P-1 coordination thread - discussion

[vjs]

Garo You stated before you were only going to do 4847

If this is the case I'll take 10223 and 19249

[garo]

yes vjs please go ahead!!

[vjs]

Great the first few 10223's are using 768k so I'll probably proceed until I hit the 1024 with these.

[Nuri]

vjs are you talking about the 9700000 - 9750000 only, or is it on a larger scale?

please also indicate the n ranges so that other users can keep track as well.

cheeers
nuri

[garo]

I believe he is only talking about the 9700000-9750000 range.

[Nuri]

ok. thx

[Nuri]

I'll also start some unreserved 4847s btw 9m & 10m when I'm finished with my current ranges. I'll let you know so that there's no duplication of work

[vjs]

Garo's correct Nuri,

I should have been more descriptive so everyone can know what I was doing.

Garo had said he was only going to do 4847, so I thought it was a good oportunity to grab a few tests. I'll probably only do 10223.

Basically I didn't want to do many tests, I'm going to be using fairly large bound for these.

B1=120K stage1 only...

As they come close to PRP I'm going to do Stage2.

Yes it's probably a waste of time but I just want to see what happens. 8 tests already no joy.

[Nuri]

good luck there... ;)

[vjs]

Woot! Found one

P-1 found a factor in stage #2, B1=85000, B2=1200000
10223*2^9945197+1 has a factor: 322554042371
3225 540423 715992 = 2 ^ 3 x 3 ^ 2 x 389 x 433 x 1187 x 224069

[Mystwalker]

10223*2^9945197+1 has a factor: 322554042371

Correct me if I'm wrong, but shouldn't this factor have been found a while ago by sieving (--> 322G)?

[Nuri]

a simple coy paste error. ;)

it has 5993 at the end, so it's 3225T

[garo]

Nuri: Are you doing any 4847s or 10223s? I have finished with 4847s on 970-975 and am now doing all 10223's above 972, ahead of PRP testing which BTW is already at 9716500!!

[Nuri]

Just go ahead...

I'm interested in doing all undone 67607s within 9m-10m range AFTER PRP reaches 10m, hoping to reduce their number below 1000.

It's a though task though.

May be I'll attack 4m-5m range or 6m-7m range instead....

I'll have to check with the odds and CPU time required first.

[vjs]

We seem to have a little bit of a lead on prp at the moment.

If anyone wish to use higher bounds like Greenbank please feel free.

Using Prime95 the worktodo.ini could be change to the following.

From

Pfactor=19249,2,9947282,1,49.8,1.7

(basically prime 95 will assign bounds of B1 and B2 dependant on the 49.4,1.7 you chose.

Knowing these bounds, if you'd like to increase them the worktodo.ini lines have to be changed to reflex your desired bounds

Pminus1=19249,2,9947282,1,50000,700000

Increasing stage1 does not require additional memory, where as increasing B2 will. For my machine time increasing linearly with bX-values, b1=60K takes twice as long as B2=30K. Same is true for B2.

If also found using B2=14 x B1 each test takes an equal time in each stage.

Your experiences will vary, I have not yet found any additional factors using
B1=85K
B2=1700k

that would not have been found using B1=45K B2=500k in half the time.

-----------------------

Also if anyone wishes to state their B1,B2 or prime95 values I will add them after the [reserved/complete]b1,b2 or 49.8,1.7

[OmbooHankvald]

When I tried sieving depth 49.5 and factor value 1.7 Prime95 gave me b1=30000 b2=30000

? Is that supposed to happen=?

[Greenbank]

Well I've taken those 24 values in the range and I'm going to apply as high a bounds as possible in the remaining time.

PRP is running at about a 7000 increase in n every 24 hours.

So from now (9779651) to the start of my range (10100000) the n difference is about 320349.

At about 7000 n a day that's about 45 days.

So that's 45 hours I can give each k,n pair.

On my machine B1=45000 takes just over 3 hours (I'm using manual bounds to prevent B2 testing).

I've got a script that endlessly loops increasing B1 by 10000 and then running the P-1 program. (The script also checks to see if anything appears in fact.txt and makes an appropriate entry in results.txt so that it is filtered out on the next iteration).

When I get near the end of this time limit I'll manually edit in a B2 range and see what happens.

[vjs]

I'd suggest increasing by more than 10K, I think it will take a significant portion of your time to load start the test etc.