with the script
and openpfgw first

[rebirther]

Originally Posted by lalera

hi,
did you look at the conjecture of R15 ?
http://www.noprimeleftbehind.net/crus/tab/CRUS_tab.htm
... around 36T k´s ...
maybe i should do a range from k=10M to ?M, n=to 25k with the script
and openpfgw first

Hmm, I missed that part, there are a lot left, can you do the current range for the last 14k's at k<10M?

[lalera]

Originally Posted by rebirther

Hmm, I missed that part, there are a lot left, can you do the current range for the last 14k's at k<10M?

hi,
yes i will do

[lalera]

hi,
reserving R15 n=200k-400k to sieve for BOINC

[gd_barnes]

I would suggest against testing k>10M on R15 at this point. We are unlikely to get all k tested on it in the foreseeable future. Testing k<10M to n=400K is fine. Some bases are just not worth attempting to test all k's.

[lalera]

Originally Posted by gd_barnes

I would suggest against testing k>10M on R15 at this point. We are unlikely to get all k tested on it in the foreseeable future. Testing k<10M to n=400K is fine. Some bases are just not worth attempting to test all k's.

hi,
ok, thank you

[VBCurtis]

Originally Posted by gd_barnes

I would suggest against testing k>10M on R15 at this point. We are unlikely to get all k tested on it in the foreseeable future. Testing k<10M to n=400K is fine. Some bases are just not worth attempting to test all k's.

k = 2^31 is a natural bound for these monster bases, as that's the limit for using sr2sieve. There's a reasonable argument for running the script on R7 and R15 to 2^31 (or just 2e9 for bookkeeping cleanliness), to maximize the use of sr2sieve; such an effort could reach top-5000 level somewhat efficiently.

I agree with Gary that it's ridiculous to contemplate testing all k's for R7 or R15, but I'm interested in testing to 2^31- I'll help with sieving if Lalera et al wish to proceed.

[lalera]

Originally Posted by VBCurtis

k = 2^31 is a natural bound for these monster bases, as that's the limit for using sr2sieve. There's a reasonable argument for running the script on R7 and R15 to 2^31 (or just 2e9 for bookkeeping cleanliness), to maximize the use of sr2sieve; such an effort could reach top-5000 level somewhat efficiently.

I agree with Gary that it's ridiculous to contemplate testing all k's for R7 or R15, but I'm interested in testing to 2^31- I'll help with sieving if Lalera et al wish to proceed.

i will proceed later with R7 and R15
at this time i have no free resources
feel free to sieve R7 and R15

for sr2sieve
For sequences k*b^n+/-1 and prime factor p, the following limits apply:

1 < k < 2^32.
1 < b < 2^32.
core(k)*core(b) < 2^31. (core(k) is the squarefree part of k).
0 < n < 2^32.
k,b < p < 2^52. (or 2^62 for x86/x86-64 builds, or 2^63 for ppc64).
(taken from the readme)
k=up to 4e9

[lalera]

hi,
the following range is done
R145
n= 25k to 100k
sieved to 7T
results emailed