Mersenne Digest Wednesday, February 2 2000 Volume 01 : Number 686 ---------------------------------------------------------------------- Date: Mon, 31 Jan 2000 12:20:59 +0100 From: Sylvain PEREZ <[EMAIL PROTECTED]> Subject: Mersenne: Sorry, but ... ... I don't receive any message from list for a few days ... so that's just a "test". I owe you the equivalent bandwidth of this message. Sylvain Perez _________________________________________________________________ Unsubscribe & list info -- http://www.scruz.net/~luke/signup.htm Mersenne Prime FAQ -- http://www.tasam.com/~lrwiman/FAQ-mers ------------------------------ Date: Mon, 31 Jan 2000 14:21:50 +0000 From: Alexander Kruppa <[EMAIL PROTECTED]> Subject: Mersenne: Chance of P-1 factor Hi all, I've been trying to figure out what the chance of finding a P-1 factor is, given a certain amount of trial factoring and a P-1 bound. I started with: The chance of Mp having a factor in [2^n, 2^(n+1)] is 1/n. The chance of a number N being B-smooth is (log(B)/log(N))^(log(N)/log(B)) . Since we know factors of Mp are f = 2kp+1, the difficulty of the problem of finding f is reduced to factoring k. So the chance that a factor of Mp can be found by P-1 with bound B is (log(B)/log(f/2p)) ^ (log(f/2p)/log(B)) . I get the probabilities of finding any factor by running over factor sizes in bits (powers of two) and add up the product of the prob of Mp having a factor of n bits and the prob of P-1 finding such a factor. (In fact, I make two lists and get the dot product) Mathematica: Pfactor[L_] := Table[If[n>L,1/n,0],{n,200}]; // The probabilities of Mp having a factor of n bits, after trial factoring up to 2^L Psmooth[B_,p_] := Table[ If [(n-Log[2,2p])<Log[2,B], 1, (Log[2,B]/(n-Log[2,2p]))^((n-Log[2,2p])/Log[2,B])], {n,200}]; //The chance of P-1 finding a factor of size n of Mp with stage 1 bound B N[Psmooth[1000000,33000000].Pfactor[68]] 0.026342 This would mean that we'd find factors of only 2.6% of the 10 million exponents, even with a bound of 1M ! Thats pretty disappointing. Now, is the above about correct ?? I'm pretty unsure about my math, so.. at least N[Psmooth[4000000,300000].Pfactor[53]] ~= 10% which is close to my empirical data, if a little low ( I get about 11-12% success rate with B1=100k, B2=4M). How much error does my quantizing factor sizes to powers of 2 introduce? And what is the probability of a number being B1-smooth with a single factor f B1<f<B2 (for P-1 with stage 2) ? And it just occurred to me: if we know that Mp is composite (because we did an LL test), after trial factoring the chance for larger factors must increase, because the interval in which the factor must lie is narrowed down. How can I take that into account? Any help would be much appreciated! Ciao, Alex. _________________________________________________________________ Unsubscribe & list info -- http://www.scruz.net/~luke/signup.htm Mersenne Prime FAQ -- http://www.tasam.com/~lrwiman/FAQ-mers ------------------------------ Date: Mon, 31 Jan 2000 10:42:33 -0500 From: George Woltman <[EMAIL PROTECTED]> Subject: Re: Mersenne: Chance of P-1 factor Hi, At 02:21 PM 1/31/00 +0000, Alexander Kruppa wrote: >I started with: > >The chance of a number N being B-smooth is >(log(B)/log(N))^(log(N)/log(B)) . I don't think this is correct. Look up Dickman's function in Knuth vol 2 pages 382 and 383. You can also look at a July 10, 1996 post to this mailing list from Peter-Lawrence Montgomery. I've been grappling with this same problem in prime95 version 20. I'll let you independently implement it and see if out results agree! Regards, George _________________________________________________________________ Unsubscribe & list info -- http://www.scruz.net/~luke/signup.htm Mersenne Prime FAQ -- http://www.tasam.com/~lrwiman/FAQ-mers ------------------------------ Date: Mon, 31 Jan 2000 11:38:09 -0500 From: George Woltman <[EMAIL PROTECTED]> Subject: Mersenne: Version 20 memory questions Hi all, I'm working on prime95 version 20. The important new feature is a P-1 factoring step prior to a Lucas-Lehmer test. The P-1 factoring step has a 3-5% chance of finding a factor at a cost of 2-4% of an LL test. The net effect is we speed up GIMPS' throughput by a percent or two (more if you factor in the fact that a double-check is avoided too). Stage 2 of P-1 factoring requires more memory than an LL test. The more memory you give the program the faster stage 2 executes. For example, an exponent around 10 million requires about 4MB of data. P-1 stage 2 would like to have 48MB, but can get by with as little as 24MB. The good news is that the LL test takes a month, P-1 factoring takes a day, and the memory hungry stage 2 takes only half a day. GIMPS has always had a good reputation for not interfering with your normal work. To preserve GIMPS' reputation, I'm thinking of implementing the following. In the Options/CPU dialog, prime95 will let you select the maximum amount of memory the program can use and the hours of the day it can use it. The default would be 80% of RAM (divided by the number of CPUs) during nighttime hours only. Finally, the questions: Would we be better off disabling P-1 factoring unless the user explicitly activates it (knowing that most users won't read enough to turn it on)? Are there better solutions? It would be nice if prime95 could detect that memory thrashing was happening and pause itself until more memory was available. Can Windows programs do this? Are the defaults too aggressive (especially the 80% of RAM)? You can send your comments and suggestions to me privately or to the entire mailing list. Having fun, George _________________________________________________________________ Unsubscribe & list info -- http://www.scruz.net/~luke/signup.htm Mersenne Prime FAQ -- http://www.tasam.com/~lrwiman/FAQ-mers ------------------------------ Date: Mon, 31 Jan 2000 12:30:50 -0500 From: Jud McCranie <[EMAIL PROTECTED]> Subject: Re: Mersenne: Version 20 memory questions At 11:38 AM 1/31/00 -0500, George Woltman wrote: > Would we be better off disabling P-1 factoring unless the user explicitly >activates it (knowing that most users won't read enough to turn it on)? That sounds like a good idea to me. Giving up 24-48MB would interfere with some people's work, and you don't want to do that. Prime95 needs to be transparent. It probably should be in the advanced menu, and allow you to set the hours and days. +--------------------------------------------------------+ | Jud McCranie | | | | 137*2^197783+1 is prime! (59,541 digits, 11/11/99) | | 137*2^224879+1 is prime! (67,687 digits, 1/00) | +--------------------------------------------------------+ _________________________________________________________________ Unsubscribe & list info -- http://www.scruz.net/~luke/signup.htm Mersenne Prime FAQ -- http://www.tasam.com/~lrwiman/FAQ-mers ------------------------------ Date: Mon, 31 Jan 2000 13:44:51 -0500 From: "Louis Towles" <[EMAIL PROTECTED]> Subject: Mersenne: Version 20 memory questions This is one vote for memory conumsion by default stays at 4ish MB, and only if you change it on the advanced menu will it use more. I know too many people who never read the docs and will have no idea why their computer "got slow" when they put on ver. 20 and will just want it taken off. Louis Towles _________________________________________________________________ Unsubscribe & list info -- http://www.scruz.net/~luke/signup.htm Mersenne Prime FAQ -- http://www.tasam.com/~lrwiman/FAQ-mers ------------------------------ Date: Mon, 31 Jan 2000 14:14:27 -0500 From: Alan Powell <[EMAIL PROTECTED]> Subject: Re: Mersenne: Version 20 memory questions George Here is another vote for P-1 to default to off unless it is explicitly enabled using the Advanced Menu like ECM. Regards Alan Powell [EMAIL PROTECTED] _________________________________________________________________ Unsubscribe & list info -- http://www.scruz.net/~luke/signup.htm Mersenne Prime FAQ -- http://www.tasam.com/~lrwiman/FAQ-mers ------------------------------ Date: Mon, 31 Jan 2000 13:53:34 -0500 From: Pierre Abbat <[EMAIL PROTECTED]> Subject: Re: Mersenne: Version 20 memory questions >Are there better solutions? It would be nice if prime95 could detect that >memory thrashing was happening and pause itself until more memory >was available. Can Windows programs do this? Watch the time it takes to do an iteration; if it jumps up, go to sleep for a few minutes. This won't distinguish being swapped out from being pushed to the background, though. On Linux you can look in /proc/self/. I once ran a program which took the more memory the longer it ran. By the end of the night, it was taking only a few percent of the CPU, even though it was not running nicely, and sieve was taking the rest. >Are the defaults too aggressive (especially the 80% of RAM)? I would set the RAM to 62% on smaller machines, 70% on larger. I have GIMP (not to be confused with GIMPS) set up for 10 meg on a 16 meg box, and 90 meg on 128. phma _________________________________________________________________ Unsubscribe & list info -- http://www.scruz.net/~luke/signup.htm Mersenne Prime FAQ -- http://www.tasam.com/~lrwiman/FAQ-mers ------------------------------ Date: Mon, 31 Jan 2000 12:26:48 -0700 From: "Alan Vidmar" <[EMAIL PROTECTED]> Subject: Re: Mersenne: Version 20 memory questions I have to agree with Louis on this issue. Alan On 31 Jan 00, at 13:44, Louis Towles wrote: From: "Louis Towles" <[EMAIL PROTECTED]> To: "Prime Mailing List" <[EMAIL PROTECTED]> Subject: Mersenne: Version 20 memory questions Date sent: Mon, 31 Jan 2000 13:44:51 -0500 > This is one vote for memory conumsion by default stays at 4ish MB, and only > if you change it on the advanced menu will it use more. > > I know too many people who never read the docs and will have no idea why > their computer "got slow" when they put on ver. 20 and will just want it > taken off. > > > Louis Towles > > > _________________________________________________________________ > Unsubscribe & list info -- http://www.scruz.net/~luke/signup.htm > Mersenne Prime FAQ -- http://www.tasam.com/~lrwiman/FAQ-mers "A programmer is a person who turns coffee into software." Alan R. Vidmar Assistant Director of IT Office of Financial Aid University of Colorado [EMAIL PROTECTED] (303)492-3598 *** This message printed with 100% recycled electrons *** _________________________________________________________________ Unsubscribe & list info -- http://www.scruz.net/~luke/signup.htm Mersenne Prime FAQ -- http://www.tasam.com/~lrwiman/FAQ-mers ------------------------------ Date: Mon, 31 Jan 2000 19:44:28 -0000 From: "Andy Steward" <[EMAIL PROTECTED]> Subject: Re: Mersenne: Chance of P-1 factor - ----- Original Message ----- From: Alexander Kruppa <[EMAIL PROTECTED]> > I've been trying to figure out what the chance of finding a P-1 factor > is, given a certain amount of trial factoring and a P-1 bound. [...] > This would mean that we'd find factors of only 2.6% of the 10 million > exponents, even with a bound of 1M ! Thats pretty disappointing. To throw some experimental data in with the theory: I use p-1 with B1= 1e7 and B2 = 1e8 to try and factorize Generalized Repunits. The composites that I deal with range from say 75 to 150 digits so are much smaller than those Mersenne composites you are working on while having a similar restriction on the form of factors. The results: 8 successes in 1452 trials or 0.55% Andy Steward _________________________________________________________________ Unsubscribe & list info -- http://www.scruz.net/~luke/signup.htm Mersenne Prime FAQ -- http://www.tasam.com/~lrwiman/FAQ-mers ------------------------------ Date: Mon, 31 Jan 2000 20:08:22 +0000 (GMT) From: Chris Jefferson <[EMAIL PROTECTED]> Subject: Re: Mersenne: Version 20 memory questions > > GIMPS has always had a good reputation for not interfering with > your normal work. To preserve GIMPS' reputation, I'm thinking of implementing > the following. In the Options/CPU dialog, prime95 will let you select the > maximum > amount of memory the program can use and the hours of the day it can use it. > The default would be 80% of RAM (divided by the number of CPUs) during > nighttime hours only. > > Finally, the questions: > > Would we be better off disabling P-1 factoring unless the user explicitly > activates it (knowing that most users won't read enough to turn it on)? Hmm.. I think it should be turned on, if a quick check shows the user has enough Physical memory (I'm sure windows can do this) how about only if you have >64MB? > > Are there better solutions? It would be nice if prime95 could detect that > memory thrashing was happening and pause itself until more memory > was available. Can Windows programs do this? If this is possible, great! I'm not sure it is tho... > > Are the defaults too aggressive (especially the 80% of RAM)? > Hmm... I think it is. I currently run Prime95 and War-FTPd in 32MB of ram and they are quite happy living together, along with some no too demanding apps. However, if it started to noticably intferere with my work, off it would go! I know this kind of thing has been bought up before, but how about activating as a screen saver / when the screen saver comes on. Can this be detected? > You can send your comments and suggestions to me privately > or to the entire mailing list. > > Having fun, > George > > > _________________________________________________________________ > Unsubscribe & list info -- http://www.scruz.net/~luke/signup.htm > Mersenne Prime FAQ -- http://www.tasam.com/~lrwiman/FAQ-mers > _________________________________________________________________ Unsubscribe & list info -- http://www.scruz.net/~luke/signup.htm Mersenne Prime FAQ -- http://www.tasam.com/~lrwiman/FAQ-mers ------------------------------ Date: Mon, 31 Jan 2000 17:32:18 -0500 From: Pierre Abbat <[EMAIL PROTECTED]> Subject: Re: Mersenne: Version 20 memory questions >I know this kind of thing has been bought up before, but how about >activating as a screen saver / when the screen saver comes on. Can this be >detected? This is not a good idea, because the screen saver could be running on the local display while someone is logged in from another computer or terminal and busy using the computer. phma _________________________________________________________________ Unsubscribe & list info -- http://www.scruz.net/~luke/signup.htm Mersenne Prime FAQ -- http://www.tasam.com/~lrwiman/FAQ-mers ------------------------------ Date: Mon, 31 Jan 2000 16:35:39 -0800 From: Luke Welsh <[EMAIL PROTECTED]> Subject: Re: Mersenne: Version 20 memory questions At 01:53 PM 1/31/00 -0500, Pierre Abbat wrote: >>Are there better solutions? It would be nice if prime95 could detect that >>memory thrashing was happening and pause itself until more memory >>was available. Can Windows programs do this? > >Watch the time it takes to do an iteration; if it jumps up, go to sleep >for a few minutes. Run P-1 and LL in parallel? Execute the P-1 thread when there are enough "resources", otherwise LL. If, after 1 week, P-1 has not run to completion, give up on it. Assuming 24x7, one week's worth of overnights and weekends should be enough? Maybe default to execute P-1 during "off hours" unless overridden via a menu option? - --Luke _________________________________________________________________ Unsubscribe & list info -- http://www.scruz.net/~luke/signup.htm Mersenne Prime FAQ -- http://www.tasam.com/~lrwiman/FAQ-mers ------------------------------ Date: Mon, 31 Jan 2000 20:58:06 -0500 From: "David Campeau" <[EMAIL PROTECTED]> Subject: Re: Mersenne: Version 20 memory questions > GIMPS has always had a good reputation for not interfering with > your normal work. To preserve GIMPS' reputation, I'm thinking of implementing > the following. In the Options/CPU dialog, prime95 will let you select the > maximum > amount of memory the program can use and the hours of the day it can use it. > The default would be 80% of RAM (divided by the number of CPUs) during > nighttime hours only. > > Finally, the questions: > > Would we be better off disabling P-1 factoring unless the user explicitly > activates it (knowing that most users won't read enough to turn it on)? > Should we? I don't think so... or at least P-1 factoring should be a new type of work on PimeNet. > Are there better solutions? It would be nice if prime95 could detect that > memory thrashing was happening and pause itself until more memory > was available. Can Windows programs do this? > Yes it can, but I don't think this is the best way to go. The system will from time to time put infrequently used page file on the hard drive when he approach is total phisical memory limit. But if you leave the system with a healty amount (lets say 20-25%of total) of phisical memory to work with, hard page fault (swapping) are going to be nearly non-existant. Start a timer (each second?), and call GlobalMemoryStatus This will give you the current memory state of the machine (Total phisical, Available phisical, Total PageFile, Available PageFile, Total Virtual, Available Virtual) >From there you adjust your memory usage (lower it) according to the available memory. When you write a file to disk, it should be a good place to reclaim memory (if available and needed). > Are the defaults too aggressive (especially the 80% of RAM)? > If you always leave 20% of ram to the system to work with, there should not be a problem. This is a nice problem, it's going to be interesting to see what you come up with. Regards, David Campeau, _________________________________________________________________ Unsubscribe & list info -- http://www.scruz.net/~luke/signup.htm Mersenne Prime FAQ -- http://www.tasam.com/~lrwiman/FAQ-mers ------------------------------ Date: Mon, 31 Jan 2000 18:30:25 -0800 From: "John R Pierce" <[EMAIL PROTECTED]> Subject: Re: Mersenne: Version 20 memory questions > GIMPS has always had a good reputation for not interfering with > your normal work. To preserve GIMPS' reputation, I'm thinking of implementing > the following. In the Options/CPU dialog, prime95 will let you select the > maximum > amount of memory the program can use and the hours of the day it can use it. > The default would be 80% of RAM (divided by the number of CPUs) during > nighttime hours only. Um, 80%? I have 384MB in this box. 80% of 384MB is like 300MB! I hope that would be limited to the min/max amount of memory you need to do useful work? Since I have 384MB, I would have no problem dedicating 48MB to this 24/7. > Finally, the questions: > > Would we be better off disabling P-1 factoring unless the user explicitly > activates it (knowing that most users won't read enough to turn it on)? > > Are there better solutions? It would be nice if prime95 could detect that > memory thrashing was happening and pause itself until more memory > was available. Can Windows programs do this? Yes, you can get various statistics out of the system re: paging etc, but I believe the exact method is different on win9X and NT/w2k. I think someone elses suggestion might be better, detect when the system slows down and back off for 5 minutes or something. I would think 20% or something would be a better default. if someone has 128MB, thats 24MB. If they have 256MB, thats 48MB. if they have less than 128MB, they really can't afford to run this at all unless its on a dedicated basis. - -jrp _________________________________________________________________ Unsubscribe & list info -- http://www.scruz.net/~luke/signup.htm Mersenne Prime FAQ -- http://www.tasam.com/~lrwiman/FAQ-mers ------------------------------ Date: Tue, 01 Feb 2000 08:26:13 +0100 From: Harald Tveit Alvestrand <[EMAIL PROTECTED]> Subject: Re: Mersenne: Version 20 memory questions At 20:58 31.01.00 -0500, David Campeau wrote: >If you always leave 20% of ram to the system to work with, there should not >be a problem. I wonder how you're going to find out that there's anything left. *ALL* of my NT systems routinely run with a commit charge around double physical memory (this 96 MB laptop currently has 157M commit charge). *I* know that this is due to an enormous amount of rarely used daemons. But how is Prime95 to know that? Harald A - -- Harald Tveit Alvestrand, EDB Maxware, Norway [EMAIL PROTECTED] _________________________________________________________________ Unsubscribe & list info -- http://www.scruz.net/~luke/signup.htm Mersenne Prime FAQ -- http://www.tasam.com/~lrwiman/FAQ-mers ------------------------------ Date: Tue, 1 Feb 2000 13:43:42 +0100 (MET) From: Wojciech Florek <[EMAIL PROTECTED]> Subject: Mersenne: Re: Is it new? (fwd) - ---------- Forwarded message ---------- Date: Mon, 31 Jan 2000 14:59:38 -0500 From: George Woltman <[EMAIL PROTECTED]> To: Wojciech Florek <[EMAIL PROTECTED]> Subject: Re: Is it new? Hi, At 01:39 PM 1/30/00 +0100, you wrote: >Playing with factors of mersennes I've found the following >relation: > >Let p=4q+b, b=1,3, and a=4-b (so a=3,1, respectively) then >p divides M(p-3)-aq. >E.g. >p=41, q=10 b=1 a=3 >M(38)-3*10=274877906913=6704339193*41 > >p=43, q=10, b=3 a=1 >M(40)-10=1099511627775=25570037855*43 > >Is it something new or well known? I don't know. I'm much more a computer programmer than a mathematician. You might ask on the Mersenne mailing list. Perhaps someone there can answer your question, prove your observation, and tell us if it is of any value. Best regards, George Wojciech Florek (WsF) [EMAIL PROTECTED] _________________________________________________________________ Unsubscribe & list info -- http://www.scruz.net/~luke/signup.htm Mersenne Prime FAQ -- http://www.tasam.com/~lrwiman/FAQ-mers ------------------------------ Date: Tue, 1 Feb 2000 19:53:37 -0000 From: "Daniel Grace" <[EMAIL PROTECTED]> Subject: Re: Mersenne: Re: Is it new? (fwd) - ----- Original Message ----- From: Wojciech Florek <[EMAIL PROTECTED]> To: <[EMAIL PROTECTED]> Sent: Tuesday, February 01, 2000 12:43 PM Subject: Mersenne: Re: Is it new? (fwd) > >Playing with factors of mersennes I've found the following > >relation: > > > >Let p=4q+b, b=1,3, and a=4-b (so a=3,1, respectively) then > >p divides M(p-3)-aq. > >E.g. > >p=41, q=10 b=1 a=3 > >M(38)-3*10=274877906913=6704339193*41 > > > >p=43, q=10, b=3 a=1 > >M(40)-10=1099511627775=25570037855*43 > > > >Is it something new or well known? > > I don't know. I'm much more a computer programmer than a mathematician. > You might ask on the Mersenne mailing list. Perhaps someone there can > answer your question, prove your observation, and tell us if it is of any > value. > > Best regards, > George > If p=1 mod 4 then let x=2^(p-3)-1-3(p-1)/4 so 4x=2^(p-1)-4-3(p-1)=1-4+3=0 mod p. If p=3 mod 4 then let x=2^(p-3)-1-(p-3)/4 so 4x=2^(p-1)-4-(p-3)=1-4+3=0 mod p. Yea, it works for any odd prime - you could prove it more consisely. I cannot see how this helps to factor Mersennes though. Rgds, - ---------------------------------------------------------- Daniel W. Grace e-mail: [EMAIL PROTECTED] _________________________________________________________________ Unsubscribe & list info -- http://www.scruz.net/~luke/signup.htm Mersenne Prime FAQ -- http://www.tasam.com/~lrwiman/FAQ-mers ------------------------------ Date: Tue, 01 Feb 2000 13:35:36 -0600 From: "David J. Zook" <[EMAIL PROTECTED]> Subject: Re: Mersenne: Version 20 memory questions For a machine with 64 MB and up, defaulting to ON during night hours would seem OK to me (I have 144 MB). However, I assume that it will be possible to specify (change) what hours are the ON hours. I suppose that midnight to 06:00 would be fine for most people, as a default. It should be possible for me to PAUSE the P-1 test when 48 MB is too much for the moment. If PRIME95 could do some other work while P-1 test was paused we would be more likely to approve of the 48 MB footprint. Less than 64 MB should default to OFF at all times. I am afraid of thrashing on the small machine. If there is any point to doing Stage 1 processing (24 MB maximum) without doing Stage 2 processing, this could be implemented on machines with 32 MB - 64 MB, during default (night) hours. I have been looking for a method to determine how much RAM is "really" available (without using swap file), but have never seen it mentioned in any way. This includes DDJ and MSJ. There are many apps which would benefit from using maximum RAM without going to swap file. Perhaps PRIMENET should gather machine descriptors (with permission, of course): processor descriptor from CPUID, L1 and L2 cache sizes, speeds, and RAM size. I do not have the Intel/Athlon manuals at hand, but my memory says that much detailed data is available from the processor, or from tables keyed to processor step value. - -- The Internet is a library the size of the world. _________________________________________________________________ Unsubscribe & list info -- http://www.scruz.net/~luke/signup.htm Mersenne Prime FAQ -- http://www.tasam.com/~lrwiman/FAQ-mers ------------------------------ Date: Tue, 1 Feb 2000 22:59:57 +0000 From: "Steinar H. Gunderson" <[EMAIL PROTECTED]> Subject: Mersenne: Re: Version 20 memory questions On Tue, Feb 01, 2000 at 01:35:36PM -0600, David J. Zook wrote: >I have been looking for a method to determine how much RAM is "really" >available >(without using swap file), but have never seen it mentioned in any way. A quick and dirty test would be using ReCache (see previous discussion on mersenne, or mail me for the URL) and just checking when it starts to swap. However, that would be rather hard for a program, but I guess you could try to see just when the swap usage starts to go up, up, up... Remember -- swapping isn't the only problem either. That spare memory is used for disk cache. /* Steinar */ - -- Homepage: http://members.xoom.com/sneeze/ _________________________________________________________________ Unsubscribe & list info -- http://www.scruz.net/~luke/signup.htm Mersenne Prime FAQ -- http://www.tasam.com/~lrwiman/FAQ-mers ------------------------------ Date: Tue, 1 Feb 2000 23:29:47 -0000 From: "Brian J. Beesley" <[EMAIL PROTECTED]> Subject: Re: Mersenne: Version 20 memory questions On 31 Jan 00, at 11:38, George Woltman wrote: > GIMPS has always had a good reputation for not interfering with > your normal work. To preserve GIMPS' reputation, I'm thinking of > implementing the following. In the Options/CPU dialog, prime95 will let > you select the maximum amount of memory the program can use and the hours > of the day it can use it. The default would be 80% of RAM (divided by the > number of CPUs) during nighttime hours only. Hmm. Small memory systems tend to have a smaller percentage of system RAM available to applications than systems with larger memory. The point is that the OS kernel & "essential" DLLs/loadable modules are fixed in size (for a particular hardware & OS setup). > > Finally, the questions: > > Would we be better off disabling P-1 factoring unless the user explicitly > activates it (knowing that most users won't read enough to turn it on)? Sounds sensible - especially as an interim measure. When we get more experience we may be able to change this decision in favour of enabling P-1 with a small memory work space as the default. Probably we should tell users what is going on the first time they run a V20 program (i.e. create local.ini using V20 or run V20 for the first time on an existing local.ini file). Whatever the default is. We should also be looking at bringing P-1 factoring into the PrimeNet assignment allocation/results reporting system, if we want to encourage people to participate on a reasonable scale. > > Are there better solutions? It would be nice if prime95 could detect that > memory thrashing was happening and pause itself until more memory was > available. Can Windows programs do this? Yes - there must be APIs since there are existing Windows applications which measure such things. Including tuning aids supplied in the Microsoft Resource Kits for Win 95 and NT. Look at the page fault I/O rate - anything above 1 per second sustained for any period of time indicates a problem with memory quantity. Note, it is _usual_ for systems to operate with more virtual memory demand summed over active applications than physical memory available, since many applications contain static work arrays and/or major blocks of code (especially called from language support libraries) which are either shared or not used by whatever tasks the applications are doing at the time. Measuring the page fault rate (or, crudely, the I/O rate on whichever device(s) have swap files mounted) is also as good a way as any of identifying memory shortages on a linux system. Is there any realistic way of implementing the required memory space as "virtual memory" using a random access file instead of a plain memory workspace, or would that cause excessive overheads? The reason I ask is that, if it were done that way, the system would (more or less) tune itself according to memory availability on all major OS. Windows 9x, NT and linux all use "slack" memory to buffer disk I/O. > Are the defaults too aggressive (especially the 80% of RAM)? If the system is going to be "self tuning" in any way, it should endeavour to use _all_ the available RAM (not just 80%) providing that it gets itself out of the way when a foreground application demands memory. Preferably Prime95 would contract its workspace - at the expense of its own efficiency - rather than just suspending itself. Or automatically switch to something else with a smaller memory footprint (ECM on small exponents, or trial factoring - or starting Phase 1 on another P-1 job, if the problem is caused by being in phase 2) if there really isn't sufficient memory available to continue running P-1 (phase 2) profitably for the time being - automatically switching back when the crisis passes (this would mean interrupting the other task for a "system status check" every few minutes, though the overhead should be manageable). Regards Brian Beesley _________________________________________________________________ Unsubscribe & list info -- http://www.scruz.net/~luke/signup.htm Mersenne Prime FAQ -- http://www.tasam.com/~lrwiman/FAQ-mers ------------------------------ Date: Tue, 1 Feb 2000 19:06:33 -0500 From: Pierre Abbat <[EMAIL PROTECTED]> Subject: Re: Mersenne: Version 20 memory questions >However, I assume that it will be possible to specify (change) what >hours are the >ON hours. I suppose that midnight to 06:00 would be fine for most >people, as a >default. This probably depends on your distro. Red Hat and Mandrake run their crontabs between 4 and 5 in the morning, so that hour may be bad; but those processes, which consist of building a database of all the files on the disk and all the man pages, are I/O bound, not compute bound. phma _________________________________________________________________ Unsubscribe & list info -- http://www.scruz.net/~luke/signup.htm Mersenne Prime FAQ -- http://www.tasam.com/~lrwiman/FAQ-mers ------------------------------ Date: Wed, 2 Feb 2000 15:56:01 -0500 From: "Vincent J. Mooney Jr." <[EMAIL PROTECTED]> Subject: Mersenne: Icon I am using WIN 98. How do I set up an icon on the desktop to kick off PRIME95 (as I needed to do twice today when the dang computer crashed)? _________________________________________________________________ Unsubscribe & list info -- http://www.scruz.net/~luke/signup.htm Mersenne Prime FAQ -- http://www.tasam.com/~lrwiman/FAQ-mers ------------------------------ End of Mersenne Digest V1 #686 ******************************
