Sieve file boundaries demystified

[mdettweiler]

As the NPLB/PrimeGrid sieving drive advances, we are moving our team drives to the better sieved files now available. This has created some confusion as to which primes should be reported with which proof codes. Karsten came up with the great idea of making a table summarizing these boundaries (why didn't I think of that? ), so here it is.

First of all:
-Ranges processed under the old sieve files, that is, sieve files produced by NPLB's own team sieves and internal sieving efforts, should have primes reported as "[your name], Srsieve, NPLB, LLR" at the top-5000.

-Ranges processed under the new sieve files, that is, sieve files produced by the joint NPLB/PrimeGrid sieving drive, should have primes reported as "[your name], Srsieve, Psieve, NPLB, LLR" at the top-5000.

The only difference is that the new sieve file was sieved primarily with the psieve software (ppsieve and tpsieve); the sieve was initially started with srsieve, so we still include that.

If you make a mistake in reporting a prime, you can email Chris Caldwell (the maintainer of the top-5000 primes list) to have it manually corrected, though this is a hassle for everyone involved and should be avoided whenever possible.

5th Drive
n=600K-945K: old sieve file
n=945K-1M: new sieve file

6th Drive
n=600K-770K: old sieve file
n=770K-1M: new sieve file

7th Drive
n=600K-770K: old sieve file
n=770K-1M: new sieve file

10th Drive
n=500K-675K: old sieve file
n=675K-1M: new sieve file

11th Drive
n=425K-600K: old sieve file

12th Drive
No top-5000 primes to report, you're off the hook!

13th Drive
n=1M-2M (all): new sieve file

14th Drive
n=1M-2M (all): new sieve file

k=3010-3200 mini-drive
n=420259-580K: old sieve file
n=580K-600K: new sieve file

k=300-400, n=1M-2M individual-k drive
For this drive, it depends on when you were sent your sieve files by the admins. Look at the first line of your file; if it shows "140000000000000" (that's 140 plus 12 zeros) you're on the old sieve file, whereas anything higher than that is on the new sieve file.
Alternately: if you got your file any time in 2011 or later, it's definitely on the new sieve.

Max

[mdettweiler]

While the boundaries stated in the previous post are all you need to know as far as prime reporting goes, the situation is a bit more complex for the admins, who need to know exactly which version of the sieve file is used for particular ranges. My reason for posting this information in the public forum is because it serves as a nice central repository that we admins can refer to without having to dig too much; none of this is of particular relevance to the average participant, unless you're just the naturally curious type.

Note: for the NPLB/PrimeGrid sieving drive files, I will refer to them by their names as assigned by the PrimeGrid folks here.

5th Drive
n=600K-945K: NPLB p=26T file
n=945K-1M: RSP1M_75Ti with 58T-60T factors additionally removed

6th Drive
n=600K-770K: NPLB p=26T file
n=770K-800K: RSP1M_20101103
n=800K-850K: RSP1M_20101110
n=850K-1M: RSP1M_1405T

7th Drive
n=600K-770K: NPLB p=26T file
n=770K-800K: RSP1M_1405T
n=800K-950K: RSP1M_20110207
n=950K-1M: RSP1M_20110407

10th Drive
n=500K-675K: NPLB p=45T file
n=675K-700K: RSP1M_1405T
n=700K-745K: RSP1M_20110207
n=745K+: RSP1M_20110731

11th Drive
n=425K-600K: NPLB p=30T file

12th Drive
n=50K-250K: NPLB p=30T file
n=250K-350K, k=2000-2367: NPLB p=30T file
n=250K-350K, k=2369+: RSP1M_20110731

13th Drive
n=1M-1.047M: RSP2M_20110613
n=1.047M-1.05M: RSP2M_20110731
n=1.05M-1.06M: RSP2M_20110613
n=1.06M+: RSP2M_20110731

14th Drive
n=1M-2M: RSP2M_20110731

k=3010-3200 mini-drive
n=420259-580K: partly Flatlander's p=19T file, partly NPLB p=30T file (switched somewhere in there--not sure where)
n=580K-600K: RSP1M_20101103

k=300-400, n=1M-2M individual-k drive
files sent out before 2010-12-07: NPLB p=140T file
files sent out after 2010-12-07: RSP2M_1405T
files sent out after 2011-10-25: RSP2M_20110731
port 3500 6-k mini-drive: RSP2M_1405T for n=1M-1.59M, RSP2M_20110731 for n>1.59M
(yes, I know this is confusing--thankfully, this is a manual drive so rigorous matching is not quite as necessary)

Gary and I will try to keep a copy of each of these sieve files on hand wherever possible so that we have a master copy somewhere with which to compare results. If any of you other admins are missing one of these and need it, ask one of us and we'll send it to you.

Max