RE: Mersenne: It is with deep regrets that I must consider this option...
From: Tim Esau I am compelled to redirect the majority of my computing team to another target--SETI. When this was announced, I looked into it and had to decide whether to jump ship as well. The single largest reason that I'm staying 100% with GIMPS/PrimeNet is the efficiency and thought that went into the project. The biggest single factor against SETI (in my book) is that they decided to go the screensaver route. That gives up *a lot* of computing cycles for no good reason, IMO. But that's just my opinion. Rick. - [EMAIL PROTECTED] http://www.alienshore.com/ Unsubscribe list info -- http://www.scruz.net/~luke/signup.htm
RE: Mersenne: I am curious
From: Jud McCranie Why not form one big team out of all of GIMPS and split the money, Honey? While I can't speak for others, I'm not in GIMPS for the money. While the awards are reasonably new, I've got over ten years accumulated on my two home machines and I *like* to be able to look at that total and see what I've done. Sure I'll take the money if it's thrown my way, but I'm not going to change the way I'm doing this just to have a better chance to get it. That's my MO, anyway. Rick. - [EMAIL PROTECTED] http://www.alienshore.com/ Unsubscribe list info -- http://www.scruz.net/~luke/signup.htm
Re: Mersenne: LLL
Hi, At 12:15 AM 5/14/99 +0400, Alexey Khlyamkov wrote: I'am so sorry for this situation. No need to apologize Alexey. This is really a bug in mprime. My disk quota on the server was overfulled in that time and mprime running on several machines which connected to main server via NFS zeroed out some setting files such as local.ini and worktodo.ini. Prior to contacting the server, mprime calls the routine IniFileWritable, which should have returned FALSE and thereby avoiding contact with the server to get more work. I do not know why this routine failed - maybe it has something to do with NFS. I can't test an NFS setup, but I'll try debugging using a local disk. Best regards, George Unsubscribe list info -- http://www.scruz.net/~luke/signup.htm
Re: Mersenne: It is with deep regrets that I must consider this option...
On Fri, May 14, 1999 at 03:43:42PM -0400, Rick Pali [EMAIL PROTECTED] wrote: When this was announced, I looked into it and had to decide whether to jump ship as well. Ditto. The single largest reason that I'm staying 100% with GIMPS/PrimeNet is the efficiency and thought that went into the project. The biggest single factor against SETI (in my book) is that they decided to go the screensaver route. That gives up *a lot* of computing cycles for no good reason, IMO. I've decided to leave all my x86 machines (Win9x, Linux and FreeBSD) crunching away at GIMPS/PrimeNet, and run the SETI@Home client on the Sparc machines I administer. The *NIX SETI clients don't come with all that GUI crap, GIMPS/PrimeNet runs faster on x86 boxes, and I don't have time or energy to muck about with non-PrimeNet GIMPS clients. Bryan -- Bryan Fullertonhttp://www.samurai.com/ Core Competency Samurai Consulting "No, we don't do seppuku." Can you feel the Ohmu call? Unsubscribe list info -- http://www.scruz.net/~luke/signup.htm
Re: Mersenne: SETI on ABC News last night
suppose will sound more appealing to the average person--"the search for enormous prime numbers" or "the search for alien life forms"--especially now that SETI has such high-profile exposure? GIMPS has a big disadvantage in This might be true for a few months but I susupect that many will lose interest as (at last report) the SETI software will not alert you if it finds a "match." They won't tell you if your dataset has an interesting target because they don't want to cause a possible panic due to a false alarm. Would anyone care to comment on the appeal of SETI? Personally speaking, it doesn't interest me at all. I don't consider its goals to be terribly useful or important, and I don't think that it has a reasonable chance of accomplishing anything. But as number theory enthusiast I find something intrinsically interesting and worthwhile about finding factors and searching for Mersenne primes. I am probably in a minority of the general population in this regard. What's the appeal of SETI? Many people (at least in the US) seem to believe in ETs (just look at the popularity of the X-files, MIB, Independense Day, Star Trek, etc.) As for useful goals that seems to be a matter of opinion no matter what you decide to study. As a number theory enthusiast I would think that you would find it unusual if they didn't find something. Not to say that there are ETs but if you consider the number of stars we know about (and say similar to our sun) then I would expect at some point for life to form on that planet and accidently/intentionally broadcast radio signals. Yes I'm playing fast and loose here with a few things but this are just the results from a quick poll of co-workers. -Ray Unsubscribe list info -- http://www.scruz.net/~luke/signup.htm
Re: Mersenne: Screen saver killers?
Why not just write a piece of code that (during installation of Prime95) removes the screensaver start-up line in the ini (windows) files. Then, restart the PC with these new settings and the screensaver start-up lines will be invisiable to the bootup. This would automatically turn off screensavers. Once the Prime95 program is uninstalled (forbid that to happen), the screen saver can be added back into the ini file and the computer can load the screensaver files. or perhaps just copy the screensaver files into a temp directory while the Prime95 program is running. That way, the computer cannot run the files at all. When Prime95 is ended, the files are returned to the original directory. hope this would work... -oliver On Fri, 14 May 1999 21:10:17 +0200 "Steinar H. Gunderson" [EMAIL PROTECTED] writes: OK, second message in a row, I just thought it would be nice to separate them. Has anybody got experience in turning off/disabling screensavers under Win95? We run Prime95 at 40 machines (most of them 486'es, though) at school, and screen savers are CPU hoggers (I suppose... at least everybody tells me so). ___ You don't need to buy Internet access to use free Internet e-mail. Get completely free e-mail from Juno at http://www.juno.com/getjuno.html or call Juno at (800) 654-JUNO [654-5866] Unsubscribe list info -- http://www.scruz.net/~luke/signup.htm
Mersenne: Ghosting Solutions
In our labs we ghost our machines as often as we can: during winter break, spring break, beginning and ending of the summer. We also re-ghost as needed around the lab. When I first started in GIMPS I had Prime95 on each machine locally and just as you describe they lost the work they started each time I re-ghosted. My solution was to put a different copy of Prime95 on the home directory of each workstation on the server they connect to. This works quite well, no work gets lost on a re-ghost now. Also, I can remotely check up on each workstation via this directory on the server. I think this is an excellent way of implementing Prime95 in my case. Of course it does waste about 4 MB of disk space for each workstation * 275 workstations = 1.1 GB which is nothing in server space these days. BTW, we had another problem that I'm in the process of fixing: Every now and then, these machines will be reset using Ghost, losing both the [pq]* files and the worktodo.ini files (the rest of the files are not that important). I'm now writing a Linux pseudo-proxy that will give out the same exponent every time if it has not been cleared. (The [pq]* problem remains unsolved :-( ) If anybody are interested (I could make this work under Windows, too, if there is a need for it), let me know. (Of course, if there is no uncleared exponent, the request will be forwarded to the standard PrimeNet server. I'm wondering how Prime95 will cope if it wants more than one exponent, and gets the same all the time. Setting `Days of work to get' to 1 should fix this problem.) Marc Getty - [EMAIL PROTECTED] - [EMAIL PROTECTED] - ICQ: 12916278 http://www.getty.nethttp://www.vwthing.org Work: 215-204-3291 http://etc.temple.edu/Home/Cell: 215-962-5603 Unsubscribe list info -- http://www.scruz.net/~luke/signup.htm
Re: Mersenne: Stoopid Bureaucracies!
I tried at my company too (A very large Aerospace company based in Washington) and my boss said that it didn't add any value to the company so she couldn't approve it. I gotta find a way to show it does have value... -Chuck If you find one... 1) £50,000 2) Lots of free publicity Unsubscribe list info -- http://www.scruz.net/~luke/signup.htm
Mersenne: It is with deep regrets that I must consider this option...
It has finally happened: SETI@Home (http://setiathome.ssl.berkeley.edu/) has come on-line as a distributed project. While I truly admire the work done at GIMPS by George, Scott and the rest of the team, and I enjoy the 'purity of pursuit' in this scientific endeavor, I am compelled to redirect the majority of my computing team to another target--SETI. I will still leave a few systems working on primes, but there is now the hunt is on... Sadly I will depart from a probable peak of position 59 as team "timesau" with 22.832 cpu/yrs and 811.96 cpu-hrs/day and I will begin to convert to the new task at hand. Thanks again to all who make GIMPS a fantastic resource! Unsubscribe list info -- http://www.scruz.net/~luke/signup.htm
Mersenne: Apology
I publicly apologize to Mark Honey for misinterpreting his motives. +---+ | Jud McCranie [EMAIL PROTECTED] | | | | I have macular stars in my eyes. | +---+ Unsubscribe list info -- http://www.scruz.net/~luke/signup.htm
Mersenne: SETI on ABC News last night
The GIMPS project may be in for some serious competition for America's CPU time now, as ABC news did a story last night on the SETI project and the impending release of software for Mac and Wintel systems. Which do you suppose will sound more appealing to the average person--"the search for enormous prime numbers" or "the search for alien life forms"--especially now that SETI has such high-profile exposure? GIMPS has a big disadvantage in that area. Would anyone care to comment on the appeal of SETI? Personally speaking, it doesn't interest me at all. I don't consider its goals to be terribly useful or important, and I don't think that it has a reasonable chance of accomplishing anything. But as number theory enthusiast I find something intrinsically interesting and worthwhile about finding factors and searching for Mersenne primes. I am probably in a minority of the general population in this regard. Let's just hope we don't lose too many existing GIMPS accounts. ___ Get Free Email and Do More On The Web. Visit http://www.msn.com Unsubscribe list info -- http://www.scruz.net/~luke/signup.htm
RE: Mersenne: I am curious
---QUOTE--- I was being a little sarcastic. I think that the original poster who was wanting to team up with other people running at least 10 machines is not in the spirit of the adventure. I think it is fine to make a team if everyone on the team has regular physical access to each machine, but I don't like the idea of teaming up with unknown people. That's why I said "why not make everyone a big team", with a little sarcasm that was probably too subtle. ---QUOTE--- Yes, I would have to agree here. Trying to create groups to get nice big numbers of CPU years I do agree with, but not for the money. By the way, aren't we forgetting something? I hope if anyone DOES win, they will give a reasonable portion to the people who wrote the very highly optimised software to do it, and the people who made sure they weren't re-checking an exponent that hadn't been checked a hundred times before Just out of interest, can I have someone demand I give them a share of the money / stop being in GIMPS if they really wanted to (not that I should think they would...) Unsubscribe list info -- http://www.scruz.net/~luke/signup.htm
RE: Mersenne: Screen saver killers?
-Original Message- From: Steinar H. Gunderson [SMTP:[EMAIL PROTECTED]] Sent: Friday, May 14, 1999 12:10 PM To: [EMAIL PROTECTED] Subject: Mersenne: Screen saver killers? OK, second message in a row, I just thought it would be nice to separate them. Has anybody got experience in turning off/disabling screensavers under Win95? [Gilmore, John (AZ75)] Start - Settings - Control Panel - Activate "Display" Icon - ScreenSaver Tab - then in the Screen Saver window, select either "Blank Screen" or "(None)" Unsubscribe list info -- http://www.scruz.net/~luke/signup.htm
Mersenne Digest V1 #556
Mersenne Digest Friday, May 14 1999 Volume 01 : Number 556 -- Date: Thu, 13 May 1999 04:02:16 -0400 From: [EMAIL PROTECTED] Subject: Mersenne: M617 sieving This weekend sieving for M617 has started. Binaries for DOS and linux are available from ftp://ftp.netdoor.com/users/acurry/nfs Let me know if you want binaries for other platforms. - - I will provide source code for anyone who wants it.Bob I'm messing around trying to construct more "user-friendly" factoring site. However, I'm perpetually stymied by the nitty-gritty details of most of the algorithms. If you don't mind me reading the code in an effort to fix this, I'd love a copy. On a somewhat related note, is there a mailing list anywhere which deals with factoring? I hate to admit it, but I don't have nearly enough time to plow through sci.math with any sort of regularity. On top of which, my newsfeed seems to be getting flakier by the day. I also noticed alot of exponents in excess of 7m cropping up in the account, does this indicate that we've caught up on the bug problem? Unsubscribe list info -- http://www.scruz.net/~luke/signup.htm -- Date: Thu, 13 May 1999 12:19:09 +0200 (CEST) From: Henrik Olsen [EMAIL PROTECTED] Subject: Re: Mersenne: Factoring bignums On Wed, 12 May 1999, Pierre Abbat wrote: Is there a program for factoring numbers up to, say, 2^128 in a reasonable time? I tried bc but it doesn't have a factor command, so I wrote a loop and it spent all its time outputting. Get Richard Crandall's giantint package, it contains factor, which will factor "any size" numbers, using a variety of algorithms. As for time, I just did a quick test: echo 123456789012345678901234567890123456789012345678901234 |time ./factor Sieving... 2 * 73 * 1723 * Commencing Pollard rho... .. .. .. Commencing Pollard (p-1)... .. Commencing ECM... Choosing curve 1, with s = 352116908, B = 1000, C = 5: .. 17108860903 * 28685059068699533197755335074782923141 14.11user 0.01system 0:15.72elapsed 89%CPU This on a 188MHz Pentium. You can get giantint from http://www.perfsci.com/ - -- Henrik Olsen, Dawn Solutions I/S URL=http://www.iaeste.dk/~henrik/ `Can you count, Banjo?' He looked smug. `Yes, miss. On m'fingers, miss.' `So you can count up to ...?' Susan prompted. `Thirteen, miss,' said Banjo proudly. Terry Pratchett, Hogfather Unsubscribe list info -- http://www.scruz.net/~luke/signup.htm -- Date: Thu, 13 May 1999 10:29:07 -0600 From: "Aaron Blosser" [EMAIL PROTECTED] Subject: RE: Mersenne: LLL Perhaps Scott could contact this fellow and see whats up? I know when I had a bunch of machines checking out numbers, Scott emailed me to be sure it wasn't just a bug or something. At any rate, it looks like all the LL exponents have been chewed up. Moreover, judging by the estimated time to complete, these are not fast machines... LLL would probably have been better off leaving the auto assignment turned on because I'd guess it would have taken DoubleCheck exponents. I hope this person knows what they're doing! Some of those numbers checked out all had the same ComputerID...hmmm... Aaron -Original Message- From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED]]On Behalf Of david campeau Sent: Wednesday, May 12, 1999 11:45 PM To: [EMAIL PROTECTED] Subject: Mersenne: LLL I don`t know what is going on, but LLL as about 6000 exponnent reserved and it`is still asking for more! If this continue for the night there should not be any exponnent available this morning. David, P.S. current number of exponents available 5351 (5AM UTC) Unsubscribe list info -- http://www.scruz.net/~luke/signup.htm -- Date: Thu, 13 May 1999 13:59:00 -0500 From: Amy and Shane Sanford [EMAIL PROTECTED] Subject: RE: Mersenne: LLL Maybe there something wrong with the PrimeNet server? Earlier today I had some of my machines check in and it seemed a little wacky. Also, entropia.com seems to be pretty flacky especially the status pages. When I loaded my individual account report it went completely nuts and told me I had 1000's 1000's of numbers check out (then I did a reload and it back to normal). Shane At 10:29 AM 5/13/99 -0600, Aaron Blosser wrote: Perhaps Scott could contact this fellow and see whats up? I know when I had a bunch of machines checking out numbers, Scott emailed me to be sure it wasn't just a bug or something. At any rate, it
Re: Mersenne: Getting maximum speed out of a Linux machine
Brian (and any list members that might be interested): I've ported ReCache to Linux, and tested it out. Oddly enough, it didn't help. I'm not sure if I've ported the spawnl() call in a wrong way (I'm doing a fork() and then an exec()), but it certainly doesn't help (the iteration time goes up from 0.201 to 0.203 secs). Perhaps the Linux MM is better than Windows after all. What _does_ help, however, is killing everything (by going to a different runlevel) and starting again (by going back). This is a bit surprising; I thought this kind of thing was reserved Windows users. (The iteration time is still at 0.201 secs, but it is in fact needed to get maximum speed back after running ReCache.) Result: I don't have any clue about what's going on. /* Steinar */ Unsubscribe list info -- http://www.scruz.net/~luke/signup.htm
RE: Mersenne: I am curious
On Fri, 14 May 1999, Chris Jefferson wrote: ---QUOTE--- I was being a little sarcastic. I think that the original poster who was wanting to team up with other people running at least 10 machines is not in the spirit of the adventure. I think it is fine to make a team if everyone on the team has regular physical access to each machine, but I don't like the idea of teaming up with unknown people. That's why I said "why not make everyone a big team", with a little sarcasm that was probably too subtle. ---QUOTE--- Yes, I would have to agree here. Trying to create groups to get nice big numbers of CPU years I do agree with, but not for the money. By the way, aren't we forgetting something? I hope if anyone DOES win, they will give a reasonable portion to the people who wrote the very highly optimised software to do it, and the people who made sure they weren't re-checking an exponent that hadn't been checked a hundred times before Just out of interest, can I have someone demand I give them a share of the money / stop being in GIMPS if they really wanted to (not that I should think they would...) One thing most people seems to have forgotten when it comes to talk about the money, is that according to the common scientific discovery rules George Woltman and Scott Kurowski will be co-discoverers of all primes found using mprime/prime95, and the client/server setup, so should rightfully get a big part of the money. Either half to George and Scott, half to the "winning" team, to split it equally between the runner(s) and the coders, or split equally between all involved, with George and Scott getting equal shares with the people in the team. Personally I have a problem with the idea of forgetting to credit the people who made it all possible, though I can understand if people think giving up 25.000$ for a principle sounds idealistic. :) -- Henrik Olsen, Dawn Solutions I/S URL=http://www.iaeste.dk/~henrik/ `Can you count, Banjo?' He looked smug. `Yes, miss. On m'fingers, miss.' `So you can count up to ...?' Susan prompted. `Thirteen, miss,' said Banjo proudly. Terry Pratchett, Hogfather Unsubscribe list info -- http://www.scruz.net/~luke/signup.htm
RE: Mersenne: usefulness and 486's
On Fri, 14 May 1999, Aaron Blosser wrote: It does seem lopsided that your 486 takes 2 weeks to do what one of those beastly PIII-550's can do in 12 hours (just a guess), but that's still 12 hours less, and that beastly CPU can do a lot more LL iterations in that time, so you'd still be contributing to the effort. On a related note, I've found that for LL testing, the speed of a Pentium MMX and a Pentium II is about the same adjusted for clock speed, but for factoring, the P-II seems to finish in about half the time. This indicates to me that there's room for more improvement, though how that would work in details isn't clear, since I couldn't find the source for the factoring part of the program when I looked. -- Henrik Olsen, Dawn Solutions I/S URL=http://www.iaeste.dk/~henrik/ `Can you count, Banjo?' He looked smug. `Yes, miss. On m'fingers, miss.' `So you can count up to ...?' Susan prompted. `Thirteen, miss,' said Banjo proudly. Terry Pratchett, Hogfather Unsubscribe list info -- http://www.scruz.net/~luke/signup.htm
RE: Mersenne: usefulness and 486's
A few weeks ago, after growing tired of my mother worrying about her 486DX2-66 machine blowing up on 1/1/00, I gave in and let her have my P120 machine (with a Y2K complient bios) in exchange for her 486 machine. I now have this 486 machine doing the factoring that was being done by the P120. Generally, it is taking almost two weeks per exponent. The 486 really isn't as good at doing those types of calculations, but still... Therefore, my question is even with this two week time, is the 486 machine doing "useful work" for GIMPS, or is it merely heating up the CPU? I ask this because in a couple or three weeks I may have access to a quantity (30 to 75) of 486DX-50 machines. If these machines can contribute useful work to GIMPS, I will happily give them each a mouthful of exponents to factor :-) I'd say most definitely. Factoring is a good workout for a 486, but considering that current factoring assignments are approaching 10M, it will still be quite some time before the LL tests get up that far, so the fact that it takes a 486 a couple weeks just to trial factor those numbers is of little consequence. It does seem lopsided that your 486 takes 2 weeks to do what one of those beastly PIII-550's can do in 12 hours (just a guess), but that's still 12 hours less, and that beastly CPU can do a lot more LL iterations in that time, so you'd still be contributing to the effort. You could also do something more conducive to the 486 environment...I don't know - is ECM good work for a 486? Aaron Unsubscribe list info -- http://www.scruz.net/~luke/signup.htm
Re: Mersenne: Getting maximum speed out of a Linux machine
On Sat, 15 May 1999 [EMAIL PROTECTED] wrote: snips Me neither, so far as linux is concerned. I did find that ReCache "worked" in the sense of at least making things no worse on a wide selection of systems running Win 9x NT. It had least effect on systems which had minimal physical memory, and most effect on systems with lots of memory high clock rate multipliers. e.g. on my PII-333 system (96 MB, NT WS 4.0) on a 256K FFT a "random" start of Prime95 gets an iteration time somewhere between 0.190 0.195 - towards the high end if Prime95 is started automatically by means of a shortcut in the "Startup" folder - whereas using ReCache I get 0.188 _consistently_. If you find ReCache doesn't work for you - even on a Windoze machine - then I'm sorry, but you do have the option not to use it! This really sounds like it's a result of Intel's policy of making their chipsets as cheap as possible. It's a well known problem that several of the widely used Intel chipsets can't cache memory above 64MB. What isn't so well known is that this makes a real difference on Windows, since Windows uses memory from the high addresses first, so the first programs to be started ends up in uncacheable memory. This is why it's sometimes possible to see machine performance drop when you add more memory. I suspect the real reason why ReCache makes prime95 faster is that it uses up all the non cacheable memory, then loads prime85 in chacheable. This will also be the reason for the difference on Linux, since the VM model is completely different, and Linux specifically works around some of the stupidities of the chipsets. -- Henrik Olsen, Dawn Solutions I/S URL=http://www.iaeste.dk/~henrik/ `Can you count, Banjo?' He looked smug. `Yes, miss. On m'fingers, miss.' `So you can count up to ...?' Susan prompted. `Thirteen, miss,' said Banjo proudly. Terry Pratchett, Hogfather Unsubscribe list info -- http://www.scruz.net/~luke/signup.htm
Re: Mersenne: Stoopid Bureaucracies!
I tried at my company too (A very large Aerospace company based in Washington) and my boss said that it didn't add any value to the company so she couldn't approve it. I gotta find a way to show it does have value... -Chuck If you find one... 1) £50,000 2) Lots of free publicity Those are precisely the reasons that scared the hell out of her. Publicity is one of those things that large companies rarely want, only because it rarely benefits them. I can't help but to agree with her on the publicity thing. What do the stockholders care about mersenne primes. They wanna see more airplanes not more mersenne primes. More airplanes = more money... On the bright side, Boe... 'er I mean the large Washington based aerospace company I work for, is involved in some very interesting pure mathematical research because it does benefit many of the projects they work on. My goal is to find one of those researchers and make them a sponsor of this project. This way it would be seen as an innovation rather than a waste of resources. Disclaimer: If you think anything on this page represents the opinions of anyone but me, you're absolutely crazy. These are my opinions and mine alone. -- ~~~ : WWW: http://www.silverlink.net/poke : : E-Mail: [EMAIL PROTECTED] : ~~ : Ask Mike! Aviation's response to Dear : : Abby. http://www.avstarair.com: ~~~ Unsubscribe list info -- http://www.scruz.net/~luke/signup.htm
Mersenne: Re: Mersenne Digest V1 #557
On Sat, May 15, 1999 at 12:52:56AM -0700, Mersenne Digest wrote: Why not just write a piece of code that (during installation of Prime95) removes the screensaver start-up line in the ini (windows) files. Well, as Prime95 is only installed once, and the users are adding screen savers all the time, this will help little. --- snip --- My solution was to put a different copy of Prime95 on the home directory of each workstation on the server they connect to. On our system, each workstation doesn't have their own home directory, and there's nothing I can do about it :-( Of course it does waste about 4 MB of disk space for each workstation * 275 workstations = 1.1 GB which is nothing in server space these days. Well, considering that our server is only a bit over 10 GB, and we're already hard pushed ;-) --- snip --- I now have this 486 machine doing the factoring that was being done by the P120. Generally, it is taking almost two weeks per exponent. The machines at school are estimated to use about three months each (not 24/7) :-) Just have patience. Therefore, my question is even with this two week time, is the 486 machine doing "useful work" for GIMPS, or is it merely heating up the CPU? I ask this because in a couple or three weeks I may have access to a quantity (30 to 75) of 486DX-50 machines. If these machines can contribute useful work to GIMPS, I will happily give them each a mouthful of exponents to factor :-) I asked George about this once. His answer (not 100% accurate, but the wording was similiar): `Every little bit helps!' Yes, they will be doing useful work. --- snip --- On a related note, I've found that for LL testing, the speed of a Pentium MMX and a Pentium II is about the same adjusted for clock speed, but for factoring, the P-II seems to finish in about half the time. Curious -- I'm just now talking to George about improving the LL code for P6 (PPro/PII/PIII). Some volunteer already improved the factoring code. Look at the whatsnew.txt file from Prime95: === New features in Version 14.3 of prime95.exe --- 1) The Pentium Pro factoring code is nearly twice as fast compared to version 14.2. === In fact, 486 code is often better for P6 than Pentium code. If you don't believe me, look at what GNU did for glibc2 (the now-standard Linux C library): In the bottom, you have C code. In addition, there are some special 386-optimized routines. The 486 `inherits' these routines, and adds some more 486-specific ones. Pentium inherits the 386 and 486 routines. BUT... Pentium Pro inherits _only_ 386 and 486 code, not Pentium code. (Trying to pipeline FPU-code the same way for P6 as for Pentium is generally not helping, since the P6 does out-of-order execution already. Therefore, all the FXCHs are not useful, and just eating up time.) The problem is that I don't have access to masm :-( George has promised to look into PII optimization for v19. This indicates to me that there's room for more improvement, though how that would work in details isn't clear, since I couldn't find the source for the factoring part of the program when I looked. It's in factor64.asm (at least part of it). --- I'm not in favour of "forcing" this solution on to users, it sounds a bit draconian to me. Also, I've been known to criticise vehemently software vendors whose setup programs trample on users' personalizations. The problem is that our `users' (pupils) generally know nothing about PCs, most of them can barely surf the web. Trying to explain to them that a maths program (there are not many of us knowing that Prime95 is running anyway, I've used `No Icon' so people won't tamper with it) doesn't like screen savers will not be very constructive. I would much prefer a programme of user education - either convince them that animated screensavers are a waste of resources which could be used more profitably, or at least get them to change the priority of Prime95 to 4 so that it's guaranteed a reasonable chance of getting CPU cycles. For `normal' users, yes. For those, sorry, no :-) BTW experiments indicate that screensavers usually don't consume more than 25% of the available CPU cycles anyway. I'd rather have a user who feels they really need their animated screensaver run Prime95 at 75% of its potential than not run Prime95 at all. What about letting users keep their screen savers all day, and just reset them the first time they're rebooted in the morning? --- snip --- Perhaps because the LL test is critical on the FPU speed, whereas the factoring code is critical on the integer part of the CPU, in particular the efficiency of the (I)MUL instruction. There are two factoring versions: FPU (for Pentium/P6) and integer (all others). This would appear to adequately explain the performance difference between P5 and P6, and also give an explanation as to why the 486 is apparently so much less efficient even after correction for clock speed. P6 is a
