mersenneforum.org

mersenneforum.org (https://www.mersenneforum.org/index.php)
-   Conjectures 'R Us (https://www.mersenneforum.org/forumdisplay.php?f=81)
-   -   Riesel base 3 reservations/statuses/primes (https://www.mersenneforum.org/showthread.php?t=11151)

gd_barnes 2010-02-08 10:13

1 Attachment(s)
KEP had completed k=580M-600M for n<=25K but I wanted to "fill in the hole" for the range on the pages so I did it also. There are 102 k's remaining, which are now listed on the pages. He has the same # of k's remaining and will check against my list, which is also attached here.

I am also in the process of double checking the k=500M-580M range. I should be done in 2 days.

KEP 2010-02-08 17:28

[QUOTE=gd_barnes;204872]KEP had completed k=580M-600M for n<=25K but I wanted to "fill in the hole" for the range on the pages so I did it also. There are 102 k's remaining, which are now listed on the pages. He has the same # of k's remaining and will check against my list, which is also attached here.

I am also in the process of double checking the k=500M-580M range. I should be done in 2 days.[/QUOTE]

Just to let you know, there is a 100% match between my 102 k's and those 102 k's you've listed on the pages :smile:

KEP

gd_barnes 2010-02-09 14:09

[quote=KEP;204918]Just to let you know, there is a 100% match between my 102 k's and those 102 k's you've listed on the pages :smile:

KEP[/quote]

Very good. Nice work.

[quote=gd_barnes;204872]I am also in the process of double checking the k=500M-580M range. I should be done in 2 days.[/quote]

I've completed the double check of k=500M-580M to n=25K. The total count of 460 k's remaining matched but there were 2 problems. This is really very good considering that there were 3 different searchers for the range, that R3 was being transitioned to the Riesel base 3 attack and then back to the regular threads here, and that we didn't have a completely stable starting bases script that automatically removed MOB at that point.

k that I have remaining at n=25K that was not shown previously:
511387054

k that was shown as remaining but that I found a prime for (prime n-value in parens):
519213098 (1760)

I subsequently triple checked k=511387054 to n=25K and it is definitely remaining. (Perhaps it has an n>25K prime?)

At this point, I can't easily tell if these problems were in the original searches, my listing of the k's remaining on the pages, or KEP's listing on his old pages. Regardless, it would only matter if the problems were numerous.

KEP, you'll need to add k=511387054 to and delete k=519213098 from the k's that you are sieving for k=200M-700M / n=25K-100K.

I'm also going to double check the rest of k<=810M to n=25K over the next 1-2 months.

I've been double checking quite a few things lately and this is only one of them. You'll also see a post in the bases 4-32 thread where I missed removing a couple of k's on Sierp base 19 as well as a couple of missing primes that were found on R138 a little while back.


Gary

KEP 2010-02-09 18:06

OK, is adding k=511387054 and removing k=519213098 from the sieve file.

Regards

KEP

KEP 2010-02-11 15:35

1 Attachment(s)
k=511387054 is completed for n=25K+1 to n<=100K :smile:

Attached is the residuals.

Regards

KEP

KEP 2010-02-26 13:12

Detailed and way more factbased status update as of February 26th 2010:

Regarding Riesel base 3, k=3677878: ETA March 16 2009, if a prime isn't found. (< 1190 tests remain or about 1.8 billion iterations)

Regarding Riesel base 3, k>200M to k<700M range: I have removed k=519213098 from testfile aswell from sievefile. k=511387054 hasn't been added to the sievefile, since it was completed seperately shortly after I got aware of the missed k. ETA for completion will be around 31 May 2010.

KEP

gd_barnes 2010-03-24 07:23

I have tested k=700M-800M to n=25K to fill in the hole. 499 k's remaining now shown on the web pages.

Karsten is going to double check me.

gd_barnes 2010-03-25 07:01

KEP is releasing k=3677878 as well as k=200M-700M due to a computer problem.

He reported that k=3677878 is at n=950K but I am in the process of following up on that after reviewing a partial results file. The work for k=200M-700M was lost.

All k's are now available on this base. Here is their search depth:
k=3677878, n=950K
k=4M-200M, n=100K
k=200M-820M, n=25K
all squared k's, n=100K

kar_bon 2010-03-25 16:30

200109022*3^33392-1 is prime

gd_barnes 2010-03-26 06:43

[quote=kar_bon;209504]200109022*3^33392-1 is prime[/quote]

As my daughter would say:

That was random! :smile:

Edit: Or maybe not so random. As the lowest remaining k that has not been searched for n>25K, did you start there and begin searching n>25K one k at a time?

kar_bon 2010-03-26 08:12

[QUOTE=gd_barnes;209559]Edit: Or maybe not so random. As the lowest remaining k that has not been searched for n>25K, did you start there and begin searching n>25K one k at a time?[/QUOTE]

only a small test playing around with. no further reservations.

Note: the page with all R3-primes n> 25k: use <pre>...</pre> instead of <p>...</p> for the prime list and delete all <br>'s
and "<spaces>"-> html is about 5k smaller (18.7k -> 13.4k).

kar_bon 2010-03-26 09:30

1 Attachment(s)
how about this:

KEP 2010-03-26 11:08

[QUOTE=gd_barnes;209472]KEP is releasing k=3677878 as well as k=200M-700M due to a computer problem.

He reported that k=3677878 is at n=950K but I am in the process of following up on that after reviewing a partial results file. The work for k=200M-700M was lost.

All k's are now available on this base. Here is their search depth:
k=3677878, n=950K
k=4M-200M, n=100K
k=200M-820M, n=25K
all squared k's, n=100K[/QUOTE]

k=3677878 is tested to n=950K, even though there was only partial residuals send to you. The residuals that you got was what I long time ago uploaded to my e-mail as a backup, sadly it was the only back-up I ever did. So guys:

[SIZE="6"]Don't do like I did, remember to do occasional backups[/SIZE]:cry:

Take care

KEP

gd_barnes 2010-03-26 16:59

[quote=kar_bon;209570]how about this:[/quote]

Hum. Most interesting. OK, I'll incorporate a variation of that into the pages and do the same for the Sierp side. I think base 3 (and bases 7/15 in the future) are the only bases that it makes sense to do that on. I think some base has k's < 10K that are searched to n=100K or 200K and the rest at n=25K (I think) but that's much more clear. For most bases with varying search depths and/or being searched by various other efforts/projects, I list all the k's individually like on bases 5, 16, and 25, but obviously that's not an option for bases 3/7/15. So your idea is a good alternative. I'll also highlight the squared-k at the end (with the * by it) to show it searched to n=100K.

About saving size on the R3 primes n>25K page, I'll have to do manual changes to it and then see if Microsoft FrontOffice will default to it. Since I usually just copy a spacing character from the end of some other row in order to insert a prime, I think it should work OK.


Gary

gd_barnes 2010-03-26 22:58

k=600M-700M has been doublechecked to n=25K by me. There are 513 k's remaining and no problems were found.

This is very good for such a large amount of testing. Nice work everyone. :smile:

I have now doublechecked or tested all k=500M-820M to n=25K with just the 2 small problems found that were mentioned back a few weeks ago.

I had stated that I would doublecheck all k<820M a couple of months ago and would do it over the next 1-2 month span. Since KEP has now released k=200M-700M for n>25K, I'll finish k<500M at a more leisurely pace over the next few months to a year.

Eventually I'll do the same for the Sierp side to n=25K. Since it is only been tested to k=310M, it won't take too long.


Gary

Siemelink 2010-09-28 18:05

doublechecking
 
Despite 800984378 being listed on the Riesel base 3 reservation page, 800984378*3^5-1 is prime. All the others have no hidden prime with n < 100.

Willem.

gd_barnes 2010-10-01 06:36

[QUOTE=Siemelink;231797]Despite 800984378 being listed on the Riesel base 3 reservation page, 800984378*3^5-1 is prime. All the others have no hidden prime with n < 100.

Willem.[/QUOTE]

The k remaining should be 800983478 not 800984378. I have an automated script to convert k's remaining to comma-delimited format on the pages but I sometimes won't use it for small #'s of k's. In this case, Tim did two k=5M ranges for k=800M-810M and listed the k's correctly. Apparently I manually typed them in on the page with a transposition error.

This has now been corrected on the pages.

An interesting side note of this: I have even double-checked all k=500M-820M to n=25K and still missed the error. I can personally guarantee that the # of k's remaining is correct (assuming that the older version of PFGW didn't miss any primes) for this range because I checked counts for each k=25M range but I cannot 100% guarantee that there is not a typo somewhere on the pages for the range.

I think in the future I will not manually type anything for k's remaining on base 3, no matter how few k's there are. 9 (and soon to be 10) digit k's are too error-prone to type in manually. For Ian's smallish k=10M ranges on the Sierp side, I have been using the automated script so hopefully there shouldn't be such errors like this there.

Thanks for effectively triple checking k=500M-820M Willem! :-)


Gary

Siemelink 2010-10-02 14:22

[QUOTE=gd_barnes;232190]The k remaining should be 800983478 not 800984378..
This has now been corrected on the pages.

An interesting side note of this: I have even double-checked all k=500M-820M to n=25K
..
Thanks for effectively triple checking k=500M-820M Willem! :-)

Gary[/QUOTE]

Excellent.
Did you do the doublecheck with 335 or with 3.3.6? I was playing with 335 vs 336 when I found this. If you've done the range with 336 I'll leave it alone and do the lower half.

Thanks, Willem.

gd_barnes 2010-10-03 07:00

No. I doublechecked it long ago; sometime in late 2009; likely with PFGW version 3.2.7, 3.3.0, or thereabouts.

Fortunately Mark has indicated that smaller bases have little chance of the problem with fftlen's but you are certainly welcome to double/triplecheck some or all k's for n>100 with PFGW 3.4.0 or 3.4.1.

rogue 2010-10-03 12:47

[QUOTE=gd_barnes;232404]No. I doublechecked it long ago; sometime in late 2009; likely with PFGW version 3.2.7, 3.3.0, or thereabouts.

Fortunately Mark has indicated that smaller bases have little chance of the problem with fftlen's but you are certainly welcome to double/triplecheck some or all k's for n>100 with PFGW 3.4.0 or 3.4.1.[/QUOTE]

Note that PFGW 3.4.0/3.4.1 using gwnum v26.2, which has new FFT lengths. If you want to compare FFT lengths then you need to compare 3.3.5 to 3.3.6.

Siemelink 2010-10-03 18:30

[QUOTE=rogue;232416]Note that PFGW 3.4.0/3.4.1 using gwnum v26.2, which has new FFT lengths. If you want to compare FFT lengths then you need to compare 3.3.5 to 3.3.6.[/QUOTE]

I was indeed double checking the FFt lengths between 335 and 336. And the ones that turn out different can be rerun by 341 the quickest.
Were there zeropadded FFTs in 335? I may have gotten confused on versions, despite finding differences between (what I thought) was 335 and 336 runs with -F

Willem.

rogue 2010-10-03 19:39

[QUOTE=Siemelink;232431]Were there zeropadded FFTs in 335?[/QUOTE]

Yes.

henryzz 2010-10-17 13:06

reserving 820-830M to 25k

henryzz 2010-11-29 21:05

What files is it that are being collected for base 3? Is it primes and remaining? I am afraid I don't have results as I forgot the -l.
Would [url]https://www.wetransfer.com/[/url] be good way of transferring large files? I couldn't see any ads when testing it although I do have pretty good ads protection.

mdettweiler 2010-11-29 21:20

[QUOTE=henryzz;239246]What files is it that are being collected for base 3? Is it primes and remaining? I am afraid I don't have results as I forgot the -l.
Would [url]https://www.wetransfer.com/[/url] be good way of transferring large files? I couldn't see any ads when testing it although I do have pretty good ads protection.[/QUOTE]
Primes, remaining k's, and algebraic factor eliminations need to be sent in after "from scratch" work like this; it's mainly for continuations of previous work (usually from n>=25K) that the full residuals are needed.

As far as the [URL]http://www.wetransfer.com/[/URL] web site: interesting. [URL]http://www.sendspace.com/[/URL] is one of the more popular ones on this forum due to its having 300 MB of space, month-long storage times (renewed with each download), and ads that are oh-so-slightly less obnoxious than other such services. :rolleyes: WeTransfer appears to allow transfers up to 2 GB (albeit only for two weeks, though often that's enough), and uses a nontraditional ad system (instead of banners or popups, a kind of "impression advertising" via the website's background image)--though it requires Flash (an issue for some). Pick your poison... :smile:

gd_barnes 2010-11-30 05:09

[QUOTE=henryzz;239246]What files is it that are being collected for base 3? Is it primes and remaining? I am afraid I don't have results as I forgot the -l.
[/QUOTE]

All we need for n<=25 on base 3 is all of the primes and k's remaining. No residuals are needed. There are no algebraic factor eliminations.

henryzz 2010-11-30 16:57

[QUOTE=mdettweiler;239248]As far as the [URL]http://www.wetransfer.com/[/URL] web site: interesting. [URL]http://www.sendspace.com/[/URL] is one of the more popular ones on this forum due to its having 300 MB of space, month-long storage times (renewed with each download), and ads that are oh-so-slightly less obnoxious than other such services. :rolleyes: WeTransfer appears to allow transfers up to 2 GB (albeit only for two weeks, though often that's enough), and uses a nontraditional ad system (instead of banners or popups, a kind of "impression advertising" via the website's background image)--though it requires Flash (an issue for some). Pick your poison... :smile:[/QUOTE]
I hadn't notice that they were ads on wetransfer. I just thought it a picture. Gary complained of popups and many ads when I used sendspace last time. Two weeks is plenty enough for sending results although quite possibly not enough for nfs.
Primes and remaining for 820-830M sent to Gary by WeTransfer email.

It ammuses me how much we dislike the ads considering we are using a free and very useful service.:smile:

mdettweiler 2010-11-30 18:04

[QUOTE=henryzz;239360]I hadn't notice that they were ads on wetransfer. I just thought it a picture. Gary complained of popups and many ads when I used sendspace last time. Two weeks is plenty enough for sending results although quite possibly not enough for nfs.
Primes and remaining for 820-830M sent to Gary by WeTransfer email.

It ammuses me how much we dislike the ads considering we are using a free and very useful service.:smile:[/QUOTE]
Yeah, I didn't notice it being an ad at first either but when I was checking it out I read the info page and they explained it. :smile:

I guess the dislike for the ads comes not so much from their being there when one [I]must[/I] use a service like Sendspace (in which case they're an inevitable consequent of sending a file that size for free); what Gary mentioned to me before as the main annoying thing is when people use Sendspace for files that are well below 10 MB (i.e., can be emailed pretty much regardless of your email provider).

gd_barnes 2010-12-02 11:11

I got the data but haven't looked at it yet. Thanks for sending David.

I'll concur with what Max said. I have no objection to wading through ads when using a useful service for free IF we have to use that service, which we definitely need to in this case. But if the file was < 10 MB, it can be zipped up and sent in an Email so it would be a little annoying to wade through ads for a smaller-sized file.

Puzzle-Peter 2011-02-06 19:19

For a first try, I'll take a rather small range.

Reserving k=830M to 850M to n=25000

Puzzle-Peter 2011-02-08 20:33

This is running faster than I expected.

Reserving 850M<K<900M to n=25k.

gd_barnes 2011-02-09 08:35

Yes, it's surprising how quickly you can blow through the k-ranges for base 3 up to n=25K without sieving.

One annoyance here that you'll come across: In addition to k's remaining, you'll need to send me all the primes and those files are huge. What Ian does is Email them to me in k=10M ranges. I think they are in the 10-15 MB range each when zipped, which is close to as large as gmail allows. Sometimes he has sent them and I haven't gotten them so I think the time that it takes to send them causes them to get lost in cyberspace somewhere.

To be clear, that's all that I need: Primes and k's remaining. No other files are needed.

BTW, I've found that running all cores of a 2.8 Ghz I7 will blow through a k=100M range in about 10-11 days. As a general rule, the I7 is the equivalent of about 5-6 cores of a "standard" quad.


Gary

Puzzle-Peter 2011-02-09 16:09

[QUOTE=gd_barnes;251902]
BTW, I've found that running all cores of a 2.8 Ghz I7 will blow through a k=100M range in about 10-11 days. As a general rule, the I7 is the equivalent of about 5-6 cores of a "standard" quad.

Gary[/QUOTE]

I am running a Q9300 at stock speed (2.5GHz) on this. One core does 10M in 3 days, which would be 100M on all four cores in about 8 days. Is there something wrong with my ranges? Or will running four tasks slow down the machine? I triple checked my entries in the script already...

Do you need pl_primes.txt or pfgw_prime.log? What's the difference anyway? File sizes are ~5% different.

PS.: I tried .zip and .rar, the latter being almost 50% smaller than the first. Is it ok for you to use rar?

mdettweiler 2011-02-09 17:39

[QUOTE=Puzzle-Peter;251947]I am running a Q9300 at stock speed (2.5GHz) on this. One core does 10M in 3 days, which would be 100M on all four cores in about 8 days. Is there something wrong with my ranges? Or will running four tasks slow down the machine? I triple checked my entries in the script already...

Do you need pl_primes.txt or pfgw_prime.log? What's the difference anyway? File sizes are ~5% different.

PS.: I tried .zip and .rar, the latter being almost 50% smaller than the first. Is it ok for you to use rar?[/QUOTE]
Yes, it should be OK to send him .rar files. (IIRC, he installed WinRAR a while back when I tried to send him a .rar file once.)

Also, for files bigger than 10-15 MB, you can use [url=http://www.sendspace.com]Sendspace[/url] to send the files. The incessant banner ads are annoying, but it does work and so far I don't believe they've sold anyone my email address. :rolleyes: It's one of the more popular file-transfer services here on the forum since it lets you send up to 300MB. Another good one I've tried is [url=http://www.wetransfer.com]WeTransfer[/url] which lets you send up to 2 GB and has much less obtrusive "impression advertising" (though it does require Flash for both the sender and recipient).

gd_barnes 2011-02-09 20:15

[QUOTE=Puzzle-Peter;251947]I am running a Q9300 at stock speed (2.5GHz) on this. One core does 10M in 3 days, which would be 100M on all four cores in about 8 days. Is there something wrong with my ranges? Or will running four tasks slow down the machine? I triple checked my entries in the script already...

Do you need pl_primes.txt or pfgw_prime.log? What's the difference anyway? File sizes are ~5% different.

PS.: I tried .zip and .rar, the latter being almost 50% smaller than the first. Is it ok for you to use rar?[/QUOTE]

I may have been slightly mistaken on 10-11 days to run a k=100M range on an I7. That was a rough guess from memory. 8 days sounds about right.

I only need the pl_primes.txt and pl_remain.txt files. pfgw_prime.log would have primes that I don't need in it. The script checks for k - 1 being prime to decide whether to but the k in the pl_MOB.txt file. So k - 1 primes would be in the pfgw_prime.log file, which I don't need. Those primes should be the only difference between the 2 files that you asked about.

As Max said, .rar files work for me. I got your two files and see that they are 6-7 MB each compressed so that works well. I'm glad not to deal with sendspace.


Gary

gd_barnes 2011-02-10 08:41

Peter has completed k=830M-850M to n=25K and has sent the files to me. 104 k's remain.

gd_barnes 2011-02-17 06:20

Peter has completed k=850M-900M to n=25K and has sent the files to me. 257 k's remain.

Mattyp101 2011-02-26 10:48

Reserving k=900M-910M to n=25K

Mattyp101 2011-02-28 10:58

Completed the above range.
42k's remain.
Now, what address do I send the files to?

gd_barnes 2011-02-28 11:04

[QUOTE=Mattyp101;253952]Completed the above range.
42k's remain.
Now, what address do I send the files to?[/QUOTE]

gbarnes017 at
gmail dot com

All that I need is pl_remain.txt and pl_prime.txt. If you'd like, you can include the pl_compPRP.txt file also.

Mattyp101 2011-02-28 11:13

Ok, sent.
There was no pl_compPRP.txt file

Puzzle-Peter 2011-03-03 17:46

Reserving 910M - 930M

Puzzle-Peter 2011-03-05 07:05

Wow, -f30 does make a huge difference compared to -f100!

Reserving 930M - 950M

Results for 910M - 930M will be emailed today.

gd_barnes 2011-03-06 07:41

Peter has completed k=910M-930M to n=25K and has sent the files to me. 87 k's remain.

Puzzle-Peter 2011-03-06 19:22

Reserving 950M - 1000M

Puzzle-Peter 2011-03-07 15:39

Reserving k=3677878 n=950k to n=1M to make a nice round number

Results for k=930k to 950k emailed

MyDogBuster 2011-03-07 18:45

[QUOTE]Reserving k=3677878 n=950k to n=1M to make a nice round number[/QUOTE]

We like round numbers.:smile:

That should give some cores a nice workout.:max:

Puzzle-Peter 2011-03-11 17:31

[QUOTE=Puzzle-Peter;254476]Reserving 950M - 1000M[/QUOTE]

The results are being zipped right now and will be emailed within the hour.

gd_barnes 2011-03-11 23:45

Peter has Emailed the primes/k's remaining for k=930M-950M and 950M-1G to n=25K.

k's remaining:
97 for k=930M-950M
278 for k=950M-1G

Good work Peter. It's nice to have R3 fully tested to k=1G / n=25K.

Puzzle-Peter 2011-03-12 07:50

[QUOTE=gd_barnes;254964]
Good work Peter. It's nice to have R3 fully tested to k=1G / n=25K.[/QUOTE]

I am confident Mark will be able to solve the issues I'm having with running the script in Linux. If / when he does, I might make k=63G my CRUS goal for the year :grin:

rogue 2011-03-12 18:31

[QUOTE=Puzzle-Peter;254985]I am confident Mark will be able to solve the issues I'm having with running the script in Linux. If / when he does, I might make k=63G my CRUS goal for the year :grin:[/QUOTE]

:surprised

The fixed linux version should be ready later today. Note that it is only an issue on linux as it was due to a CPU specific build of GMP. I rebuilt GMP on Steven's box as "generic x86". He was linking with the Ubuntu distribution and I have no idea how that one is built. It will be better going forward as I'll have more control over the version of GMP that gets linked with the linux version.

Mattyp101 2011-03-13 15:57

Great job on getting k to 1G. I'm gonna take some lower ranges up to n=100K, starting with k=200M-210M, and in the longer run to 300M.

Puzzle-Peter 2011-03-13 16:28

Reserving 1G - 1.1G

I'll use this to test Mark's latest modification and - in case I encounter no more errors - evaluate the speed benefit of 64 bit Linux vs. 32 bit Windows. Plus the Linux boxes have higher clock frequencies (well, most of them have). After this, I will decide where to put my long term goal.

Puzzle-Peter 2011-03-14 15:30

[QUOTE=Puzzle-Peter;255070]I'll use this to test Mark's latest modification and - in case I encounter no more errors - evaluate the speed benefit of 64 bit Linux vs. 32 bit Windows. Plus the Linux boxes have higher clock frequencies (well, most of them have). After this, I will decide where to put my long term goal.[/QUOTE]

Seems to be running smoothly, so I fed a few cores. Therefore I reserve 1.1G - 5.7G to n=25k. I created a little script to update min_k, so I can stop in the morning and restart in the evening without too much hassle. Let's see how this works out...

gd_barnes 2011-03-14 18:14

[QUOTE=Puzzle-Peter;255134]Therefore I reserve 1.1G - 5.7G to n=25k.[/QUOTE]

:surprised:surprised:flex::bow:

Puzzle-Peter 2011-03-14 18:25

[QUOTE=gd_barnes;255146]:surprised:surprised:flex::bow:[/QUOTE]

:grin:

It will take several weeks. And we will have to think about a way of exchanging the results. Mailing 10M blocks seems to be rather bothersome...

gd_barnes 2011-03-14 18:31

[QUOTE=Puzzle-Peter;255148]:grin:

It will take several weeks. And we will have to think about a way of exchanging the results. Mailing 10M blocks seems to be rather bothersome...[/QUOTE]

Several weeks? How many cores will you have on it? 30? 40? 50? It took my I7 (equivalent of ~5-6 cores) 8-10 days to do a k=100M range to n=25K. So assuming 40 CPU days to do k=100M means that your range would take 40*57=2080 CPU days. So 40 cores would take ~55-60 days.

Is that in the ball park of what you are thinking?

There is one thing that I want to bring up here: The boredom factor. I and many others have run up against that on these kinds of efforts. If you can stomach finding billions of teeny primes for weeks and months at a time, then it is something that could be done.

Puzzle-Peter 2011-03-14 18:53

[QUOTE=gd_barnes;255152]Several weeks? How many cores will you have on it? 30? 40? 50? It took my I7 (equivalent of ~5-6 cores) 8-10 days to do a k=100M range to n=25K. So assuming 40 CPU days to do k=100M means that your range would take 40*57=2080 CPU days. So 40 cores would take ~55-60 days.

Is that in the ball park of what you are thinking?[/QUOTE]

That is about what I was thinking, yes. I can give better estimates tomorrow when all those machines (50 cores) have crunched for one night, so I can extrapolate the results.

[QUOTE=gd_barnes;255152]
There is one thing that I want to bring up here: The boredom factor. I and many others have run up against that on these kinds of efforts. If you can stomach finding billions of teeny primes for weeks and months at a time, then it is something that could be done.[/QUOTE]

That's why I try to always have several things running at a time. On the other hand I am nearly done with all k's for SR5 covering n ranges adding up to 45000. Before that I spent several months doing nothing but 27121 tests for Prime Grid. So it's not too easy to bore me. Plus I realized you guys don't ask for updates every other day, so I know I can always put a few cores on another task. That helps too.

But I have to admit I am relieved that R301 is close to 25k...

mdettweiler 2011-03-14 19:37

[QUOTE=Puzzle-Peter;255148]:grin:

It will take several weeks. And we will have to think about a way of exchanging the results. Mailing 10M blocks seems to be rather bothersome...[/QUOTE]
For larger files, you can use [url=http://www.sendspace.com]Sendspace[/url], though that too can be a bit of a pain if you have to use it a lot (what with all the ads it has). If you have a very large amount of data you need to transfer to Gary, it would probably be best for me to set you up to make direct SCP/FTP/SFTP transfers into the noprimeleftbehind.net server (from whence Gary can copy them directly to his external USB hard drive, which IIRC is where he keeps all these CRUS results).

Puzzle-Peter 2011-03-15 15:27

Too bad. I had tested this on one of the newer boxes. The others started automatically in the evening and most of them errored out with "FATAL - kernel too old" messages.

Therefore I UNreserve 1.1G - 5.7G and concentrate on other subprojects. Sorry for that! I will return to R3 now and then for small ranges...

Puzzle-Peter 2011-03-16 15:30

1 Attachment(s)
R3 k=3677878 tested from n=950k to n=1M, no prime. k released.

henryzz 2011-03-16 20:37

[QUOTE=Puzzle-Peter;255259]Too bad. I had tested this on one of the newer boxes. The others started automatically in the evening and most of them errored out with "FATAL - kernel too old" messages.

Therefore I UNreserve 1.1G - 5.7G and concentrate on other subprojects. Sorry for that! I will return to R3 now and then for small ranges...[/QUOTE]
It might be possible to compile a version for an older kernel. Unfortunately building pfgw is not trivial AFAIK.

rogue 2011-03-16 22:03

[QUOTE=henryzz;255348]It might be possible to compile a version for an older kernel. Unfortunately building pfgw is not trivial AFAIK.[/QUOTE]

It isn't that bad, but doing it without some guidance (from me) will definitely make it more difficult.

Mattyp101 2011-03-18 16:03

k=200M-210M complete to n=100K.
29 primes found.
13 k's remain.
Results sent.
Continuing to k=300M

Mattyp101 2011-03-30 14:42

k=210M-230M complete to n=100K.
76 primes found.
32 k's remain.
Results sent.
Continuing.

gd_barnes 2011-03-30 22:44

[QUOTE=Mattyp101;257055]k=210M-230M complete to n=100K.
76 primes found.
32 k's remain.
Results sent.
Continuing.[/QUOTE]

k=211243646 had 2 primes listed for it. Therefore only 75 k's were found prime and 33 k's remain for k=210M-230M at n=100K.

gd_barnes 2011-04-06 08:36

Peter reported in Emails dated April 1st completion of k=1G-1.1G to n=25K. 560 k's are remaining for the range. More details on the pages.

Mattyp101 2011-04-14 10:29

1 Attachment(s)
k=230M-240M complete to n=100K
30 primes found
20 k's remain
Results attached

gd_barnes 2011-04-15 06:11

[QUOTE=Mattyp101;258510]k=230M-240M complete to n=100K
30 primes found
20 k's remain
Results attached[/QUOTE]

Hum. I only see primes. Were you going to Email the results to me? Thanks.

Puzzle-Peter 2011-04-30 12:18

Reserving k=1.1G to 1.2G

Puzzle-Peter 2011-05-02 14:56

k=1100M to 1110M done, 51 k's remaining

files emailed

Mattyp101 2011-05-02 15:46

[QUOTE=gd_barnes;258560]Hum. I only see primes. Were you going to Email the results to me? Thanks.[/QUOTE]

I'll send the results along with the next block.

Puzzle-Peter 2011-05-06 14:41

k=1110M to 1125M done, 79 k's remaining

files emailed

Puzzle-Peter 2011-05-08 10:40

R3 k=1125M to 1140M complete, 78 k's remaining

results mailed

Puzzle-Peter 2011-05-11 15:56

R3 k=1140M to 1155M complete, 84 k's remaining

results emailed

Mattyp101 2011-05-13 12:21

k=240M-250M complete to n=100K
32 primes found
18 k's remain
Results sent

Puzzle-Peter 2011-05-13 19:59

k=1155M to 1170M tested to n=25k, 77 k's remaining

Results emailed

Puzzle-Peter 2011-05-15 10:00

k=1170M to 1185M tested to n=25k, 77 k's remaining

Results emailed

henryzz 2011-05-15 14:19

Reserving k=4M-10M upto n=500k
This might take a while but I am willing to share the sieve file if I start getting in someone's way(judging by activity on taking past 100k sofar I won't).

Puzzle-Peter 2011-05-16 14:30

k=1185M to 1200M tested to n=25k, 70 k's remaining

Results emailed

henryzz 2011-05-20 20:23

[QUOTE=henryzz;261478]Reserving k=4M-10M upto n=500k
This might take a while but I am willing to share the sieve file if I start getting in someone's way(judging by activity on taking past 100k sofar I won't).[/QUOTE]
Sieving is going well so far. Sieved upto 450G sofar and removing a candidate every 55 sec. Should sieve upto 4.5T(4T probable because of removals) as test at 340K(60% through) takes 556 secs. Would you suggest splitting off 100M-250M(or 200M) early?

gd_barnes 2011-05-20 20:50

[QUOTE=henryzz;261870]Sieving is going well so far. Sieved upto 450G sofar and removing a candidate every 55 sec. Should sieve upto 4.5T(4T probable because of removals) as test at 340K(60% through) takes 556 secs. Would you suggest splitting off 100M-250M(or 200M) early?[/QUOTE]

Yeah, you might test a candidate at n=190K (60% of n=100K-250K) and see how long that takes, then stop sieving when the removal rate equals that. Then do testing of n=100K-250K. Then remove all k's with primes for the n=250K-500K range, test a candidate at n=400K (60% of n=250K-500K), stop sieving at that removal rate for that n-range, and then finish up testing n=250K-500K. You could also split off n=100K-200K and follow it up with n=200K-500K but I tend to want to put the higher n-ratio (i.e. n-max/n-min) on the lower range when breaking off n-ranges so my preference would be n=100K-250K/250K-500K.

Although it involves more personal hassle, splitting up n-ranges would be a fair amount more CPU-efficient, especially since base 3 is probably the most prime-heavy of all bases.

henryzz 2011-05-20 21:15

[QUOTE=gd_barnes;261872]Yeah, you might test a candidate at n=190K (60% of n=100K-250K) and see how long that takes, then stop sieving when the removal rate equals that. Then do testing of n=100K-250K. Then remove all k's with primes for the n=250K-500K range, test a candidate at n=400K (60% of n=250K-500K), stop sieving at that removal rate for that n-range, and then finish up testing n=250K-500K. You could also split off n=100K-200K and follow it up with n=200K-500K but I tend to want to put the higher n-ratio (i.e. n-max/n-min) on the lower range when breaking off n-ranges so my preference would be n=100K-250K/250K-500K.

Although it involves more personal hassle, splitting up n-ranges would be a fair amount more CPU-efficient, especially since base 3 is probably the most prime-heavy of all bases.[/QUOTE]
I think I will do that. A test at 190M took 145 seconds which would imply a sieve depth of ~1.5T. Any hints on what depth 250k-500k will then be sieved to. What depth if 1 or 2ks are removed? It might be possible to be worked out now based on how many primes in 100k-250k but you should be able to make an educated guess much easier I think. A test at 400k takes 655 seconds.

gd_barnes 2011-05-20 21:40

[QUOTE=henryzz;261874]I think I will do that. A test at 190M took 145 seconds which would imply a sieve depth of ~1.5T. Any hints on what depth 250k-500k will then be sieved to. What depth if 1 or 2ks are removed? It might be possible to be worked out now based on how many primes in 100k-250k but you should be able to make an educated guess much easier I think. A test at 400k takes 655 seconds.[/QUOTE]

No very accurate hints. It can vary greatly depending on how many k's are found prime for the lower n-range. If optimum sieving for n=100K-250K is 1.5T, for n=250K-500K it might be 3-4T or 8T-10T. I remember one base where I hardly had to sieve the higher range at all because the lower range had so many primes. It's usually between 2 and 8 times the lower n-range depth if the initial n-max to n-min ratio is in the 4 to 5 range, i.e. here 500K / 100K = 5.

BTW, on several occassions you've used "M" in place of "k" for the search depth. Just thought I'd point that out.

henryzz 2011-05-21 10:38

[QUOTE=gd_barnes;261875]No very accurate hints. It can vary greatly depending on how many k's are found prime for the lower n-range. If optimum sieving for n=100K-250K is 1.5T, for n=250K-500K it might be 3-4T or 8T-10T. I remember one base where I hardly had to sieve the higher range at all because the lower range had so many primes. It's usually between 2 and 8 times the lower n-range depth if the initial n-max to n-min ratio is in the 4 to 5 range, i.e. here 500K / 100K = 5.[/QUOTE]
Thanks for the info. The possible 8T-10T search depth is the reason this helps a lot of the time. Hopefully I will find primes making it 3-4T.

[QUOTE=gd_barnes;261875]BTW, on several occassions you've used "M" in place of "k" for the search depth. Just thought I'd point that out.[/QUOTE]
Whoops! Obviously k was meant. I will watch out for this in the future.

henryzz 2011-05-21 10:56

Just noticed that k=200M-250M at n=100k has 84 ks compared to 224ks. That indicates multiplying n by 4 leaves 37.5% of ks remaining. I am multiplying n by 5 but they are also probably lower weight ks so it sounds like I should find 3-4 primes. I had expected 1-2 priimes so more will be welcome if they appear.

henryzz 2011-06-01 20:00

Sieving is complete upto 1.5T. I will load 100k-250k into a personal prpnet server.

henryzz 2011-06-14 19:27

[QUOTE=henryzz;262804]Sieving is complete upto 1.5T. I will load 100k-250k into a personal prpnet server.[/QUOTE]
I have reached 200k so far with no primes.

henryzz 2011-06-27 17:05

[QUOTE=henryzz;263789]I have reached 200k so far with no primes.[/QUOTE]
I have reached n=232k.
5309716*3^231434-1 is prime

henryzz 2011-07-03 16:24

[QUOTE=henryzz;264768]I have reached n=232k.
5309716*3^231434-1 is prime[/QUOTE]
Now complete to 250k. Continuing after more sieving.
Another new prime:
7985602*3^243614-1
Results emailed

henryzz 2011-07-15 17:53

Another prime 7328954*3^250101-1
That means 3 primes within 20k.

gd_barnes 2011-07-16 05:24

[QUOTE=henryzz;266514]Another prime 7328954*3^250101-1
That means 3 primes within 20k.[/QUOTE]

Wow and only 4 k's remain for k<10M for R3. Nice! :smile:

henryzz 2011-07-19 16:13

[QUOTE=gd_barnes;266570]Wow and only 4 k's remain for k<10M for R3. Nice! :smile:[/QUOTE]
And the prime run continues:
5318998*3^267482-1 is prime

gd_barnes 2011-07-30 09:15

k=250M-300M has been released due to lack of activity or response to PMs/Email.

henryzz 2011-08-16 09:57

[QUOTE=henryzz;266933]And the prime run continues:
5318998*3^267482-1 is prime[/QUOTE]
Tested to 400k. No more primes.:no:

Puzzle-Peter 2011-08-27 19:24

Reserving k=1.2G - 1.3G to n=25k.

Puzzle-Peter 2011-08-31 14:21

Reserving k=1.3G - 1.5G to n=25k.


All times are UTC. The time now is 09:00.

Powered by vBulletin® Version 3.8.11
Copyright ©2000 - 2022, Jelsoft Enterprises Ltd.