Mersenne Digest Wednesday, January 26 2000 Volume 01 : Number 684
----------------------------------------------------------------------
Date: Mon, 24 Jan 2000 01:59:42 PST
From: "Dennis Peter" <[EMAIL PROTECTED]>
Subject: Mersenne: Q:GIMPS freezes my system
Hi there, I've experienced a problem. I'm wondering if anyone else has had
this problem:
GIMPS freezes my system after it runs for about 24 hours or so.
I have Windows NT 4.0 (SP5)
Dual Intel Celeron 433Mhz (BP6 motherboard)
128MB RAM
If I quit both GIMPS processes on my machine (and I'm running one with the
- -A2 command-line option) and start them again, no problems. It's only after
both have been running for some time. But I don't want to do this because I
leave my computer for days at a time.
Is there a work-around i can use? I've experimented with the AT WinNT
task-scheduler, but I can't get the GIMPS command-line option -Tdd:hh:mm to
work. Has this been implemented?
Thanks in advance,
Dennis
______________________________________________________
Get Your Private, Free Email at http://www.hotmail.com
_________________________________________________________________
Unsubscribe & list info -- http://www.scruz.net/~luke/signup.htm
Mersenne Prime FAQ -- http://www.tasam.com/~lrwiman/FAQ-mers
------------------------------
Date: Mon, 24 Jan 2000 05:47:09 -0500 (EST)
From: "St. Dee" <[EMAIL PROTECTED]>
Subject: Re: Mersenne: Q:GIMPS freezes my system
On Mon, 24 Jan 2000, Dennis Peter wrote:
> Hi there, I've experienced a problem. I'm wondering if anyone else has had
> this problem:
>
> GIMPS freezes my system after it runs for about 24 hours or so.
>
> I have Windows NT 4.0 (SP5)
> Dual Intel Celeron 433Mhz (BP6 motherboard)
> 128MB RAM
Hi Dennis,
How sure are you that the prime search program is the culprit here? There
have been a slew of posts to the Abit USENET newsgroup, as well as a whole
lot of discussion at www.bp6.com, about problems with the BP6 randomly
freezing. People running GUI operating systems on it seem to be reporting
much more difficulty than those running something like Linux in text mode.
I'm running a BP6 with dual Celeron 400's, and since I switched it over to
running RedHat 6.0 in console mode (no XWindows), it's been the picture of
stability (it is always up running 2 instances of mprime unless I bring it
down for new hardware).
I'm not saying that the prime search program (Prime95?) you're using isn't
the problem, just that the BP6 seems to have some "stability issues" in
general.
Kel
_________________________________________________________________
Unsubscribe & list info -- http://www.scruz.net/~luke/signup.htm
Mersenne Prime FAQ -- http://www.tasam.com/~lrwiman/FAQ-mers
------------------------------
Date: Mon, 24 Jan 2000 11:48:25 +0100
From: Paul Landon <[EMAIL PROTECTED]>
Subject: Re: Mersenne: Size of largest prime factor
> Pierre Abbat wrote:
> >If I pick a huge number n at random, how much smaller than n, on average,
> is
> >its largest prime factor?
>
Jud McCranie wrote:-
> On the average, the largest prime factor of n is n^0.6065, and the second
> largest is n^0.2117. Reference: Knuth, the Art of Computer Programming,
> vol 2, section 4.5.4.
>
But for Mersennes this might not be the case.
For the size of exponents that we deal with Mersennes are less
composite than a random set of ones & zeroes.
There are many reasons for this, if 2^p-1 has any factors they
must be bigger than p. They must be +-1 mod 8 etc.
Looking at the string of ones it certainly has regularity. Indeed
there is a measure for it, the order of 2 mod 2^p-1 which is very
low, =p; and any factors have this order as well. This is not
average.
This is not new news to most people here, but I have to remind
myself, it still hasn't been proved whether there are an infinite
number of Mersenne Primes or an infinite number of Mersenne
composites.
Cheers,
Paul Landon
_________________________________________________________________
Unsubscribe & list info -- http://www.scruz.net/~luke/signup.htm
Mersenne Prime FAQ -- http://www.tasam.com/~lrwiman/FAQ-mers
------------------------------
Date: Mon, 24 Jan 2000 05:27:07 PST
From: "Dennis Peter" <[EMAIL PROTECTED]>
Subject: Re: Mersenne: Q:GIMPS freezes my system
Kel,
I have run other numerous programs (including other number crunching spare
CPU-cycle programs) and my system has never froze until I ran Prime95.exe.
I installed GIMPS 7 days ago. I've had my system for about a year or so.
I'll check out bp6.com... and yes, I know that the dual-celeron config isn't
officially supported by Intel.
- -Dennis
- ----Original Message Follows----
From: "St. Dee" <[EMAIL PROTECTED]>
To: Dennis Peter <[EMAIL PROTECTED]>
CC: [EMAIL PROTECTED]
Subject: Re: Mersenne: Q:GIMPS freezes my system
Date: Mon, 24 Jan 2000 05:47:09 -0500 (EST)
On Mon, 24 Jan 2000, Dennis Peter wrote:
> Hi there, I've experienced a problem. I'm wondering if anyone else has
had
> this problem:
>
> GIMPS freezes my system after it runs for about 24 hours or so.
>
> I have Windows NT 4.0 (SP5)
> Dual Intel Celeron 433Mhz (BP6 motherboard)
> 128MB RAM
Hi Dennis,
How sure are you that the prime search program is the culprit here? There
have been a slew of posts to the Abit USENET newsgroup, as well as a whole
lot of discussion at www.bp6.com, about problems with the BP6 randomly
freezing. People running GUI operating systems on it seem to be reporting
much more difficulty than those running something like Linux in text mode.
I'm running a BP6 with dual Celeron 400's, and since I switched it over to
running RedHat 6.0 in console mode (no XWindows), it's been the picture of
stability (it is always up running 2 instances of mprime unless I bring it
down for new hardware).
I'm not saying that the prime search program (Prime95?) you're using isn't
the problem, just that the BP6 seems to have some "stability issues" in
general.
Kel
______________________________________________________
Get Your Private, Free Email at http://www.hotmail.com
_________________________________________________________________
Unsubscribe & list info -- http://www.scruz.net/~luke/signup.htm
Mersenne Prime FAQ -- http://www.tasam.com/~lrwiman/FAQ-mers
------------------------------
Date: Mon, 24 Jan 2000 17:14:14 +0100 (CET)
From: Henrik Olsen <[EMAIL PROTECTED]>
Subject: Re: Mersenne: Size of largest prime factor
On Mon, 24 Jan 2000, Paul Landon wrote:
> Subject: Re: Mersenne: Size of largest prime factor
>
> > Pierre Abbat wrote:
> > >If I pick a huge number n at random, how much smaller than n, on average,
> > is
> > >its largest prime factor?
> >
> Jud McCranie wrote:-
>
> > On the average, the largest prime factor of n is n^0.6065, and the second
> > largest is n^0.2117. Reference: Knuth, the Art of Computer Programming,
> > vol 2, section 4.5.4.
> >
> But for Mersennes this might not be the case.
> For the size of exponents that we deal with Mersennes are less
> composite than a random set of ones & zeroes.
> There are many reasons for this, if 2^p-1 has any factors they
> must be bigger than p. They must be +-1 mod 8 etc.
> Looking at the string of ones it certainly has regularity. Indeed
> there is a measure for it, the order of 2 mod 2^p-1 which is very
> low, =p; and any factors have this order as well. This is not
> average.
> This is not new news to most people here, but I have to remind
> myself, it still hasn't been proved whether there are an infinite
> number of Mersenne Primes or an infinite number of Mersenne
> composites.
Erhm?
2^n-1 where n is composite is in itself composite, so showing that there
are infinitely many Mersenne composites is easy. :)
>
> Cheers,
> Paul Landon
- --
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
Mersenne Prime FAQ -- http://www.tasam.com/~lrwiman/FAQ-mers
------------------------------
Date: Mon, 24 Jan 2000 17:41:37 +0100
From: Paul Landon <[EMAIL PROTECTED]>
Subject: Re: Mersenne: Size of largest prime factor
Hiya Henrik,
I did mean for 2^p-1; p prime.
That's why I work in Computing not the discipline of Maths :-)
I am certain that the graph in Knuth sect 4.5.4 (which by luck
I had only read for the first time last night) is definately
not applicable to Mersenne numbers (with prime exponent).
I am certain that there is a minimum size for any divisor of a
Mersenne (conversely there is a maximum order of 2 mod f for a
candidate factor f) such that there is therefore a maximum size
for the largest factor.
This makes the cumulative frequency graph hit 1.0 before 100%!
(I had best be precise so that one day I grow up to be a good
Mathematician, - ignoring the factorisation 1 & itself).
Can we do some statistics on the (complete) factorisations
we already have?
I am also sure that in many other ways Mersennes do not behave
like random numbers as discussed in sect 4.5.4.
I think I am correct in what I meant to say, that it hasn't
been proved that there are an infinite number of Mersenne
Primes or an infinite number of Mersenne's with prime exponent
that are composite or both with a limiting ratio.
Thanks Henrik for encouraging me to be precise in the presense
of real Mathematicians.
Cheers,
Paul Landon
Henrik Olsen wrote:
> On Mon, 24 Jan 2000, Paul Landon wrote:
> > Subject: Re: Mersenne: Size of largest prime factor
> >[snip]
> > This is not new news to most people here, but I have to remind
> > myself, it still hasn't been proved whether there are an infinite
> > number of Mersenne Primes or an infinite number of Mersenne
> > composites.
> Erhm?
> 2^n-1 where n is composite is in itself composite, so showing that there
> are infinitely many Mersenne composites is easy. :)
_________________________________________________________________
Unsubscribe & list info -- http://www.scruz.net/~luke/signup.htm
Mersenne Prime FAQ -- http://www.tasam.com/~lrwiman/FAQ-mers
------------------------------
Date: Mon, 24 Jan 2000 11:55:10 -0500
From: Jud McCranie <[EMAIL PROTECTED]>
Subject: Re: Mersenne: Size of largest prime factor
At 11:48 AM 1/24/00 +0100, Paul Landon wrote:
> On the average, the largest prime factor of n is n^0.6065,
>
>But for Mersennes this might not be the case.
>For the size of exponents that we deal with Mersennes are less
>composite than a random set of ones & zeroes.
That's right, but the original question just said a large random number.
+--------------------------------------------------------+
| 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, 24 Jan 2000 09:36:04 -0800
From: Eric Hahn <[EMAIL PROTECTED]>
Subject: Mersenne: Re : Odd's on finding a factor
Dave Mullen wrote:
>Sorry, I'm no mathematician, and new to the Mersenne field.
>
>> No, in the x-y bit range (remember that n bit integers are
>> about >2^n) the first factor could be x/2 to y/2 bits long
>> (powers of a power >multiply).
>
>What I was trying to say in my disjointed way was ...
>
>(Example) M11 = 2047 (11 bits long). Now 2047 has <underline>only 2
factors</underline>
>(23 >x 89) and the square root of 2047 is approx 45. 45 is 6 bits
>long, therefore the factor lower than the square root must have
><<= 6 bits, >and the factor higher than the square root must have
>>=6 bits.
>
>23. is 5 bits long, and 89 is 7 bits long.
>
>Thus for the exponent 11650000 bits long, <underline>if it only has 2
factors</underline> , >then the first factor must be between 2 and 3413
bits long, and
>the second factor must be between 3413 and 11649999 bits long.
>Note that the bit lengths of the 2 factors added together must
>equal the bit length of the Prime (or bit length of the
>Prime + 1) !!
There only a slight error with your logic... For the exponent
11650000, the root is *not* 3413 bits long, but more like 5825000
bits long. Perhaps you forgot exponents add, not multiply.
For simplication:
2^3 * 2^3 = 2^6
8 * 8 = 64
Therefore:
2^11650000 = 2^5825000 * 2^5825000
Eric Hahn
_________________________________________________________________
Unsubscribe & list info -- http://www.scruz.net/~luke/signup.htm
Mersenne Prime FAQ -- http://www.tasam.com/~lrwiman/FAQ-mers
------------------------------
Date: Mon, 24 Jan 2000 22:09:12 -0000
From: "Brian J. Beesley" <[EMAIL PROTECTED]>
Subject: Re: Mersenne: Size of largest prime factor
On 24 Jan 00, at 11:48, Paul Landon wrote:
> This is not new news to most people here, but I have to remind
> myself, it still hasn't been proved whether there are an infinite
> number of Mersenne Primes or an infinite number of Mersenne
> composites.
The latter conjecture looks very, very probable!
Note that it would be sufficient to prove that there are an infinite
number of Sophie-Germain primes, since there is a "well known"
theorem which states that, if p is a Sophie-Germain prime, then 2^p-1
is divisible by 2p+1.
Of course, we do have heuristics which tend to indicate that the
number of Mersenne primes is infinite. If this is not so, then the
number of Mersenne composites _must_ be infinite. The same heuristics
(or even just application of the Prime Number Theorem) suggest that
the probability that 2^p-1 is prime decreases with increasing p,
which is a strong indication that there are, indeed, an infinite
number of Mersenne composites.
The contrary would be amusing - if there are a finite number of
Mersenne composites, there must be an integer P which is the exponent
of the _largest_ composite Mersenne number, i.e. 2^(P+k)-1 is prime
for every positive integer k. The challenge then would not be to find
all the Mersenne primes, but to determine the value of P.
Regards
Brian Beesley
_________________________________________________________________
Unsubscribe & list info -- http://www.scruz.net/~luke/signup.htm
Mersenne Prime FAQ -- http://www.tasam.com/~lrwiman/FAQ-mers
------------------------------
Date: Mon, 24 Jan 2000 17:42:05 -0500 (EST)
From: Lucas Wiman <[EMAIL PROTECTED]>
Subject: Re: Mersenne: Re : Odd's on finding a factor (part 2)
> Thus for the exponent 11650000 bits long, if it only has 2 factors , then the first
>factor must be between 2 and 3413 bits long, and the second factor must be between
>3413 and 11649999 bits long. Note that the bit lengths of the 2 factors added
>together must equal the bit length of the Prime (or bit length of the Prime + 1) !!
No, if the exponent=11650000 (I think this is what you mean), the mersenne
number must be this number of bits (as you know). The first factor must
be less than sqrt(M_p), not M_(sqrt(p)), which is what you were doing. The
first factor must therefore be less than 2^(p/2)=2^5825000>2^sqrt(11650000)~=2^3413.
> > You would probably get better results with Will Edgington's mersfacgmp
> program, and DJGPP (a port of g++ to dos).
>
> I'll check it out. I'm using UBASIC because for testing factors of large Mersenne, I
>don't need to actually represent the full Mersenne Prime. If you're familiar with
>UBASIC try ...
>
> Result = MODPOW(2,MersenneExponent,TrialFactor)
>
> where TrialFactor is the MersenneExponent * 2 * (some k in range 1 to 2^16) + 1.
>
> If Result = 1 then TrialFactor divides the Mersenne Prime. As UBASIC can handle
>around 2600 decimal digits, in theory (and with a lot of time), I could check all
>factors up to 2600 decimal digits for any given exponent. It's a damn sight faster
>than filling 16MB+ of memory with 1's and then trial dividing.
Will's program does this also (see Knuth volume II (I think 4.5.3, but I'm
not sure don't have my copy handy), or the mersenne FAQ at the bottom of this
message). Thinking on it in a slightly more awake state, I'm not sure how much
faster it would be, but try it anyway. You will need to download DJGPP (search
altavista for DJGPP), set it up, compile the GMP library
(ftp://metalab.unc.edu/gnu/gmp (I think), there is a dos batch file), and compile
Will's program
(I think it is in the links section of the mailling list FAQ). But, under
windows, would George's program be the fastest (or are you using very large
exponents?)
Have fun,
Lucas Wiman
_________________________________________________________________
Unsubscribe & list info -- http://www.scruz.net/~luke/signup.htm
Mersenne Prime FAQ -- http://www.tasam.com/~lrwiman/FAQ-mers
------------------------------
Date: Mon, 24 Jan 2000 17:46:17 -0500 (EST)
From: Lucas Wiman <[EMAIL PROTECTED]>
Subject: Re: Mersenne: Q:GIMPS freezes my system
> Hi there, I've experienced a problem. I'm wondering if anyone else has had
> this problem:
>
> GIMPS freezes my system after it runs for about 24 hours or so.
>
> I have Windows NT 4.0 (SP5)
> Dual Intel Celeron 433Mhz (BP6 motherboard)
> 128MB RAM
>
> If I quit both GIMPS processes on my machine (and I'm running one with the
> -A2 command-line option) and start them again, no problems. It's only after
> both have been running for some time. But I don't want to do this because I
> leave my computer for days at a time.
>
It sounds like you might have problems with overheating in the case which can
often cause lockups, since Prime95 stresses the CPU more, it produces more heat.
I would advise getting a case fan.
Just my $0.00 (the GNU version)
- -Lucas
_________________________________________________________________
Unsubscribe & list info -- http://www.scruz.net/~luke/signup.htm
Mersenne Prime FAQ -- http://www.tasam.com/~lrwiman/FAQ-mers
------------------------------
Date: Tue, 25 Jan 2000 01:04:02 +0100
From: [EMAIL PROTECTED] (Dieter Schmitt)
Subject: Mersenne: Round off error = 0.5
This is a multi-part message in MIME format.
- ------=_NextPart_000_0027_01BF66D0.11B26D40
Content-Type: text/plain;
charset="iso-8859-1"
Content-Transfer-Encoding: quoted-printable
Hi all,
I'm running Prime95 on a PII-400 for 6 days (no overclocking) at =
exponent 9409271. It's produced several outputs concerning ROUND OFF =
ERRORS - the last one is ROUND OFF [0.5] > 0.4
What to do now? Restart from iteration 1?
Regards
Dieter Schmitt
- ------=_NextPart_000_0027_01BF66D0.11B26D40
Content-Type: text/html;
charset="iso-8859-1"
Content-Transfer-Encoding: quoted-printable
<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.0 Transitional//EN">
<HTML><HEAD>
<META content=3D"text/html; charset=3Diso-8859-1" =
http-equiv=3DContent-Type>
<META content=3D"MSHTML 5.00.2614.3500" name=3DGENERATOR>
<STYLE></STYLE>
</HEAD>
<BODY bgColor=3D#ffffff>
<DIV><FONT face=3DArial size=3D2>Hi all,</FONT></DIV>
<DIV> </DIV>
<DIV><FONT face=3DArial size=3D2>I'm running Prime95 on a PII-400 for 6 =
days (no=20
overclocking) at exponent 9409271. It's produced several outputs =
concerning=20
ROUND OFF ERRORS - the last one is ROUND OFF [0.5] > 0.4</FONT></DIV>
<DIV> </DIV>
<DIV><FONT face=3DArial size=3D2>What to do now? Restart from iteration=20
1?</FONT></DIV>
<DIV> </DIV>
<DIV><FONT face=3DArial size=3D2>Regards</FONT></DIV>
<DIV> </DIV>
<DIV><FONT face=3DArial size=3D2>Dieter =
Schmitt</FONT></DIV></BODY></HTML>
- ------=_NextPart_000_0027_01BF66D0.11B26D40--
_________________________________________________________________
Unsubscribe & list info -- http://www.scruz.net/~luke/signup.htm
Mersenne Prime FAQ -- http://www.tasam.com/~lrwiman/FAQ-mers
------------------------------
Date: Mon, 24 Jan 2000 19:36:47 -0500
From: George Woltman <[EMAIL PROTECTED]>
Subject: Re: Mersenne: Round off error = 0.5
Hi,
At 01:04 AM 1/25/00 +0100, Dieter Schmitt wrote:
>I'm running Prime95 on a PII-400 for 6 days (no overclocking) at
>exponent 9409271. It's produced several outputs concerning
>ROUND OFF ERRORS - the last one is ROUND OFF [0.5] > 0.4
>What to do now? Restart from iteration 1?
The first thing to do is try and figure out if your CPU is
overheating or you have some flaky memory chips. You almost
certainly have a hardware problem of some type.
Prime95 will go back to the last good save file after the error.
So there is a chance your result is OK. The question is "did a
hardware error go undetected?" I have little hard data on this,
but I'd guess several errors of this type mean you have less than a 50-50
chance of producing a good result.
Regards,
George
_________________________________________________________________
Unsubscribe & list info -- http://www.scruz.net/~luke/signup.htm
Mersenne Prime FAQ -- http://www.tasam.com/~lrwiman/FAQ-mers
------------------------------
Date: Mon, 24 Jan 2000 19:55:47 EST
From: [EMAIL PROTECTED]
Subject: Mersenne: Red Red Red
Aaaack! Who's the one sending mail to the list that makes it appear with a
red background?
Stephan "Retinal Afterimage" Lavavej
_________________________________________________________________
Unsubscribe & list info -- http://www.scruz.net/~luke/signup.htm
Mersenne Prime FAQ -- http://www.tasam.com/~lrwiman/FAQ-mers
------------------------------
Date: Tue, 25 Jan 2000 03:32:11 -0000
From: "Andy Steward" <[EMAIL PROTECTED]>
Subject: Re: Mersenne: Odds on finding a factor ?
- ----- Original Message -----
From: Lucas Wiman <[EMAIL PROTECTED]>
> I think under windows that dos windows only run when they are "up".
> (I could be wrong, I've stopped using windows again)
No. You can set a background priority. In Win 95, right-click on the icon,
then click "Properties". Click on the "Misc" tab and move the slider for
"Idle Priority" anywhere you want between Low and High. Click "OK".
If I'm going to be away from a machine for a while, I quite often set a "B"
priority task running in a Ubasic window and minimise it, then run an "A"
priority task running in an active window. That way, I ensure that the
machine won't be idle if the "A" task completes before my return. The
downside is the reduction in resources available to the "A" task while
both are running.
HTH,
Andy Steward
Factorisations of 57,619 Generalised Repunits at:
<http://www.users.globalnet.co.uk/~aads/index.html>
_________________________________________________________________
Unsubscribe & list info -- http://www.scruz.net/~luke/signup.htm
Mersenne Prime FAQ -- http://www.tasam.com/~lrwiman/FAQ-mers
------------------------------
Date: Tue, 25 Jan 2000 03:29:34 -0000
From: "Andy Steward" <[EMAIL PROTECTED]>
Subject: Mersenne: Re: Odd's on finding a factor (part 2)
Dave Mullin wrote:
>Result = MODPOW(2,MersenneExponent,TrialFactor)
>where TrialFactor is the MersenneExponent * 2 * (some k in range 1 to 2^16)
+ 1.
>If Result = 1 then TrialFactor divides the Mersenne Prime. As UBASIC can
>handle around 2600
>decimal digits, in theory (and with a lot of time), I
>could check all factors up to 2600 decimal
>digits for any given exponent.
>It's a damn sight faster than filling 16MB+ of memory with 1's and
>then trial dividing.
Ubasic can indeed handle integers up to around 2600 digits, but modpow()
needs to hold intermediate results, so will only work up to around 1300
digits input. Good luck anyway!
Andy Steward
_________________________________________________________________
Unsubscribe & list info -- http://www.scruz.net/~luke/signup.htm
Mersenne Prime FAQ -- http://www.tasam.com/~lrwiman/FAQ-mers
------------------------------
Date: Tue, 25 Jan 2000 12:22:12 +0800
From: "Dave Mullen" <[EMAIL PROTECTED]>
Subject: Mersenne: Re: Odds on finding a factor
This is a multi-part message in MIME format.
- ------=_NextPart_000_0028_01BF672E.CED6BF80
Content-Type: text/plain;
charset="iso-8859-1"
Content-Transfer-Encoding: quoted-printable
>There only a slight error with your logic... For the exponent
11650000, the root is *not* 3413 bits long, but more like 5825000
bits long. Perhaps you forgot exponents add, not multiply.
Okay, next time I'll open my mouth a little wider, so I can fit both =
feet inside.
Sorry all, sometimes I put my mouth in gear before my brain is engaged.
>Who needs to? I have code which tries any number up to 2^95 as a=20
factor of a Mersenne number with an exponent up to approx. 600=20
million, using less than 1 KByte of code space and 32 bytes of data=20
storage. It executes in a time proportional to the logarithm of the=20
exponent - for an exponent around 35 million, it takes approximately=20
2000 CPU clocks i.e. 4 microseconds per test on a 500 MHz CPU.
My UBASIC is running thus :-
=20
In 36 hours, tested 88855 exponents with all multipliers 1 to 2357. (and =
found factors for 8860 of the exponents).
=20
88855 exponents x 2358 multipliers =3D 20952090 tests =3D 1616 tests per =
second =3D 618 microsecs per test on a P133.
Brian, I'd be very interested in a copy of that code, if you'd care to =
E-Mail it.
Regards
Dave Mullen
- ------=_NextPart_000_0028_01BF672E.CED6BF80
Content-Type: text/html;
charset="iso-8859-1"
Content-Transfer-Encoding: quoted-printable
<!DOCTYPE HTML PUBLIC "-//W3C//DTD W3 HTML//EN">
<HTML>
<HEAD>
<META content=3Dtext/html;charset=3Diso-8859-1 =
http-equiv=3DContent-Type>
<META content=3D'"MSHTML 4.72.3110.7"' name=3DGENERATOR>
</HEAD>
<BODY bgColor=3D#ffffff>
<DIV><FONT color=3D#000000 size=3D2>>There only a slight error with =
your logic...=20
For the exponent<BR>11650000, the root is *not* 3413 bits long, but more =
like=20
5825000<BR>bits long. Perhaps you forgot exponents add, not=20
multiply.</FONT></DIV>
<DIV><FONT color=3D#000000 size=3D2></FONT> </DIV>
<DIV><FONT color=3D#000000 size=3D2>Okay, next time I'll open my mouth a =
little=20
wider, so I can fit both feet inside.</FONT></DIV>
<DIV><FONT color=3D#000000 size=3D2></FONT> </DIV>
<DIV><FONT color=3D#000000 size=3D2>Sorry all, sometimes I put my mouth =
in gear=20
before my brain is engaged.</FONT></DIV>
<DIV><FONT color=3D#000000 size=3D2></FONT> </DIV>
<DIV><FONT color=3D#000000 size=3D2>>Who needs to? I have code which =
tries any=20
number up to 2^95 as a <BR>factor of a Mersenne number with an exponent =
up to=20
approx. 600 <BR>million, using less than 1 KByte of code space and 32 =
bytes of=20
data <BR>storage. It executes in a time proportional to the logarithm of =
the=20
<BR>exponent - for an exponent around 35 million, it takes approximately =
<BR>2000 CPU clocks i.e. 4 microseconds per test on a 500 MHz=20
CPU.<BR></FONT></DIV>
<DIV><FONT color=3D#000000 size=3D2></FONT> </DIV>
<DIV><FONT color=3D#000000 size=3D2>
<DIV><FONT color=3D#000000 size=3D2>My UBASIC is running thus =
:-</FONT></DIV>
<DIV><FONT color=3D#000000 size=3D2></FONT> </DIV>
<DIV><FONT size=3D2>In 36 hours, tested 88855 exponents with all =
multipliers 1 to=20
2357. (and found factors for 8860 of the exponents).</FONT></DIV>
<DIV><FONT size=3D2></FONT> </DIV>
<DIV><FONT size=3D2>88855 exponents x 2358 multipliers =3D 20952090 =
tests =3D 1616=20
tests per second =3D 618 microsecs per test on a P133.</FONT></DIV>
<DIV><FONT size=3D2></FONT> </DIV>Brian, I'd be very interested in =
a copy of=20
that code, if you'd care to E-Mail it.</FONT></DIV>
<DIV><FONT color=3D#000000 size=3D2></FONT> </DIV>
<DIV><FONT color=3D#000000 size=3D2>Regards</FONT></DIV>
<DIV><FONT color=3D#000000 size=3D2></FONT> </DIV>
<DIV><FONT color=3D#000000 size=3D2>Dave =
Mullen</FONT></DIV></BODY></HTML>
- ------=_NextPart_000_0028_01BF672E.CED6BF80--
_________________________________________________________________
Unsubscribe & list info -- http://www.scruz.net/~luke/signup.htm
Mersenne Prime FAQ -- http://www.tasam.com/~lrwiman/FAQ-mers
------------------------------
Date: Tue, 25 Jan 2000 00:35:44 -0500 (EST)
From: Lucas Wiman <[EMAIL PROTECTED]>
Subject: Re: Mersenne: Re: Odds on finding a factor
> >Who needs to? I have code which tries any number up to 2^95 as a
> factor of a Mersenne number with an exponent up to approx. 600
> million, using less than 1 KByte of code space and 32 bytes of data
> storage. It executes in a time proportional to the logarithm of the
> exponent - for an exponent around 35 million, it takes approximately
> 2000 CPU clocks i.e. 4 microseconds per test on a 500 MHz CPU.
This is a bit misleading, more accuratly, it runs in time proportional
to lg(p)*lg^2(n), where p is the exponent, and n is the potential factor.
Just call me "Mr. Pedantic". :)
- -Lucas Wiman
_________________________________________________________________
Unsubscribe & list info -- http://www.scruz.net/~luke/signup.htm
Mersenne Prime FAQ -- http://www.tasam.com/~lrwiman/FAQ-mers
------------------------------
Date: Tue, 25 Jan 2000 19:30:19 -0000
From: "Brian J. Beesley" <[EMAIL PROTECTED]>
Subject: Re: Mersenne: Q:GIMPS freezes my system
On 24 Jan 00, at 1:59, Dennis Peter wrote:
> Hi there, I've experienced a problem. I'm wondering if anyone else has
> had this problem:
>
> GIMPS freezes my system after it runs for about 24 hours or so.
>
> I have Windows NT 4.0 (SP5)
> Dual Intel Celeron 433Mhz (BP6 motherboard)
> 128MB RAM
As it happens, there is an article on the BP6 board in the March
Personal Computer World (UK) which I received today which may be
relevant. ("Hands On: Hardware") Apparently the BX chipset may need
auxilliary cooling when the system is driven hard (which 2 copies of
Prime95 running LL tests will definitely do). The gist of the article
is about overclocked systems, but the practical upshot is that
apparently cooling fans sold for obsolete 486 CPUs fit the BX chipset
package perfectly! Another suggestion is to refit the chipset and CPU
heatsinks using thermal transfer compound, which should help keep the
CPU cool by letting waste heat get to the heatsink more easily.
Maybe the trick of using the 486 CPU fan on the BX chipset package
would assist some other systems, especially those running dual
PII/III's or even fast uniprocessor systems, if they're showing any
signs of flakiness.
Regards
Brian Beesley
_________________________________________________________________
Unsubscribe & list info -- http://www.scruz.net/~luke/signup.htm
Mersenne Prime FAQ -- http://www.tasam.com/~lrwiman/FAQ-mers
------------------------------
Date: Wed, 26 Jan 2000 00:13:04 +0100
From: Michael =?iso-8859-1?Q?=D8stergaard?= Pedersen <[EMAIL PROTECTED]>
Subject: Mersenne: A Windows 2000 Question
Hi,
I am using the NT Service version of Prime95. I thought that when I shut
down windows the service would be stopped and results saved to disk, but
that's not the case. Only if I locate the service and press STOP, the results
are saved to disk.
I thought that a shutdown would stop all services so the results would be
saved, but that doesn't seem to be the case...
I'm using Windows 2000 (2195)
Anyone have similar problems or a workaround?
Regards,
Michael
_________________________________________________________________
Unsubscribe & list info -- http://www.scruz.net/~luke/signup.htm
Mersenne Prime FAQ -- http://www.tasam.com/~lrwiman/FAQ-mers
------------------------------
Date: Tue, 25 Jan 2000 19:32:49 -0800
From: Russel Brooks <[EMAIL PROTECTED]>
Subject: Mersenne: Self Test
My pc just surprised me by deciding to do a self test after it finished
its latest piece of work.
> The program will now perform a self-tesst to make sure the
> Lucas-Lehmer code is working properly on your computer.
I thought the test was only performed when you first started doing GIMPS
processing. Is this normal or has it detected an error in my results?
(later...)
I got "Self-test 256 passed!" so I guess I don't have any problems.
cheers... Russ
_________________________________________________________________
Unsubscribe & list info -- http://www.scruz.net/~luke/signup.htm
Mersenne Prime FAQ -- http://www.tasam.com/~lrwiman/FAQ-mers
------------------------------
Date: Tue, 25 Jan 2000 19:53:20 -0800
From: Kevin Sexton <[EMAIL PROTECTED]>
Subject: Mersenne: OT:Processor glued to heatsink?
This doesn't directly deal with Mersennes, except I planned
on running mprime 24 hours a day on it, but maybe someone
can help.
I got an older pentium machine, working, but no OS (except
the 4 files needed to boot dos6) no CD-ROM, no added cards,
etc. I attempted to remove the heat sink from the cpu, but
ended up with the cpu out too. The problem is that the heat
sink is oversized, covering the zif-lever, and it seems to
be glued to the processor. the "glue seems to be a waxy
substance. Has anyone dealt with this stuff before, or have
an idea how to get the processor seated without being able
to lift the lever?
_________________________________________________________________
Unsubscribe & list info -- http://www.scruz.net/~luke/signup.htm
Mersenne Prime FAQ -- http://www.tasam.com/~lrwiman/FAQ-mers
------------------------------
Date: Tue, 25 Jan 2000 20:08:22 -0800
From: Kevin Sexton <[EMAIL PROTECTED]>
Subject: Re: Mersenne: Self Test
Russel Brooks wrote:
> My pc just surprised me by deciding to do a self test after it finished
> its latest piece of work.
>
> > The program will now perform a self-tesst to make sure the
> > Lucas-Lehmer code is working properly on your computer.
>
> I thought the test was only performed when you first started doing GIMPS
> processing. Is this normal or has it detected an error in my results?
>
> (later...)
> I got "Self-test 256 passed!" so I guess I don't have any problems.
>
> cheers... Russ
This is normal, there are different self tests for different sizes of
numbers, you must have just gotten your first number in a higher range.
_________________________________________________________________
Unsubscribe & list info -- http://www.scruz.net/~luke/signup.htm
Mersenne Prime FAQ -- http://www.tasam.com/~lrwiman/FAQ-mers
------------------------------
Date: Tue, 25 Jan 2000 22:01:28 -0800
From: Kevin Sexton <[EMAIL PROTECTED]>
Subject: Mersenne: small OS setup for mprime
Did anyone figure out a simple way to run mprime from a
floppy, or small partition? I want to search for primes on a
second computer, but the total hard drive space is 650MB,
and I have and want to keep dos, and don't have windows of
any kind installed.
If I figure out how to install it I will add an LS-120
drive.
_________________________________________________________________
Unsubscribe & list info -- http://www.scruz.net/~luke/signup.htm
Mersenne Prime FAQ -- http://www.tasam.com/~lrwiman/FAQ-mers
------------------------------
Date: Wed, 26 Jan 2000 01:13:04 -0500
From: "David Campeau" <[EMAIL PROTECTED]>
Subject: Re: Mersenne: A Windows 2000 Question (Solution?)
Hi,
The behaviour you describe is expected and is the same with earlier version
of NT.
When the service is reporting status to the Service Control Manager, it as
to mention that it want to handle the shutdown event by specifying
SERVICE_ACCEPT_SHUTDOWN to the dwControlsAccepted member of SERVICE_STATUS
in a call to SetServiceStatus and then handle the SERVICE_CONTROL_SHUTDOWN
event in the handler for ServiceControl.
This should solve the problem:
change in service.c:
BOOL ReportStatusToSCMgr(DWORD dwCurrentState,
DWORD dwWin32ExitCode,
DWORD dwWaitHint)
{
static DWORD dwCheckPoint = 1;
BOOL fResult = TRUE;
if ( !bDebug ) // when debugging we don't report to the SCM
{
if (dwCurrentState == SERVICE_START_PENDING)
ssStatus.dwControlsAccepted = 0;
else
// Change here
// ssStatus.dwControlsAccepted = SERVICE_ACCEPT_STOP;
ssStatus.dwControlsAccepted = SERVICE_ACCEPT_STOP |
SERVICE_ACCEPT_SHUTDOWN;
// ....
// The rest is the same
}
AND
VOID WINAPI service_ctrl(DWORD dwCtrlCode)
{
// Handle the requested control code.
//
switch(dwCtrlCode)
{
// Stop the service.
//
// The folowing line is added
// Note: Only the system can send SERVICE_CONTROL_SHUTDOWN to
// a service, otherwise the function ControlService will fail.
// USE WITH CAUTION see note at the end.
case SERVICE_CONTROL_SHUTDOWN:
case SERVICE_CONTROL_STOP:
ssStatus.dwCurrentState = SERVICE_STOP_PENDING;
ServiceStop();
break;
// ....
// The rest is the same
}
David Campeau,
P.S: I sent a similar message to George Woltman about a month ago, but I
didn't receive any response. Perhaps there is something else I am
overlooking?
NOTE (taken from MSDN):
The SERVICE_CONTROL_SHUTDOWN control should only be processed by services
that must absolutely clean up during shutdown, because there is an extremely
limited time (about 20 seconds) available for service shutdown. After this
time expires, system shutdown proceeds regardless of whether service
shutdown is complete. If the service needs to take more time to shut down,
it should send out STOP_PENDING status messages, along with a wait hint, so
that the service controller knows how long to wait before reporting to the
system that service shutdown is complete. For example, the server service
needs to shut down so that network connections are not made when the system
is in the shutdown state.
_________________________________________________________________
Unsubscribe & list info -- http://www.scruz.net/~luke/signup.htm
Mersenne Prime FAQ -- http://www.tasam.com/~lrwiman/FAQ-mers
------------------------------
Date: Wed, 26 Jan 2000 17:45:16 +0100
From: Sylvain PEREZ <[EMAIL PROTECTED]>
Subject: RE: Mersenne: small OS setup for mprime
What I personnaly did is to install a full (but light) Windows95 system on
an IomegaZip 100 disquette, plus the regular Prime95 v19.
So now I only need a motherboard, an internal zip drive and some little
parts to make a Prime Machine. You have to set the BIOS to boot on Zip (or
LS120).
It works perfectly, even if it may be sometimes tricky to make a Zip a
system drive (and the only drive of the machine).
It's also easy for communicating with PrimeNet : shutdown your "Zip"
Windows, take the Zip disquette, put it on your machine with a modem or so,
actualize your Prime data and put the Zip disquette back to its machine,
boot. This can be done every 3 month or so.
Sylvain Perez
The guy who's late updating the french Gimps site ...
- -----Message d'origine-----
De: Kevin Sexton [SMTP:[EMAIL PROTECTED]]
Date: mercredi 26 janvier 2000 07:01
A: Mersenne List
Objet: Mersenne: small OS setup for mprime
Did anyone figure out a simple way to run mprime from a
floppy, or small partition? I want to search for primes on a
second computer, but the total hard drive space is 650MB,
and I have and want to keep dos, and don't have windows of
any kind installed.
If I figure out how to install it I will add an LS-120
drive.
_________________________________________________________________
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
------------------------------
End of Mersenne Digest V1 #684
******************************
