PDA

View Full Version : Sieve Introduction?



tandrews
03-15-2007, 01:28 PM
Any recommended reading references that detail exactly what "sieving" is (I have only a rough idea) as it applies to this project?

hhh
03-16-2007, 05:33 AM
That's how it works (basically):
You take a small prime number (e.g. 2) and look at all the k,n-pairs left in your list if they are divided by this number. These get kicked out. Then you take the next prime number, 3. Then 5. And so on. Right now, we are at around 1000000000000000. Don't blame me if there are a few 0's too much or too less, I didn't count. Should be around 2^50.
In fact it's no that easy, but your question was not very precise either.
Sieving is very fast, but you never know which k,n-pair will be removed next.
So it can happen that you find a factor for a number that has already been tested, or that will be tested in ten years only. Nevertheless, it is very effective and without it the project would be pointless.

H.

Death
03-29-2007, 05:21 AM
PRS52 v0.102b
P Range 666666000000000 <= p <= 666766000000000
N Range 991 <= n <= 50000000

PRS52 v0.102b
P Range 666677924999938 <= p <= 666766000000000
N Range 991 <= n <= 50000000

PRS52 v0.102b
P Range 666678124999918 <= p <= 666766000000000
N Range 991 <= n <= 50000000

PRS52 v0.102b
P Range 666681049999906 <= p <= 666766000000000
N Range 991 <= n <= 50000000

PRS52 v0.102b
P Range 666681174999882 <= p <= 666766000000000
N Range 991 <= n <= 50000000

PRS52 v0.102b
P Range 666682749999822 <= p <= 666766000000000
N Range 991 <= n <= 50000000

PRS52 v0.102b
P Range 666696249999802 <= p <= 666766000000000
N Range 991 <= n <= 50000000

PRS52 v0.102b
P Range 666698199999786 <= p <= 666766000000000
N Range 991 <= n <= 50000000

PRS52 v0.102b
P Range 666698199999786 <= p <= 666766000000000
N Range 991 <= n <= 50000000

PRS52 v0.102b
P Range 666732824999782 <= p <= 666766000000000
N Range 991 <= n <= 50000000

PRS52 v0.102b
P Range 666735749999746 <= p <= 666766000000000
N Range 991 <= n <= 50000000

PRS52 v0.102b
P Range 666737574999562 <= p <= 666766000000000
N Range 991 <= n <= 50000000

PRS52 v0.102b
P Range 666751874999466 <= p <= 666766000000000
N Range 991 <= n <= 50000000

PRS52 v0.102b
P Range 666752149999426 <= p <= 666766000000000
N Range 991 <= n <= 50000000

PRS52 v0.102b
P Range 666752524999398 <= p <= 666766000000000
N Range 991 <= n <= 50000000

PRS52 v0.102b
P Range 666765524999380 <= p <= 666766000000000
N Range 991 <= n <= 50000000
Statistics :
pmin : 666765524999380
pmax : 666766000000000
# Tested p : 13913501
# Tested k : 111308008
# Whacked k : 88218874
# Whacked k1 : 0
# Whacked k2 : 0
# Whacked k3 : 1994555
# Whacked p : 5078407
Setup time: 3.904603 seconds.
Sieve time: 979.622939 seconds.
Sieve rate: 484.881 kp/sec.
Next p to process 666766000000013

does this mean that's there no factors in this range? what does mean k1 k2 k3

Joe O
03-29-2007, 03:00 PM
There is no way to tell from this output whether you found any factors or not. You would have to look in the fact.txt file to see if there were any factors found.

As for the k1, k2, and k3, they are three locations in the code where a decision is made whether or not to continue looking at a particular k for the current p. If the program determines that a particular p can never be a factor of k.2^n+-1 for any n, then it stops that k and goes on to the next k.

jasong
03-29-2007, 06:23 PM
I love how the word "whack" is used.

"Oh, I see, you refuse to talk? Vinnie, get the baseball bat, I want take make absolutely sure there are no k/n pairs in this p."