Delivered-To: luke@ndatech.com Date: Fri, 14 Dec 2001 10:42:02 -0800 From: mersenne-digest-invalid-reply-address@base.com (Mersenne Digest) To: mersenne-digest@base.com Subject: Mersenne Digest V1 #919 Reply-To: mersenne@base.com Sender: mersenne-digest-invalid-reply-address@base.com Mersenne Digest Friday, December 14 2001 Volume 01 : Number 919 ---------------------------------------------------------------------- Date: Wed, 12 Dec 2001 12:13:47 -0500 From: George Woltman Subject: RE: Mersenne: Database merge At 08:09 AM 12/12/2001 -0800, Aaron Blosser wrote: >Just out of curiosity though, I'm wondering why some of my cleared >exponents are still there... like these ones: > >8664941 65 0x87838CA31D64B5__ 16-Jun-00 07:44 WorkerBoy2 >9885119 65 0x86BA14ED5C3785__ 25-Dec-00 13:45 NYFS-1 Merging is a messy inexact little-tested science. The binary database has a list of all exponents below 10,000,000 that need double-checking. Thus, these exponents were not removed from the server (and the old LL result is still visible even though we should have cleared it). _________________________________________________________________________ Unsubscribe & list info -- http://www.ndatech.com/mersenne/signup.htm Mersenne Prime FAQ -- http://www.tasam.com/~lrwiman/FAQ-mers ------------------------------ Date: Wed, 12 Dec 2001 09:33:50 -0800 From: "John R Pierce" Subject: Re: Mersenne: Database merge > Just out of curiosity though, I'm wondering why some of my cleared > exponents are still there... like these ones: yeah, me too... ------- Exponents Cleared since last Synchronization ------- prime fact Lucas-Lehmer residue or factor exponent bits [residues partially masked] date returned computer ID - -------- ---- -- -------------------------------- --------------- -------- - ---- 9375469 65 0x489597E6ADA791__ 23-Nov-00 19:29 p2-400-NTb 9383323 64 0xA5D5F903C745B4__ 02-Jun-01 01:24 p6-200-NT-B 9455059 64 0x0005A222568021__ 28-Nov-00 00:42 p6-200-NT-B 9728249 64 0x7E3112226FEE55__ 25-Nov-00 08:19 dell4400 9956579 64 0x30613BF765EEA3__ 08-Feb-01 17:03 dell4400b btw. phew, put a new machine up *yesterday*, a P4-1.7Mhz, and its ALREADY returned a factor of a 2^1.4M range number. - -jrp _________________________________________________________________________ Unsubscribe & list info -- http://www.ndatech.com/mersenne/signup.htm Mersenne Prime FAQ -- http://www.tasam.com/~lrwiman/FAQ-mers ------------------------------ Date: Wed, 12 Dec 2001 12:41:22 -0800 (PST) From: Mary Conner Subject: Mersenne: Moving an assignment How would I go about giving an exponent I've been assigned to someone else? From previous discussions, it seems as though PrimeNet would reject the assignment as not belonging to him, but I've seen exponents moved from one account to another without expiring, so I know it can be done. I haven't been able to find any option on Prime95's menus, or anything on the web page about how to do this so that PrimeNet will accept the reassignment. Mary Conner _________________________________________________________________________ Unsubscribe & list info -- http://www.ndatech.com/mersenne/signup.htm Mersenne Prime FAQ -- http://www.tasam.com/~lrwiman/FAQ-mers ------------------------------ Date: Wed, 12 Dec 2001 17:17:04 -0500 From: Nathan Russell Subject: Re: Mersenne: Moving an assignment At 12:41 PM 12/12/2001 -0800, Mary Conner wrote: >How would I go about giving an exponent I've been assigned to someone >else? From previous discussions, it seems as though PrimeNet would reject >the assignment as not belonging to him, but I've seen exponents moved from >one account to another without expiring, so I know it can be done. I >haven't been able to find any option on Prime95's menus, or anything on >the web page about how to do this so that PrimeNet will accept the >reassignment. To begin with, several important people in the GIMPS project, and many people who know far more than I do about it, read this list. It'd likely be best not to follow my advice until you see if any of them have a better idea. That said, the server only assigns a new assignment every few minutes. If you were to release the assignment at a time of the day when the server isn't busy (check the assignments out report to get a feeling for when that is) and, coordinating by instant messaging, on the phone, etc, the other user were to (having increased his number of days of work to queue significantly) manually request work (in the advanced menu), there's a fair chance he'd get the exponent you'd just released. You could then send him the save file. The other alternative, of course, and likely a better one, would be to email the folks at PrimeNet and ask them if they could manually transfer the assignment to the other user's account. Nathan _________________________________________________________________________ Unsubscribe & list info -- http://www.ndatech.com/mersenne/signup.htm Mersenne Prime FAQ -- http://www.tasam.com/~lrwiman/FAQ-mers ------------------------------ Date: Wed, 12 Dec 2001 15:35:18 -0800 From: "John R Pierce" Subject: Re: Mersenne: Moving an assignment > How would I go about giving an exponent I've been assigned to someone > else? From previous discussions, it seems as though PrimeNet would reject > the assignment as not belonging to him, but I've seen exponents moved from > one account to another without expiring, so I know it can be done. I > haven't been able to find any option on Prime95's menus, or anything on > the web page about how to do this so that PrimeNet will accept the > reassignment. AFAIK, take it out of your worktodo.ini file, and hand it off to teh other user, along with any Q or P files if they exist. when they check back into the server, it should be happy to let them continue. I could be wrong here, but I don't think it cares. - -jrp _________________________________________________________________________ Unsubscribe & list info -- http://www.ndatech.com/mersenne/signup.htm Mersenne Prime FAQ -- http://www.tasam.com/~lrwiman/FAQ-mers ------------------------------ Date: Wed, 12 Dec 2001 15:54:34 -0800 From: "Aaron Blosser" Subject: RE: Mersenne: Moving an assignment If it's just one exponent, cut and paste that line from the worktodo.ini file and copy your temp files (p????? and q?????) to the other machine. And just start 'er up. Even though it's not assigned to you, an exponent that has already been started work on will not be released. It'll just complain each time you update results that the exponent is not assigned to you, but it will still finish it up and you'll still get credit for it when you're all done. I know this from my "poaching" days. ;) If work hasn't already begun on it, then there's probably no good reason to transfer it anyway, so I'm sure we're talking about work in progress. Much easier than the methods below. :) Aaron > -----Original Message----- > From: mersenne-invalid-reply-address@base.com [mailto:mersenne-invalid- > reply-address@base.com] On Behalf Of Nathan Russell > Sent: Wednesday, December 12, 2001 2:17 PM > To: Mary Conner; mersenne@base.com > Subject: Re: Mersenne: Moving an assignment > > At 12:41 PM 12/12/2001 -0800, Mary Conner wrote: > >How would I go about giving an exponent I've been assigned to someone > >else? From previous discussions, it seems as though PrimeNet would > reject > >the assignment as not belonging to him, but I've seen exponents moved > from > >one account to another without expiring, so I know it can be done. I > >haven't been able to find any option on Prime95's menus, or anything on > >the web page about how to do this so that PrimeNet will accept the > >reassignment. > > To begin with, several important people in the GIMPS project, and many > people who know far more than I do about it, read this list. > > It'd likely be best not to follow my advice until you see if any of them > have a better idea. > > That said, the server only assigns a new assignment every few minutes. If > you were to release the assignment at a time of the day when the server > isn't busy (check the assignments out report to get a feeling for when > that > is) and, coordinating by instant messaging, on the phone, etc, the other > user were to (having increased his number of days of work to queue > significantly) manually request work (in the advanced menu), there's a > fair > chance he'd get the exponent you'd just released. You could then send him > the save file. > > The other alternative, of course, and likely a better one, would be to > email the folks at PrimeNet and ask them if they could manually transfer > the assignment to the other user's account. > > Nathan > > ________________________________________________________________________ _ > Unsubscribe & list info -- http://www.ndatech.com/mersenne/signup.htm > Mersenne Prime FAQ -- http://www.tasam.com/~lrwiman/FAQ-mers _________________________________________________________________________ Unsubscribe & list info -- http://www.ndatech.com/mersenne/signup.htm Mersenne Prime FAQ -- http://www.tasam.com/~lrwiman/FAQ-mers ------------------------------ Date: Wed, 12 Dec 2001 18:47:07 -0500 From: George Woltman Subject: Re: Mersenne: Why is the default page... At 06:35 PM 12/9/2001 +0000, Daran wrote: >... of the GIMPS website http://www.mersenne.org/ a list of other >distributed computing projects? By all means link to and support these >projects, but wouldn't prime.htm be a more sensible starting point? I've fixed it. Let me know if any links are now broken. Thanks _________________________________________________________________________ Unsubscribe & list info -- http://www.ndatech.com/mersenne/signup.htm Mersenne Prime FAQ -- http://www.tasam.com/~lrwiman/FAQ-mers ------------------------------ Date: Wed, 12 Dec 2001 17:02:24 -0800 (PST) From: Mary Conner Subject: Re: Mersenne: Moving an assignment On Wed, 12 Dec 2001, John R Pierce wrote: > AFAIK, take it out of your worktodo.ini file, and hand it off to teh other > user, along with any Q or P files if they exist. > > when they check back into the server, it should be happy to let them > continue. I could be wrong here, but I don't think it cares. I surely hope not, or that would allow someone to steal an exponent from someone else. I had assumed that the exponents that were changing hands on the assignment report were a bunch of guys helping a buddy out with some low exponent double checks, but if the server just lets someone start reporting an exponent they weren't assigned, they could be stolen. _________________________________________________________________________ Unsubscribe & list info -- http://www.ndatech.com/mersenne/signup.htm Mersenne Prime FAQ -- http://www.tasam.com/~lrwiman/FAQ-mers ------------------------------ Date: Wed, 12 Dec 2001 18:13:52 -0800 From: "Aaron Blosser" Subject: RE: Mersenne: Moving an assignment I think we settled on the term "poached" long ago. There's really nothing stopping anyone from working on any exponent they want. If they communicate with Primenet and it reports that an exponent is assigned to someone else, the software will remove that exponent from your worktodo list. The only exception is if that exponent has already begun work, in which case it will still report the error, but it remains in the worktodo file and when done, it will show up in your cleared exponent list. I broached the 'taboo' topic of poaching exponents quite some years ago when I mentioned that sometimes, exponents are checked out and no work is ever done on them, so I was "helping out" by snagging those and just finishing them off. However, at the time, there was no expiration period of 60 days, so sometimes exponents were many months or even over a year old, had never had any check-ins, and were only cleared when Scott or whoever would manually go through and clear those out. Nowadays, exponents must be checked in every 60 days with some sort of status update, otherwise it's automatically released back into the pool. I can only imagine that our poaching thread from back when is what prompted that change in the procedure, and it's a good change. I don't recall the subject of deliberate exponent poaching coming up at all recently, except musings about what if two people work on the same exponent unknowingly, and it happens to be prime. :) But that was more of an 'accidental' poaching, when a client hasn't checked in for a very long time but is actually still working on a number, and meanwhile some other machine gets the same one and finishes it as well. It's also worth noting that GIMPS/Primenet has no claim on the exponents being handed out. They are simply prime numbers which anyone can run an LL test on. GIMPS/Primenet is great because it coordinates the efforts of thousands. Way back when, George kept the master list... you'd download a file that had untested #'s (George would update that every so often), then email George telling him what number you were working on, then email him back the results when it was done. Man... primenet is SO much better! So much better that it landed me in a bit of trouble, but oh well, still a vast improvement. :) Aaron > -----Original Message----- > From: mersenne-invalid-reply-address@base.com [mailto:mersenne-invalid- > reply-address@base.com] On Behalf Of Mary Conner > Sent: Wednesday, December 12, 2001 5:02 PM > To: mersenne@base.com > Subject: Re: Mersenne: Moving an assignment > > On Wed, 12 Dec 2001, John R Pierce wrote: > > AFAIK, take it out of your worktodo.ini file, and hand it off to teh > other > > user, along with any Q or P files if they exist. > > > > when they check back into the server, it should be happy to let them > > continue. I could be wrong here, but I don't think it cares. > > I surely hope not, or that would allow someone to steal an exponent from > someone else. I had assumed that the exponents that were changing hands > on the assignment report were a bunch of guys helping a buddy out with > some low exponent double checks, but if the server just lets someone start > reporting an exponent they weren't assigned, they could be stolen. _________________________________________________________________________ Unsubscribe & list info -- http://www.ndatech.com/mersenne/signup.htm Mersenne Prime FAQ -- http://www.tasam.com/~lrwiman/FAQ-mers ------------------------------ Date: Thu, 13 Dec 2001 04:27:12 +0000 (GMT) From: Russel Brooks Subject: Re: Mersenne: Teams and AOL Philip WHITTINGTON wrote: > If you have multiple PCs and other friends who want to join GIMPS why don't > you just make each PC and each of your friends create their own account; > they will still be donating as much processing power to the problem, and it > would seem to me that this is equally beneficial. I am suggesting they join gimps, as individuals if they wish. If they have no interest in stats or top producers list I'm offering to have them join my 'team' so I can get it. My primary objective is to get their pcs running Prime95; stat credit is a minor secondary concern. Cheers... Russ _________________________________________________________________________ Unsubscribe & list info -- http://www.ndatech.com/mersenne/signup.htm Mersenne Prime FAQ -- http://www.tasam.com/~lrwiman/FAQ-mers ------------------------------ Date: Thu, 13 Dec 2001 07:30:19 -0000 From: bjb@bbhvig.uklinux.net Subject: Re: Mersenne: Moving an assignment On 12 Dec 2001, at 12:41, Mary Conner wrote: > How would I go about giving an exponent I've been assigned to someone > else? From previous discussions, it seems as though PrimeNet would reject > the assignment as not belonging to him, but I've seen exponents moved from > one account to another without expiring, so I know it can be done. I > haven't been able to find any option on Prime95's menus, or anything on > the web page about how to do this so that PrimeNet will accept the > reassignment. Simple - just let the other person run the assignment. The server will accept the result though it will probably mutter about "assignment not valid for this user", and CPU credit will not be awarded in the PrimeNet table. Regards Brian Beesley _________________________________________________________________________ Unsubscribe & list info -- http://www.ndatech.com/mersenne/signup.htm Mersenne Prime FAQ -- http://www.tasam.com/~lrwiman/FAQ-mers ------------------------------ Date: Thu, 13 Dec 2001 07:36:00 -0000 From: "Daran" Subject: Re: Mersenne: P-1 Stage 2 - ----- Original Message ----- From: "Alexander Kruppa" To: "Daran" Cc: Sent: Wednesday, December 12, 2001 1:39 AM Subject: Re: Mersenne: P-1 Stage 2 > Daran wrote: > > So if we were using P-1 to factorise 1000 using B1=15 we would calculate > > > > E = 2^9 * 3^6 * 5^4 * 7^3 * 11^2 * 13^2 > > > > Or did you mean > > > > E = 2^3 * 3^2 * 5 * 7 * 11 * 13 ? > > The programmer of the P-1 algorithm is free to choose, either way will > find factors. > But in practice one uses the second alternative for efficiency. A prime > power p^n has a probability of 1/p^n of dividing a random integer, so we > include all those primes and prime powers with a probability >= 1/B1, > that is all primes and prime powers p^n <= B1. Of course. The math.htm page is misleading, both in its explanation of the algorithm, and its definition of 'smooth'. This also throws a spanner into the works of another idea I've had. [...] > Actually, it wouldn't work well. If we wanted to allow two large primes > in factor-1, one from [B1, B2] and one from [B1, B3], we have to include > all _combinations_ of such primes, i.e. > > for all p_1 in [B1, B2] > compute x^p_1 > for all p_2 in [B1, B3] > compute (x^p_1)^p_2 Which is the same as (x^p_2)^p_1, so this is easily improved by up to a factor of 2. > multiply that to accu > gcd(accu, N) > > The inner loop would be iterated (pi(B2)-pi(B1))*(pi(B3)-pi(B1)) times, > where pi(n) is the # of primes <=n, and that would be a terribly large > number. I haven't heard of a practical three-stage algorithm yet. Yes. At first I thought you could decouple the loops by replacing x with the result of the first stage 2, but of course, all those -1's spoil that idea. Still, I can't help thinking there ought to be a better way of doing it than this, though I can't see what it might be. I just feel that inside of every O(N^2) algorithm is an O(N*log2(N)) algorithm trying to get out. However, there's another generalisation that occurs to me that would be quite cheap to implement. The purpose of stage two is to catch those factors which are 'just out of reach' of stage 1. However another way a factor could be 'just out of reach' would be if the power of exactly one of the primes < B1 was exactly 1 more than stage 1 would find. Stage 2 would find these factors *as well* if the algorithm were run for primes in [2,B2] instead of [B1,B2] Has this been investigated? > Alex Daran _________________________________________________________________________ Unsubscribe & list info -- http://www.ndatech.com/mersenne/signup.htm Mersenne Prime FAQ -- http://www.tasam.com/~lrwiman/FAQ-mers ------------------------------ Date: Thu, 13 Dec 2001 13:44:25 -0500 From: Jeff Woods Subject: Mersenne: CNET coverage of M39: Very good Fairly decent and accurate coverage on news.com (CNet) at the following link. Apologies for the HTML that is embedded in the text pasted in below. If your mail software can't read the HTML, follow the link to the original site. http://news.cnet.com/news/0-1003-200-8159590.html?tag=tp_pr Distributed computing strikes gold By Stephen Shankland Staff Writer, CNET News.com December 13, 2001, 4:00 a.m. PT A 20-year-old in Owen Sound, Canada, has found the world's largest known prime number using a mere desktop computer. But he didn't work alone: His system was part of a 210,000-machine quasi-supercomputer stretched across the globe. Using a computer with an 800MHz chip from Advanced Micro Devices, Michael Cameron found the prime number on Nov. 14, according to Entropia. The San Diego company sells software to enable "distributed computing," which harnesses the unused processing abilities of computers scattered across the Internet. Although the arrival of profit motive has transformed distributed computing, its roots remain in academic pursuits such finding optimal Golomb rulers or alien radio signals. Cameron's computer found the number, but he shares credit with others: George Woltman, who founded the Great Internet Mersenne Prime Search (GIMPS) and wrote the search software, and Entropia founder Scott Korowski, who created the network system called PrimeNet that governs the 210,000 computers that are part of the effort. Prime numbers, once a mathematical curiosity but now crucial to encrypted communications, are numbers greater than one that are divisible only by one and the number itself. Cameron was participating in a project to search for a particular type of prime number called a Mersenne prime. The number that Cameron discovered--2 to the 13,466,917th power minus 1--has 4,053,946 digits. In order to cram his discovery onto a 29-inch-by-40-inch poster sold by Perfectly Scientific, the number is printed in a tiny 1.37-point font and read with a magnifying glass. Mersenne primes are named after Marin Mersenne, a French monk born in 1588 who investigated a particular type of prime number: 2 to the power of "p" minus one, in which "p" is an ordinary prime number. Mersenne primes are much rarer than ordinary primes. The GIMPS effort, exhaustively searching for possible candidates since 1996, has been responsible for discovering the five most recent examples. Altogether, 39 have been discovered so far. Cameron's computer took 42 days to verify that the number was a Mersenne prime. After that, researchers using a workstation took three weeks to confirm the work. Prime numbers are needed for encrypted communications such as a Web browser's Secure Sockets Layer (SSL) technology that makes it harder to sniff out credit card numbers or other private information. But those systems typically use primes that are merely 300 or so digits, said Stanford University mathematician Dan Boneh. "The large Mersenne primes are not very useful," Boneh said, though finding one will grant a person 15 minutes of fame. Mathematical hobbyists have provided online versions of Cameron's number written out in decimal form or in words. Searching for Mersenne primes is computationally intense, but it is a problem that's known as "embarrassingly parallel," which means it can easily be broken down into independent parts that separate computers tackle. Many supercomputer problems take another form, requiring high-speed communication between separate computers or requiring that a problem be solved one step at a time with little opportunity for sharing among many systems. Parallel computing tasks aren't merely academic. Sun Microsystems and Intel use distributed computing software to help design microprocessors, and companies such as Entropia, Turbolinux, Platform Computing, Parabon Computation and United Devices have software that can be used for work in genetics, pharmaceuticals or financial services. Typically, this software is used within a single corporation rather than on strangers' computers across the Internet. The concept of distributed computing is closely related to "grid" computing, which unites computers and storage systems into a single pool of resources. The National Science Foundation is among those interested in the concept, devoting $53 million to one grid. Entropia, IBM, Sun, Platform Computing and others are working with the open-source Globus Project to define standards and software for controlling grids. Ultimately, researchers envision a future in which all computers are linked into a single mammoth resource that can be tapped when needed. The Mersenne prime search is moving in that direction. Each day, its network of computers does work that would take a single 90MHz Pentium computer 200 years to accomplish. On average, the network of computers performs 2.4 trillion calculations per second. The most popular model in the prime number hunt is a computer with an Intel Pentium III chip, with AMD Athlon chips coming in a close second. Although the search effort is voluntary, there is an incentive to participate. The Electronic Frontier Foundation paid $50,000 to GIMPS participant Nayan Hajratwala of Plymouth, Mich., upon discovery of the first prime number with more than a million digits. Funded by an individual's specific donation, the organization also is offering $100,000 for the first 10 million digit prime, $150,000 for the first 100 million digit prime, and $250,000 for the first 1 billion digit prime. Searchers are now looking for the 40th Mersenne prime. Software for the task can be downloaded from the Mersenne Web site. _________________________________________________________________________ Unsubscribe & list info -- http://www.ndatech.com/mersenne/signup.htm Mersenne Prime FAQ -- http://www.tasam.com/~lrwiman/FAQ-mers ------------------------------ Date: Thu, 13 Dec 2001 13:49:04 -0500 From: Jeff Woods Subject: Mersenne: Entropia site down? http://www.entropia.com/ips doesn't work. I get this: The page cannot be found The page you are looking for might have been removed, had its name changed, or is temporarily unavailable. HTTP 404 - File not found Internet Information Services - ----------- However, the www.entropia.com main page works fine. It has been like this for several days now. Anyone else see this? _________________________________________________________________________ Unsubscribe & list info -- http://www.ndatech.com/mersenne/signup.htm Mersenne Prime FAQ -- http://www.tasam.com/~lrwiman/FAQ-mers ------------------------------ Date: Thu, 13 Dec 2001 12:07:53 -0800 (PST) From: Mary Conner Subject: RE: Mersenne: Moving an assignment On Wed, 12 Dec 2001, Aaron Blosser wrote: > I think we settled on the term "poached" long ago. > > There's really nothing stopping anyone from working on any exponent they > want. If they communicate with Primenet and it reports that an exponent > is assigned to someone else, the software will remove that exponent from > your worktodo list. This is true, although I would hope that the PrimeNet server does not "cooperate" in "poaching". I know all about the expiration period and whatnot, and that someone who was assigned an exponent previously might have it expire and then still complete it. I noted in an earlier message that I'd gotten an expired exponent for a doublecheck and wondered what would happen if we both started checking in work on this exponent. Well, before my computer could start work on it, the previous holder checked it in completed, and Prime95 removed it from my queue. So that part works. But I'm still back to trying to figure out how to move an exponent from one machine to another. Perhaps a little background is in order. I've been talking with one of the math teachers at my daughter's school about doing some work with distributed computing. I printed out some documentation from some projects and took it in to him, including sample pages from the PrimeNet status reports. He's considering working it in after Christmas break, but he said the example work units the students would work on would have to be short. I could set Prime95 up with a small dummy exponent, but I'd rather see them working on something real (and so they could see themselves on the PrimeNet reports), so I decided to try capturing a small doublecheck exponent that was expiring that they could polish off in a week. I planned to set my computer not to update while their computer was masquerading as mine for the week. I know trying to grab expiring exponents is hit or miss, but since they are recycled at 6am GMT (10pm here), I figured I'd have a good shot at getting a few, hopefully to get one that was truly abandoned. I was following several candidates on the assignment report when one of them changed accounts without having expired. It still had a few days to go until expiry, and when it changed accounts, the original date assigned remained the same, as well as the run time, so this was not an expired exponent being reassigned. So I figured that there must be a way to give an exponent from one machine to another (and even one account to another) and have PrimeNet know about it and not squawk. That would simplify things and allow the kids to run their own account with their school's name or something. This exponent is not an isolated example. I have a printout of the first page of the Assigned Exponents report from Dec 7th. I compared it with the Assigned Exponents report from a few minutes ago, and three exponents off that first page have moved from their old accounts (three different accounts) to this other person's account (same account for all three). Their assignment dates have remained the same, and the run times were not reset. I don't know if George manually moved these exponents, but I do know at least one was being regularly updated every few days, and another had over 3 weeks to go until expiring. If George didn't do it, then there's got to be a way for me to move an exponent, by hook or by crook. Mary Conner _________________________________________________________________________ Unsubscribe & list info -- http://www.ndatech.com/mersenne/signup.htm Mersenne Prime FAQ -- http://www.tasam.com/~lrwiman/FAQ-mers ------------------------------ Date: Thu, 13 Dec 2001 17:34:21 -0500 From: "C. Garrison" Subject: Re: Mersenne: Moving an assignment Mary, I haven't done this myself, but give it a try. 1. Exit your client. 2. Copy your prime95 directory to your recipient's hard drive in a directory called prime95_tmp. 3. Edit that temporary directory's worktodo.ini and take out the exponents you do not want to transfer (leave the ones you do want to transfer). 4. Exit their primary client. 4. Start their "temporary" client from that directory. 5. Have them go to the Test Menu/User Information screen and change your info to theirs. 6. Have them Stop and then Start the client - I would expect that the info change would then get transmitted. 7. Exit their "temporary" client. 8. Edit their primary worktodo.ini to add the worktodo entries of their "temporary" client. 9. Also copy any p and q files corresponding to those transferred exponents, if any. 10. Start their primary server. Hopefully the are now running okay. 11. Edit your primary worktodo.ini and remove those transferred exponent work items. 12. Start you primary client. 13. If everything is okay, remove your p and q files corresponding to those transferred exponents, if any. 14. Remove both temporary directories. You said by hook or by crook. Simple . Carleton Garrison wwwteamprimeribcom - ----- Original Message ----- From: "Mary Conner" To: Sent: Thursday, December 13, 2001 3:07 PM Subject: RE: Mersenne: Moving an assignment [...] > If George didn't do it, then > there's got to be a way for me to move an exponent, by hook or by crook. > > Mary Conner _________________________________________________________________________ Unsubscribe & list info -- http://www.ndatech.com/mersenne/signup.htm Mersenne Prime FAQ -- http://www.tasam.com/~lrwiman/FAQ-mers ------------------------------ Date: Fri, 14 Dec 2001 04:57:06 +0000 (GMT) From: Philip Whittington Subject: Mersenne: Speed of RAM Hello All, I'm sorry to trouble GIMPSers with what might be more of a computer question. I am using an Athon Thunderbird 1.3GHz and 256 MB of 133 MHz (DDR) Ram. My computer has always emitted a high pitched squeaking noise when processes like GIMPS start, and also when I do things like use the scroll wheel of my mouse etc. Recently the computer started hanging randomly on certain tasks (not GIMPS). After having taken the RAM down to 100MHz the high pitched squeak and the hanging has stopped. 1) Will the hanging have any impact on the credibility of what PRIME95 finds - I am around 86% through a L-L Test? 2) Can anyone shed light on whether it is my ram at fault or something on the motherboard? (sorry, I realise this is not really GIMPS related) Thank you, Philip Whittington _________________________________________________________________________ Unsubscribe & list info -- http://www.ndatech.com/mersenne/signup.htm Mersenne Prime FAQ -- http://www.tasam.com/~lrwiman/FAQ-mers ------------------------------ Date: Thu, 13 Dec 2001 21:45:45 -0800 From: "Aaron Blosser" Subject: RE: Mersenne: Speed of RAM Questions like this are fine. :) We talk about stuff like this from time to time. F'rinstance, I have some Compaq servers where the video goes totally haywire when Prime95 is running on them (only at 1024x768... works fine at 800x600). Turns out the built in graphics stuff is a bit too close to the processors, so there's a bit of interference going on. I would suspect the same behaviour in your case. Do you notice the sound coming from the built in speaker? Or is it somewhere else? Capacitors have a neat trick of making high pitched noises like you described, though the type found on motherboards are usually a bit too small for such shenanigans, but you never know. Slowing down the bus speed must have eliminated whatever resonance was going on. As for whether your results will be okay or not, there are a few built in checks in George's algorithm, so most hardware related things are generally caught (correct George?) and it is able to revert to the last time it saved the temp file in such cases. You can take a look in your results.txt file to see if there are any unusual errors in there. No real way to tell what your lockup problems are. If it only happened when you did certain things (mouse movements or something), then maybe you have some issues with the hardware related to the PS/2 interface... just a wild guess based on what you've said. Aaron > -----Original Message----- > From: mersenne-invalid-reply-address@base.com [mailto:mersenne-invalid- > reply-address@base.com] On Behalf Of Philip Whittington > Sent: Thursday, December 13, 2001 8:57 PM > To: mersenne@base.com > Subject: Mersenne: Speed of RAM > > Hello All, > > I'm sorry to trouble GIMPSers with what might be more of a computer > question. I > am using an Athon Thunderbird 1.3GHz and 256 MB of 133 MHz (DDR) Ram. > > My computer has always emitted a high pitched squeaking noise when > processes > like GIMPS start, and also when I do things like use the scroll wheel of > my > mouse etc. Recently the computer started hanging randomly on certain tasks > (not > GIMPS). After having taken the RAM down to 100MHz the high pitched squeak > and > the hanging has stopped. > > 1) Will the hanging have any impact on the credibility of what PRIME95 > finds - > I am around 86% through a L-L Test? > > 2) Can anyone shed light on whether it is my ram at fault or something on > the > motherboard? (sorry, I realise this is not really GIMPS related) > > Thank you, > > Philip Whittington _________________________________________________________________________ Unsubscribe & list info -- http://www.ndatech.com/mersenne/signup.htm Mersenne Prime FAQ -- http://www.tasam.com/~lrwiman/FAQ-mers ------------------------------ Date: Fri, 14 Dec 2001 07:13:13 -0000 From: "Daran" Subject: Re: Mersenne: CNET coverage of M39: Very good - ----- Original Message ----- From: "Jeff Woods" To: Sent: Thursday, December 13, 2001 6:44 PM Subject: Mersenne: CNET coverage of M39: Very good > ...His > system was part of a 210,000-machine quasi-supercomputer stretched across > the globe. [...] > The Mersenne prime search is moving in that direction. Each day, its > network of computers does work that would take a single 90MHz Pentium > computer 200 years to accomplish... 200 X 365 = 73000 p90 days/day So what are the other 137,000 faster than P90 machines doing? Regards Daran _________________________________________________________________________ Unsubscribe & list info -- http://www.ndatech.com/mersenne/signup.htm Mersenne Prime FAQ -- http://www.tasam.com/~lrwiman/FAQ-mers ------------------------------ Date: Fri, 14 Dec 2001 07:13:39 -0000 From: "Daran" Subject: Re: Mersenne: Teams and AOL - ----- Original Message ----- From: "Russel Brooks" To: Sent: Thursday, December 13, 2001 4:27 AM Subject: Re: Mersenne: Teams and AOL > My primary objective is to get their pcs running Prime95; stat > credit is a minor secondary concern. Or any other distributed computing project. An idling computer is like a running tap. > Cheers... Russ Daran _________________________________________________________________________ Unsubscribe & list info -- http://www.ndatech.com/mersenne/signup.htm Mersenne Prime FAQ -- http://www.tasam.com/~lrwiman/FAQ-mers ------------------------------ Date: Fri, 14 Dec 2001 03:36:50 -0500 From: Nathan Russell Subject: Re: Mersenne: CNET coverage of M39: Very good At 07:13 AM 12/14/2001 +0000, Daran wrote: >So what are the other 137,000 faster than P90 machines doing? > >Regards > >Daran That's a great point. Are there huge numbers of GIMPS members out there running machines not capable of running windows 98? Nathan _________________________________________________________________________ Unsubscribe & list info -- http://www.ndatech.com/mersenne/signup.htm Mersenne Prime FAQ -- http://www.tasam.com/~lrwiman/FAQ-mers ------------------------------ Date: Fri, 14 Dec 2001 18:40:01 -0000 From: bjb@bbhvig.uklinux.net Subject: Re: Mersenne: Speed of RAM On 14 Dec 2001, at 4:57, Philip Whittington wrote: > I'm sorry to trouble GIMPSers with what might be more of a computer question. I > am using an Athon Thunderbird 1.3GHz and 256 MB of 133 MHz (DDR) Ram. > > My computer has always emitted a high pitched squeaking noise when processes > like GIMPS start, and also when I do things like use the scroll wheel of my > mouse etc. Recently the computer started hanging randomly on certain tasks (not > GIMPS). After having taken the RAM down to 100MHz the high pitched squeak and > the hanging has stopped. The most likely reasons for the "squeak" are: (a) some sort of mechanical resonance, perhaps involving the heatsink/CPU coupling, driven by magnetic linking from the "undertones" of CPU activity. Unless the noise is LOUD this is most probably harmless. (b) possibly the power supply is having difficulty providing sufficient current. Switched mode power supplies usually operate "supersonically" (around 50 KHz) but the frequency can be dragged down into the audible range by excess current draw. This would cause a squeal. Now a Tbird system does draw more current when the FPU is very active i.e. when Prime95 is running, but this should persist whilst Prime95 is running, rather than being transient as the program starts, as your message suggests. (?) Excess current draw would tend to cause system instability, getting worse as the smoothing capacitors age rapidly due to running under heavy strain. If your PSU is less than 250W or not marked as "Athlon approved" (see the AMD website) then I'd reccomend changing it for a 300W+ AMD/P4 approved PSU. Changing the system bus from 133 MHz to 100 MHz will underclock the system substantially (1.33 GHz will run at only 1 GHz) which will have the effect of reducing the current draw. Even changing the memory bus from 133 MHz to 100 MHz leaving the system bus at 133 MHz will reduce the load since the FPU will be inactive some of the time - the memory bandwidth is in any case inadequate to keep the FPU fully occupied, cutting the memory bandwidth by 25% obviously doesn't improve matters! > > 1) Will the hanging have any impact on the credibility of what PRIME95 finds - > I am around 86% through a L-L Test? Might be OK. Try running the full selftest, or the torture test. If these show no problems (despite the system hanging occasionally) you should be allright. In any case there's not much point in aborting 86% of the way through. Suggest you switch to DC assignments (after finishing this one) until the problem is sorted, though. > > 2) Can anyone shed light on whether it is my ram at fault or something on the > motherboard? (sorry, I realise this is not really GIMPS related) More likely M/B rather than RAM. Some Abit KT7s have weak smoothing capacitors, though others seem to be fine. This may depend on having an adequate PSU and/or adequate CPU cooling - the caps which blow are those next to the CPU, which therefore run hot if the CPU is inadequately ventilated. You might want to check out the memory using Memt25 but I doubt you will find a problem. If it is a PSU / smoothing cap problem, expect it to get worse, slowly at first but eventually getting so bad that the system won't boot. Repair is impractical; replacing the M/B will fix the problem temporarily, but it will recur, sooner rather than later, unless the PSU is upgraded. Regards Brian Beesley _________________________________________________________________________ Unsubscribe & list info -- http://www.ndatech.com/mersenne/signup.htm Mersenne Prime FAQ -- http://www.tasam.com/~lrwiman/FAQ-mers ------------------------------ End of Mersenne Digest V1 #919 ******************************