Mersenne: V19 Bug in Factoring Timing?
I just upgraded this morning to the latest Linux version of Linux, v19.0.2. The box with the problem is a 486-66 running Red Hat 5.0. I'm noticing some strange numbers in its output: Factoring M10533203 to 2^64 is 39.83% complete. 310.800 sec. Factoring M10533203 to 2^64 is 39.84% complete. 310.790 sec. Factoring M10533203 to 2^64 is 39.84% complete. 310.630 sec. Factoring M10533203 to 2^64 is 39.85% complete. -3984.147 sec. Factoring M10533203 to 2^64 is 39.86% complete. 310.720 sec. Factoring M10533203 to 2^64 is 39.86% complete. 311.920 sec. About one out of every 15 lines says it completed in negative time. _ Unsubscribe list info -- http://www.scruz.net/~luke/signup.htm Mersenne Prime FAQ -- http://www.tasam.com/~lrwiman/FAQ-mers
Mersenne: Error 2250 connecting to PrimeNet
I have a 486-66 running RedHat 5.0 which I just upgraded to mprime 19.0.2 this morning. It's getting error 2250 when connecting to PrimeNet. Here's what I see when I select "Help/About PrimeNet Server" from the mprime menu: Your choice: 20 Contacting PrimeNet Server. ERROR 2250: Server unavailable The FAQ at http://www.entropia.com/ips/faq.html may have more information. Unable to get version information from PrimeNet server. Hit enter to continue: This returns immediately and it appears that it's not even trying to connect. When I do the same from the old (18.1.2) version it works just fine. I can also connect to PrimeNet from 19.0.2 on another machine. The FAQ suggests I might be using RPC instead of HTTP to communicate with PrimeNet. As far as I can tell I'm not; I have a line that says "UseHTTP=1" in my prime.ini file. One change I had to make between versions is that I had to switch to using sprime.tar.gz since the old RedHat box doesn't have glibc2.1 installed. _ Unsubscribe list info -- http://www.scruz.net/~luke/signup.htm Mersenne Prime FAQ -- http://www.tasam.com/~lrwiman/FAQ-mers
Mersenne: Re: V19 Bug in Factoring Timing?
On Sun, Oct 10, 1999 at 01:24:35PM -0400, Walt Mankowski wrote: I just upgraded this morning to the latest Linux version of Linux, The latest Linux version of Linux? Great! :-) About one out of every 15 lines says it completed in negative time. There was a similiar bug discovered during the QA phase, although it was during ECM, not factoring. I'm sure George will fix it somehow... /* 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
Mersenne: Status report: battery......
I got George's latest newsletter today, with the wonderful news that
Prime95 now supports my most wanted feature: The ability to turn itself off
when my laptop is running on battery.
Unfortunately, I'm out of luck:
- my laptop runs NT (version 4)
- the NT version does not support the feature
- the Win95/98 version doesn't stop when run under NT
So it's back to the manual method ("net stop" and "net start") for me
again.if the person who understands how APM stuff works on NT reads
this, I'm sure he can patch it into the source in a few hours..
for the TODO list, I suppose.
Harald A
--
Harald Tveit Alvestrand, Maxware, Norway
[EMAIL PROTECTED]
_
Unsubscribe list info -- http://www.scruz.net/~luke/signup.htm
Mersenne Prime FAQ -- http://www.tasam.com/~lrwiman/FAQ-mers
Mersenne: glitches in mprime v19?
I just grabbed sprime.tar.gz off the ftp site, and to my dismay, I have run into a couple of problems: a) When I start the new version, I keep getting "Error 2250: Server unavailable" messages. I have "UseHTTP=1" in prime.ini, so what gives? I switched back to v18.1.2 and it connects to entropia.com just fine. b) When I killed mprime (v19), I realized that it had been *factoring* my highest exponent (which was not the current one being tested), when it's explicitly set up to do Lucas-Lehmer tests only. Is this the intended behavior? The number that it was factoring is 710, if that's pertinent. I can't recall if these problems were discussed on the list already during beta-testing (out of curiosity, is there an archive of recent messages?). Thanks in advance for any pointers or insights. -jason _ Unsubscribe list info -- http://www.scruz.net/~luke/signup.htm Mersenne Prime FAQ -- http://www.tasam.com/~lrwiman/FAQ-mers
Re: Mersenne: glitches in mprime v19?
On Sun, 10 Oct 1999, jason wrote: I just grabbed sprime.tar.gz off the ftp site, and to my dismay, I have run into a couple of problems: a) When I start the new version, I keep getting "Error 2250: Server unavailable" messages. I have "UseHTTP=1" in prime.ini, so what gives? I switched back to v18.1.2 and it connects to entropia.com just fine. No idea, I haven't seen a problem with either the LInux or the Win95 versions. b) When I killed mprime (v19), I realized that it had been *factoring* my highest exponent (which was not the current one being tested), when it's explicitly set up to do Lucas-Lehmer tests only. Is this the intended behavior? The number that it was factoring is 710, if that's pertinent. The documentation explicitely mentions that it's factoring to a higher limit in the new version, and since catching up in factoring before going on to LL testing has been a feature for a while, it's not _that_ surprising. Note that this catching up on factoring includes factoring newly assigned exponents BEFORE continuing with the current LL test and it isn't a new thing, but has been there as long as I've been on this project. I can't recall if these problems were discussed on the list already during beta-testing (out of curiosity, is there an archive of recent messages?). Thanks in advance for any pointers or insights. -jason _ Unsubscribe list info -- http://www.scruz.net/~luke/signup.htm Mersenne Prime FAQ -- http://www.tasam.com/~lrwiman/FAQ-mers -- Henrik Olsen, Dawn Solutions I/S URL=http://www.iaeste.dk/~henrik/ A Pentium is a terrible thing to waste, http://www.mersenne.org/prime.htm _ Unsubscribe list info -- http://www.scruz.net/~luke/signup.htm Mersenne Prime FAQ -- http://www.tasam.com/~lrwiman/FAQ-mers
Re: Mersenne: glitches in mprime v19?
Jason wrote: b) When I killed mprime (v19), I realized that it had been *factoring* my highest exponent (which was not the current one being tested), when it's explicitly set up to do Lucas-Lehmer tests only. Is this the intended behavior? The number that it was factoring is 710, if that's pertinent. Was this exponent newly assigned or already had it been assigned at some point in the past? All versions attempt to trial-factor exponents (if necessary) to a certain point before beginning the L-L tests. To give an accurate amount of work queued up, they will perform any necessary trial-factoring (for those exponents to be L-L tested) before they begin (or continue) any further L-L testing. As it happens, v19 has new breakpoints in trail-factoring. For example, all exponents in the 8-9 million range had previously been trial-factored only to 2^63, however v19's breakpoints will cause additional trial-factoring to 2^64 to be attempted for exponents above 8.25 million. There are also some exponents (between 7-7.27 million) that has only been trial-factored thru 2^62. V19 will attempt additional trial-factoring thru 2^63 for these exponents. The limits for v18 and v19 trail-factoring (in millions) are: Trial-FactoringV18V19 thrurange(s) range(s) === = === 2^72 --- 71.00 - 79.30 2^71 --- 57.02 - 71.00 2^70 --- 44.15 - 57.02 2^69 --- 35.10 - 44.15 2^68 --- 28.13 - 35.10 2^67 --- 21.59 - 28.13 2^66 --- 17.85 - 21.59 2^65 --- 13.38 - 17.85 2^64 9.15 - 20.40 8.25 - 13.38 2^63 7.27 - 9.15 6.515 - 8.25 2^62 5.16 - 7.27 5.16 - 6.515 2^61 3.96 - 5.16 3.96 - 5.16 2^60 2.95 - 3.96 2.95 - 3.96 2^59 2.655 - 2.95 2.36 - 2.95 2^58 2.135 - 2.655 1.93 - 2.36 2^57 1.675 - 2.135 1.48 - 1.93 Eric Hahn _ Unsubscribe list info -- http://www.scruz.net/~luke/signup.htm Mersenne Prime FAQ -- http://www.tasam.com/~lrwiman/FAQ-mers
