Mersenne Digest Sunday, October 17 1999 Volume 01 : Number 646
----------------------------------------------------------------------
Date: Sat, 16 Oct 1999 11:08:14 -0400 (EDT)
From: Darxus <[EMAIL PROTECTED]>
Subject: Re: Mersenne: Islands of Truth
On Fri, 15 Oct 1999, Jud McCranie wrote:
> At 05:45 PM 10/15/99 -0400, you wrote:
>
> >I've put a graph of these "pairs" up on my web page.
>
> You can't really tell much from that graph - most of the points are hugging
> the x-axis.
Most of those are pretty random... like, the higher your numbers get, the
more grouped they are.
I have more thoughts on this, from reading
http://www.utm.edu/research/primes/notes/faq/NextMersenne.html
a few times, but I gotta get going.
__________________________________________________________________
PGP fingerprint = 03 5B 9B A0 16 33 91 2F A5 77 BC EE 43 71 98 D4
[EMAIL PROTECTED] / http://www.op.net/~darxus
Join the Great Internet Mersenne Prime Search
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
------------------------------
Date: Sat, 16 Oct 1999 19:35:32 +0300
From: Jukka Santala <[EMAIL PROTECTED]>
Subject: Re: Mersenne: The Mysterious Ways of S.T.L.
"Brian J. Beesley" wrote:
> They will, and they're very likely to be artificial.
Umm. I'm moderately familiar with statistics, that's actually part of why I'm not
looking for significance in the patterns. (For one thing, the wole concept of
statistics kinda excludes the possibility of using it to "find the right one", like
some people seem to be assuming they can do. Kinda like the various countries hosting a
celebration for the birth of the 6 billionth person on Earth;) However, I was
suggesting to those actually looking for patterns in there not to automatically assume
they need to work with the set of all Mersenne exponents vs. exponents of prime
Mersennes. For one thing, this will very rarely yield a result that is even a candidate
(As some have found out). On the other hand, this could be a way to check the working
of the conjencture... We can artificially generate this situation, thouhg. However,
small correction: With Mersennes, the Mersenne numbers size is directly propotional to
the exponent or number of bits in the number, so obiviously there's not much sense in
testing the two separately ;)
-Donwulff
_________________________________________________________________
Unsubscribe & list info -- http://www.scruz.net/~luke/signup.htm
Mersenne Prime FAQ -- http://www.tasam.com/~lrwiman/FAQ-mers
------------------------------
Date: Sat, 16 Oct 1999 12:45:41 -0400 (EDT)
From: Darxus <[EMAIL PROTECTED]>
Subject: Re: Mersenne: Islands of Truth
On Fri, 15 Oct 1999, John R Pierce wrote:
> > Darxus wrote:
> >
> > > I've put a graph of these "pairs" up on my web page.
> > > http://www.op.net/~darxus/p_pairs.bmp
> >
> > Never use BMP on the WWW, please!
> > There's no way for me to look at that.
Interesting, what browser are you using ?
> here, I whacked it for you...
> http://hogranch.com/files/Bitmaps/p_pairs.gif
>
> since it was a very-few color chart, GIF was the logical format. BMP was
> 71k. GIF is 4K. nuff said?
I do not like .bmp because of microsoft and the fact that Jukka can't view
it (whatever the reason is).
I do not like .gif because of the copyright/patent thing on its
compression.
As far as I'm concerned, that only leaves .png, who's support is lacking.
Unfortunately, at this moment, I do not have the ability to do any
conversions, so thank you for putting up a .gif.
__________________________________________________________________
PGP fingerprint = 03 5B 9B A0 16 33 91 2F A5 77 BC EE 43 71 98 D4
[EMAIL PROTECTED] / http://www.op.net/~darxus
Join the Great Internet Mersenne Prime Search
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
------------------------------
Date: Sat, 16 Oct 1999 12:53:37 -0400 (EDT)
From: Darxus <[EMAIL PROTECTED]>
Subject: Re: Mersenne: Re: splitting up 10m digit primes
On Sat, 16 Oct 1999, Jukka Santala wrote:
> Except that PrimeNet doesn't control the prize. This is the error everybody is doing.
> EFF is adminstrating the competition and prize, given by anonymous donaters to
The EFF is requiring full disclosure. So if somebody used the GIMPS
client to find the 1st 10m digit prime, they'd have to say so, which would
then give GIMPS/Primenet the opportunity to point out that this was done
dishonestly.
I think you are all being rather silly about this whole thing. Progress
in the middle of checking the primality of a mersenne number looks to be
too cumbersone to keep track of centrally, and I believe a distributed
solution -- while I've contemplated such things and find them intreaguing
- -- would be more trouble than they'd be worth. Of course, I incourage you
to continue to ponder the problem, and have no objection to it being
discussed here. I just do not expect you to find a reasonable solution,
other than just letting it be, any time soon. And this is coming from a
person who is not often content to just let things be. But I think your
energies could be more effectively spent in othere areas.
I'd be significantly more interested in a client that had the
(optional) ability to upgrade itself.
__________________________________________________________________
PGP fingerprint = 03 5B 9B A0 16 33 91 2F A5 77 BC EE 43 71 98 D4
[EMAIL PROTECTED] / http://www.op.net/~darxus
Join the Great Internet Mersenne Prime Search
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
------------------------------
Date: Sat, 16 Oct 1999 13:07:06 -0400 (EDT)
From: Darxus <[EMAIL PROTECTED]>
Subject: Re: Mersenne: Islands of Truth
On Sat, 16 Oct 1999, Darxus wrote:
> On Fri, 15 Oct 1999, Jud McCranie wrote:
>
> > At 05:45 PM 10/15/99 -0400, you wrote:
> >
> > >I've put a graph of these "pairs" up on my web page.
> >
> > You can't really tell much from that graph - most of the points are hugging
> > the x-axis.
>
> Most of those are pretty random... like, the higher your numbers get, the
> more grouped they are.
>
> I have more thoughts on this, from reading
> http://www.utm.edu/research/primes/notes/faq/NextMersenne.html
> a few times, but I gotta get going.
Okay, that page is a bit overwhelming for me, but it looks like it's
saying that, when you graph log(log(p)) vs. n, where the logs are base 2,
p is part of 2^p-1, and n is the number of the mersenne prime (did I say
that right?), you'll get an approximation of a straight line. All
mersenne primes will fit that pattern, but they are less likely to be *on*
the line than they are to be *near* the line.
Geez, I'm having difficulty putting this into words.
Like... primes are not likely to fall extreemly close to the line,
but they're probably going to be near it -- like, they're magnetically
repelled from the line. This would cause random clumping.
It would also cause random clumping when you graph P vs. N (where P is
part of 2^P-1, and N is the number of the merseene prime).
I have no idea what would cause this, but if it's correct, we should be
able to calculate the probability of the distance from the line (if you're
doing the log log thing, or if you're fitting an exponential curve), and
predict more accurately where a prime would fall (you'd be predicting 2
points, one on either side of the line).
Gonna go regression test (proper term?) the extrapolations I've been
playing with assuming the distribution is totally random but approximating
an exponential curve....
__________________________________________________________________
PGP fingerprint = 03 5B 9B A0 16 33 91 2F A5 77 BC EE 43 71 98 D4
[EMAIL PROTECTED] / http://www.op.net/~darxus
Join the Great Internet Mersenne Prime Search
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
------------------------------
Date: Sat, 16 Oct 1999 13:37:20 -0400
From: Jud McCranie <[EMAIL PROTECTED]>
Subject: Re: Mersenne: Islands of Truth
At 01:07 PM 10/16/99 -0400, Darxus wrote:
> Like... primes are not likely to fall extreemly close to the line,
>but they're probably going to be near it -- like, they're magnetically
>repelled from the line.
No, that isn't the case.
>I have no idea what would cause this, but if it's correct, we should be
>able to calculate the probability of the distance from the line (if you're
>doing the log log thing, or if you're fitting an exponential curve), and
>predict more accurately where a prime would fall (you'd be predicting 2
>points, one on either side of the line).
That won't work either.
+---------------------------------------------------------+
| Jud McCranie |
| |
| Programming Achieved with Structure, Clarity, And Logic |
+---------------------------------------------------------+
_________________________________________________________________
Unsubscribe & list info -- http://www.scruz.net/~luke/signup.htm
Mersenne Prime FAQ -- http://www.tasam.com/~lrwiman/FAQ-mers
------------------------------
Date: Sat, 16 Oct 1999 18:40:22 GMT
From: [EMAIL PROTECTED] (Michael Oates)
Subject: Re: Mersenne: problem with prme95 - spl file - ME TOO !
Brian,
This will probably make you laugh, if it does don't tell me, I was not
ammused when I found out what had happened, read on...
>On 15 Oct 99, at 13:16, Michael Oates wrote:
>
>> I am also having the same problem, but only on one machine, I have 9 others
>> that are fine, all are using the same version of v19 beta 4
>
>What's different? Must be _something_ ...
I have been running Prime95 on this machine for about 4 weeks. It is on a
LAN with other machines also running Prime95.
I have a shortcut to run the program in the Startup folder.
Today I realised that not only was the server not being updated, but the
save files had not been altered for over a week. This was very odd as the
program was re-starting correctly where it left off. So I thought, it must
be storing the save file somewhere else. Then I also noticed that some of
the setting details were wrong, not just wrong, but were for a different
machine. I ran the program from the directory with File Explorer and it used
a different exponent and was at a different stage ! Eh !!!!!!!
I checked the shortcut that was in the Startup folder... well would you
believe it... it was pointing to another machine.
Some how Windows95 had changed the path to the shortcut all on it's own, and
had been running another exponent on another machine over the network. And
for some odd reason the details stored in local.ini were taken from the
local machine and the details from prime.ini were from the remote.
Now I know what you are probably thinking, it was not Windows95 that changed
the path, and that I made a mistake when I set it up... WRONG, very wrong. I
have seen this before... if you have a shortcut to a file, and remove the
file, and replace it with another one Windows95 inserts another path in any
shortcut to that program, this has cause no end of problems in the past.
This probably happend when I upgraded Prime95 with a newer verion, I moved
the directory else where by mistake, instead of copying it to make a backup.
I of course just copied it back again straight away, then copied the new
updated program to it... from... wait for it... the machine that the new
shortcut was pointing to. So somehow windows changed the shortcut path to
the directory on the remote machine, great eh !
All is now working ok, but I have lost over a weeks LL testing because of
it. No harm seems to have been done to the remote machine.
Regards,
Mike,
- --
ATLAS CELESTE - Bevis Star Atlas - & "The CD-ROM"
A very rare atlas found at the Godlee Observatory
http://www.u-net.com/ph/mas/bevis/
Astronomy in the UK http://www.ph.u-net.com
_________________________________________________________________
Unsubscribe & list info -- http://www.scruz.net/~luke/signup.htm
Mersenne Prime FAQ -- http://www.tasam.com/~lrwiman/FAQ-mers
------------------------------
Date: Sat, 16 Oct 1999 14:49:42 -0400
From: George Woltman <[EMAIL PROTECTED]>
Subject: Mersenne: AMD K6 and Athlon
Hi all,
As you now, prime95 v19 has two different code paths. One optimized for
Pentiums the other for Pentium Pro and later CPUs.
A user reports that the AMD K6 is faster using the Pentium code, but
v19 selects the Pentium Pro code instead. I will fix this, but in the
meantime K6 owners should change their CPU type to Pentium.
If there are any Athlon owners out there, can you let me know which
code path is faster for your CPU?
Thanks,
George
_________________________________________________________________
Unsubscribe & list info -- http://www.scruz.net/~luke/signup.htm
Mersenne Prime FAQ -- http://www.tasam.com/~lrwiman/FAQ-mers
------------------------------
Date: Sat, 16 Oct 1999 15:47:42 -0400
From: Jud McCranie <[EMAIL PROTECTED]>
Subject: Re: Mersenne: AMD K6 and Athlon
At 02:49 PM 10/16/99 -0400, George Woltman wrote:
>v19 selects the Pentium Pro code instead.
I tried version 19 on a PII and a Celeron, and in both cases it thought
they were P-Pros. It got the MHz correct.
+---------------------------------------------------------+
| Jud McCranie |
| |
| Programming Achieved with Structure, Clarity, And Logic |
+---------------------------------------------------------+
_________________________________________________________________
Unsubscribe & list info -- http://www.scruz.net/~luke/signup.htm
Mersenne Prime FAQ -- http://www.tasam.com/~lrwiman/FAQ-mers
------------------------------
Date: Sun, 17 Oct 1999 02:34:11 +0200 (CEST)
From: Henrik Olsen <[EMAIL PROTECTED]>
Subject: Re: Mersenne: problem with prme95 - spl file - ME TOO !
On Sat, 16 Oct 1999, Michael Oates wrote:
> I have been running Prime95 on this machine for about 4 weeks. It is on a
> LAN with other machines also running Prime95.
>
> I have a shortcut to run the program in the Startup folder.
For Windows 95 and prime95, this is the wrong approach, since as the
included story shows, this makes you vulnerable to the mangling of
shortcuts that Windows loves and we hate.
Instead, remove the shortcut from the startup folder, start
prime95 manually, choose "Options" and make sure "Windows 98/95 Service"
is checked, then it will start automagically, this is especially
important in networked setups, as the service runs even when noone's
logged in.
> Today I realised that not only was the server not being updated, but the
> save files had not been altered for over a week. This was very odd as the
> program was re-starting correctly where it left off. So I thought, it must
> be storing the save file somewhere else. Then I also noticed that some of
> the setting details were wrong, not just wrong, but were for a different
> machine. I ran the program from the directory with File Explorer and it used
> a different exponent and was at a different stage ! Eh !!!!!!!
>
> I checked the shortcut that was in the Startup folder... well would you
> believe it... it was pointing to another machine.
>
> Some how Windows95 had changed the path to the shortcut all on it's own, and
> had been running another exponent on another machine over the network. And
> for some odd reason the details stored in local.ini were taken from the
> local machine and the details from prime.ini were from the remote.
>
> Now I know what you are probably thinking, it was not Windows95 that changed
> the path, and that I made a mistake when I set it up... WRONG, very wrong. I
> have seen this before... if you have a shortcut to a file, and remove the
> file, and replace it with another one Windows95 inserts another path in any
> shortcut to that program, this has cause no end of problems in the past.
>
> This probably happend when I upgraded Prime95 with a newer verion, I moved
> the directory else where by mistake, instead of copying it to make a backup.
> I of course just copied it back again straight away, then copied the new
> updated program to it... from... wait for it... the machine that the new
> shortcut was pointing to. So somehow windows changed the shortcut path to
> the directory on the remote machine, great eh !
>
> All is now working ok, but I have lost over a weeks LL testing because of
> it. No harm seems to have been done to the remote machine.
>
> Regards,
>
> Mike,
>
> --
> ATLAS CELESTE - Bevis Star Atlas - & "The CD-ROM"
> A very rare atlas found at the Godlee Observatory
> http://www.u-net.com/ph/mas/bevis/
> Astronomy in the UK http://www.ph.u-net.com
>
> _________________________________________________________________
> 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/
Catherine: Go to hell!
Gabriel: Heaven, heaven. At least get the zip code right. The Prophecy
_________________________________________________________________
Unsubscribe & list info -- http://www.scruz.net/~luke/signup.htm
Mersenne Prime FAQ -- http://www.tasam.com/~lrwiman/FAQ-mers
------------------------------
Date: Sat, 16 Oct 1999 20:50:20 EDT
From: [EMAIL PROTECTED]
Subject: Mersenne: Bold Predictions From STL's Mysterious Ways
(I'm quite poor at choosing subjects, you see.)
Well, I found my notecard of predictions that I had calculated a while ago
with my conjecture. Here are some of the values I computed. These can either
be used for a good laugh (M100 in particular is nice to look at), or you can
write these values down and see how close they hit the mark (live long and
find Mersennes?). At the time I only had data for Mersenne prime exponents up
to 3021377 (i.e. not 6972593), and I calculated what M37 should have been as
a test. It's pretty close. My prediction for the 6M prime was also close.
Now, to find the missing Mersenne....
Here I call M# to be q instead of 2^q-1 for brevity.
M37 (known at the time to actually be 3021377): 3166795
M38: 4673434 (the elusive missing Mersenne?!? If there is no Mersenne prime
around here, then the 3021377-6972593 gap is almost as large as the 127-521
gap!)
M39: 6896873 (There is a prime at 6972593 in this region. But is it M38 or
M39?)
M40: 10178139
M41: 15020505
M42: 22166682
M43: 32712733
M44: 48276189 (hence my conjecture that the decamegaprime is either right at
the 10M digit limit, or has an exponent around 48M)
M50: 498689073
M75: 8379797036294
M100 ~ 140811183105000000 (that's its EXPONENT, whoo hoo. The inaccuracy is
due to the precision limitations of my calculator)
As for myself, I'm *really* hoping that there's a missing Mersenne. Time will
tell....
- -*---*-------
S.T.L.
_________________________________________________________________
Unsubscribe & list info -- http://www.scruz.net/~luke/signup.htm
Mersenne Prime FAQ -- http://www.tasam.com/~lrwiman/FAQ-mers
------------------------------
Date: Sun, 17 Oct 1999 02:22:58 +0200
From: "Steinar H. Gunderson" <[EMAIL PROTECTED]>
Subject: Mersenne: Re: AMD K6 and Athlon
On Sat, Oct 16, 1999 at 03:47:42PM -0400, Jud McCranie wrote:
>I tried version 19 on a PII and a Celeron, and in both cases it thought
>they were P-Pros. It got the MHz correct.
Were these upgraded, or fresh installs? The GIMPS software (still no
collective name for Prime95/Saver95/mprime/ntprime?) won't try a
redetection if there already is a CPU type entry available.
/* 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: Sun, 17 Oct 1999 02:21:26 +0200
From: "Steinar H. Gunderson" <[EMAIL PROTECTED]>
Subject: Mersenne: Re: problem with prme95 - spl file - ME TOO !
On Sat, Oct 16, 1999 at 06:40:22PM +0000, Michael Oates wrote:
>This will probably make you laugh, if it does don't tell me, I was not
>ammused when I found out what had happened, read on...
This story reminds me of something!
After installing Office 97 on my own PC (this was at the time I was using
a NIC Windows actually could use, and at the time I used Windows a lot),
suddenly Word 6.0 refused to run on my mother's computer.
Reason: Windows has found Word 6.0 (over the network) and found it to be
`obsolete', and removed it immediately.
Just shows that Microsoft can't keep their local and remote disks apart :-)
(Not that Linux does any better, but at least it doesn't assume you want
lots of stuff which you don't! :-) )
/* 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: Sun, 17 Oct 1999 02:07:44 +0200
From: "Steinar H. Gunderson" <[EMAIL PROTECTED]>
Subject: Mersenne: Re: Reliability (was Re: splitting up 10m digit primes) + Possible
Wish List item
On Sat, Oct 16, 1999 at 06:27:24AM +0300, Jukka Santala wrote:
>If you can afford the bandwidth and storage
>space, you can check the box.
The problem probably doesn't lie at the user's end. The server is the one
with the storage problem, and probably the one with the bandwidth problem
as well...
/* 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: Sun, 17 Oct 1999 02:13:33 +0200
From: "Steinar H. Gunderson" <[EMAIL PROTECTED]>
Subject: Mersenne: Re: splitting up 10m digit primes
On Sat, Oct 16, 1999 at 03:46:30PM +0100, Chris Jefferson wrote:
>However, proving that they has used the
>interim file generated by this particular program might be difficult...
Not very. Save files are incompatible from program to program. Or at least
so I think...
Converting a savefile is also going to be a lot of work :-)
/* 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: Sat, 16 Oct 1999 18:13:17 -0700
From: Spike Jones <[EMAIL PROTECTED]>
Subject: Mersenne: leaf gnawing caterpillar graphic
When I suggested the caterpillar graphic for the GIMPS front
end, I should have mentioned somewhere that my own coding
skills are quite modest. Should some coder guru in GIMPS land
created such a graphic, I would gladly pay 25 bucks for a copy.
If 10% of the GIMPSers were to pay 25 bucks each, someone
would make 20k. Just a suggestion. {8-] spike
_________________________________________________________________
Unsubscribe & list info -- http://www.scruz.net/~luke/signup.htm
Mersenne Prime FAQ -- http://www.tasam.com/~lrwiman/FAQ-mers
------------------------------
Date: Sat, 16 Oct 1999 22:30:25 -0400 (EDT)
From: Darxus <[EMAIL PROTECTED]>
Subject: Re: Mersenne: Bold Predictions From STL's Mysterious Ways
STL: what was the line you fit to the log log of the known mersenne's
exponents, and what was the error (r^2) ?
I'm assuming r^2 is like, the amount of error for a given line fitting a
set of data.
When I fit an exponential line to the 1st 37 mersenne prime's exponents
(since I believe that 6972593 is actually the 39th, but have no proof, so
I figured I'd leave it out), the line I got was y=1.7661e^0.301x (hope I
wrote that right), and r^2=0.9925.
I have a feeling the log log thing will actually be more useful.. easier
to predict a straight line than an exponential curve.
__________________________________________________________________
PGP fingerprint = 03 5B 9B A0 16 33 91 2F A5 77 BC EE 43 71 98 D4
[EMAIL PROTECTED] / http://www.op.net/~darxus
Join the Great Internet Mersenne Prime Search
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
------------------------------
Date: Sat, 16 Oct 1999 22:45:17 -0400
From: George Woltman <[EMAIL PROTECTED]>
Subject: Re: Mersenne: Re: AMD K6 and Athlon
Hi,
At 02:22 AM 10/17/99 +0200, Steinar H. Gunderson wrote:
>On Sat, Oct 16, 1999 at 03:47:42PM -0400, Jud McCranie wrote:
>>I tried version 19 on a PII and a Celeron, and in both cases it thought
>>they were P-Pros. It got the MHz correct.
>
>Were these upgraded, or fresh installs? The GIMPS software (still no
>collective name for Prime95/Saver95/mprime/ntprime?) won't try a
>redetection if there already is a CPU type entry available.
Prime95 does not detect Celerons very well. However, if a PIII, PII or
Celeron is detected as a Pentium Pro - its no big deal. All those CPU
types use the same code.
Regards,
George
_________________________________________________________________
Unsubscribe & list info -- http://www.scruz.net/~luke/signup.htm
Mersenne Prime FAQ -- http://www.tasam.com/~lrwiman/FAQ-mers
------------------------------
Date: Sat, 16 Oct 1999 23:14:16 -0400 (EDT)
From: Darxus <[EMAIL PROTECTED]>
Subject: Re: Mersenne: The Mysterious Ways of S.T.L.
On Fri, 15 Oct 1999 [EMAIL PROTECTED] wrote:
> Then, I plotted e^gamma log[2] (mersenne) versus the list of 1-37. Alongside
> this I graphed y=x. This is because the y=x line represents the Wagstaff
y=x would be a slope of 1/1.
According to the "Where is the next larger Mersenne prime?" page --
http://www.utm.edu/research/primes/notes/faq/NextMersenne.html the
Wagstaff conjecture suggests a slope of 3/2, which I believe wouldn't look
so bad.
> So, I graphed e^gamma log [2] (mersenne) - (1, 2, 3, 4, etc). This represents
> how far off the Wagstaff conjecture is when applied to the data. (The
> Wagstaff conjecture *should* say that M(3021377) = 37, but it doesn't. This
> is why I graph this jibberish). This graph was INCREDIBLY disturbing. Save
> for one Mersenne prime, all these "errors" were above 0, and often big. Ech!
> So, I used my TI-92+ to take a linear regession line of this data (because I
> had recently learned how to do regression lines and correlation
> coefficients). This line was Y = .004769x + 1.4615. See what's happening
> here? It seems that there's a consistent error (1.4615) in the Wagstaff
> conjecture that doesn't change as the Mersenne primes grow (the .004769). So
> I went back and applied this correction to the graph "that seemed a little
> strange" and it fit y=x much better.
Sorry it didn't register to me that you'd mentioned the equation for this
line in this post, thanks. But what was r^2 for it ? I'm very curious.
On the previously mentioned web page, there are similar computations, but
I believe he used M38 (which you and I believe will actually turn out to
be M39), so I believe his numbers will be less accurate than yours.
I would really like to try your calculations myself, but I haven't seen my
graphing calculator for a while, I'm not sure it'd work, and I'd prefer to
use the power of my computer. Can anybody suggest any programs ?
Preferably for Linux, even though that would mean I'd have to wait to get
my Linux drive back.
I am most anxious to take M1-M36 & extrapolate M37, down through having
just, say, M1-M10 & extrapolating M11 & see how accurate that process is.
I really thing the GIMPS client should have an option to test the mersenne
number closest to this estimate of M38 that has not been and is not being
tested -- implemented in a scalable fasion, so that when/if the missing
prime is found, and it finishes the number it's on (since knowing what
isn't prime is valuble), it'll then go to the next estimated prime.
__________________________________________________________________
PGP fingerprint = 03 5B 9B A0 16 33 91 2F A5 77 BC EE 43 71 98 D4
[EMAIL PROTECTED] / http://www.op.net/~darxus
Join the Great Internet Mersenne Prime Search
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
------------------------------
Date: Sun, 17 Oct 1999 03:09:06 -0400 (EDT)
From: Lucas Wiman <[EMAIL PROTECTED]>
Subject: Re: Mersenne: Bold Predictions From STL's Mysterious Ways
> STL: what was the line you fit to the log log of the known mersenne's
> exponents, and what was the error (r^2) ?
>
> I'm assuming r^2 is like, the amount of error for a given line fitting a
> set of data.
>
> When I fit an exponential line to the 1st 37 mersenne prime's exponents
> (since I believe that 6972593 is actually the 39th, but have no proof, so
> I figured I'd leave it out), the line I got was y=1.7661e^0.301x (hope I
> wrote that right), and r^2=0.9925.
>
> I have a feeling the log log thing will actually be more useful.. easier
> to predict a straight line than an exponential curve.
They are actually equally easy to predict. A straight line is just a log
of an exponential curve.
- -Lucas
_________________________________________________________________
Unsubscribe & list info -- http://www.scruz.net/~luke/signup.htm
Mersenne Prime FAQ -- http://www.tasam.com/~lrwiman/FAQ-mers
------------------------------
Date: Sun, 17 Oct 1999 03:13:28 -0400 (EDT)
From: Lucas Wiman <[EMAIL PROTECTED]>
Subject: Re: Mersenne: The Mysterious Ways of S.T.L.
> According to the "Where is the next larger Mersenne prime?" page --
> http://www.utm.edu/research/primes/notes/faq/NextMersenne.html the
> Wagstaff conjecture suggests a slope of 3/2, which I believe wouldn't look
> so bad.
No. Wagstaff's conjecture predicts a slope of around 1.47, (2^exp(-Gamma)).
> I would really like to try your calculations myself, but I haven't seen my
> graphing calculator for a while, I'm not sure it'd work, and I'd prefer to
> use the power of my computer. Can anybody suggest any programs ?
> Preferably for Linux, even though that would mean I'd have to wait to get
> my Linux drive back.
GNU plot (I've never had a need to use it, but I've heard good things about it).
I know that it comes with Redhat and Debian.
- -Lucas
_________________________________________________________________
Unsubscribe & list info -- http://www.scruz.net/~luke/signup.htm
Mersenne Prime FAQ -- http://www.tasam.com/~lrwiman/FAQ-mers
------------------------------
Date: Sun, 17 Oct 1999 08:22:37 +0100
From: Nick Craig-Wood <[EMAIL PROTECTED]>
Subject: Re: Mersenne: Re: splitting up 10m digit primes
On Sun, Oct 17, 1999 at 02:13:33AM +0200, Steinar H. Gunderson wrote:
> On Sat, Oct 16, 1999 at 03:46:30PM +0100, Chris Jefferson wrote:
> >However, proving that they has used the
> >interim file generated by this particular program might be difficult...
>
> Not very. Save files are incompatible from program to program. Or at least
> so I think...
Some time ago I proposed a standard format for save files. It also
had the advantage that it only saved the actual bits used so would
make the save files 1MByte rather than 1.8 MByte for an 8,000,000
exponent in prime95.
> Converting a savefile is also going to be a lot of work :-)
It needs to be done exactly right - hence by the program which
generated it probably.
Here is what I wrote :-
....
Reading in other threads about save files it seems to me that it would be
worth while standardising the format of save files so that they can be ported
across platforms.
At the moment I think both Prime95 and the unix programs write the floating
point values of the FFT array to a file which is inherently non-portable
across machine architectures. It also makes the save files quite a bit
larger than necessary (an important consideration if we ever want to hold a
central repository of these files).
I'd like to propose for discussion a simple cross-platform save file format
as below.
Width Value Description
8 "MersSave" in ASCII as a file identifier
4 int32 version number of file
16 ASCII space filled name of creating program
4 int32 exponent being tested (n)
4 int32 iteration that this residue represents
4 int32 length of platform specific area x words
4x int32 platform specific
len int32 binary data for the residue
len = ((n + 0x1F) & ~ 0x1F) >> 3
4 int32 CCITT CRC-32 of the entire file
All integers should be represented as little-endian in the file (the majority
of the clients are little-endian). The residue should be little endian with
the unused 1-31 bits at the end set to 0.
A file format like this would have the following advantages
1) You could take a save file from one platform/program to another (via a
central repository at primenet maybe)
2) You could take a save file from one FFT size to another (imagine a
software upgrade where the FFT limits have changed)
3) They are much smaller (eg 600k vs 2M)
It would have the disadvantages
1) Slightly slower and more complex to read and write. However this is an
infrequent operation. (That said time to write a save file may be IO
limited in which case a shorter save file will be quicker.)
2) We've all got to agree on it ;-)
Discuss!
- --
Nick Craig-Wood
[EMAIL PROTECTED]
http://www.axis.demon.co.uk/
_________________________________________________________________
Unsubscribe & list info -- http://www.scruz.net/~luke/signup.htm
Mersenne Prime FAQ -- http://www.tasam.com/~lrwiman/FAQ-mers
------------------------------
Date: Sun, 17 Oct 1999 15:21:49 +0100
From: "Brian J. Beesley" <[EMAIL PROTECTED]>
Subject: Re: Mersenne: Re: splitting up 10m digit primes
On 17 Oct 99, at 8:22, Nick Craig-Wood wrote:
> Some time ago I proposed a standard format for save files. It also
> had the advantage that it only saved the actual bits used so would
> make the save files 1MByte rather than 1.8 MByte for an 8,000,000
> exponent in prime95.
Yes - this idea is useful ...
> > Converting a savefile is also going to be a lot of work :-)
>
> It needs to be done exactly right - hence by the program which
> generated it probably.
>
If it's worthwhile, there is of course source for George's programs
(which represent the vast majority of clients). So there's existing
source code for reading Prime95/mprime format save files.
>
> Here is what I wrote :-
>
> I'd like to propose for discussion a simple cross-platform save file format
> as below.
>
> Width Value Description
>
> 8 "MersSave" in ASCII as a file identifier
> 4 int32 version number of file
> 16 ASCII space filled name of creating program
I want this longer, & to include version major/minor/release & build
numbers as well. 32 bytes?
> 4 int32 exponent being tested (n)
> 4 int32 iteration that this residue represents
> 4 int32 length of platform specific area x words
> 4x int32 platform specific
e.g. George's programs would want to store the offset here.
Would it be too much to ask to have an int32 count of "platform
specific" bytes following in a variable length field here? Some
platforms will require none (just a zero count), others might want to
embed a (currently unknown) vector of arbitary length at this point.
> len int32 binary data for the residue
> len = ((n + 0x1F) & ~ 0x1F) >> 3
> 4 int32 CCITT CRC-32 of the entire file
>
> All integers should be represented as little-endian in the file (the majority
> of the clients are little-endian). The residue should be little endian with
> the unused 1-31 bits at the end set to 0.
Since the "majority of clients" which are "little endian" are
overwhelmingly IA32 processors, which have byte reversal instructions
specifically designed to do the job, I'd like to propose that the
integers be stored in "big-endian" format. Or, alternatively, that we
devote _two_ bits of the file version number - both the MS and LS
bits - to flags, set to 0 for little-endian format or 1 for big-
endian format. Still leaves 30 bits for file version number info, and
allows clients to write whatever format is easiest. Write operations
outnumber read operations by a big margin!
Regards
Brian Beesley
_________________________________________________________________
Unsubscribe & list info -- http://www.scruz.net/~luke/signup.htm
Mersenne Prime FAQ -- http://www.tasam.com/~lrwiman/FAQ-mers
------------------------------
Date: Sun, 17 Oct 1999 20:49:39 +0100
From: Nick Craig-Wood <[EMAIL PROTECTED]>
Subject: Re: Mersenne: Re: splitting up 10m digit primes
On Sun, Oct 17, 1999 at 03:21:49PM +0100, Brian J. Beesley wrote:
> On 17 Oct 99, at 8:22, Nick Craig-Wood wrote:
> > I'd like to propose for discussion a simple cross-platform save file format
> > as below.
> >
> > Width Value Description
> >
> > 8 "MersSave" in ASCII as a file identifier
> > 4 int32 version number of file
> > 16 ASCII space filled name of creating program
>
> I want this longer, & to include version major/minor/release & build
> numbers as well. 32 bytes?
Yes sounds like a good idea. Perhaps we should formalise the version
number like this?
1 int8 Major
1 int8 Minor
1 int8 Micro
1 int8 Patch level
> > 4 int32 exponent being tested (n)
> > 4 int32 iteration that this residue represents
> > 4 int32 length of platform specific area x words
> > 4x int32 platform specific
>
> e.g. George's programs would want to store the offset here.
Ah, offset! I wrote this before offset was introduced. This would
need to be part of the save file I think because different offset =>
different save file data. I think offset would fit into an int32?
> Would it be too much to ask to have an int32 count of "platform
> specific" bytes following in a variable length field here? Some
> platforms will require none (just a zero count), others might want to
> embed a (currently unknown) vector of arbitary length at this point.
That was what I intended by the last two entries above, but I may not
have been clear enough.
4 int32 length of platform specific area x words
4x int32 platform specific (x words of data may be 0 length)
> > len int32 binary data for the residue
> > len = ((n + 0x1F) & ~ 0x1F) >> 3
> > 4 int32 CCITT CRC-32 of the entire file
> >
> > All integers should be represented as little-endian in the file (the majority
> > of the clients are little-endian). The residue should be little endian with
> > the unused 1-31 bits at the end set to 0.
>
> Since the "majority of clients" which are "little endian" are
> overwhelmingly IA32 processors, which have byte reversal instructions
> specifically designed to do the job, I'd like to propose that the
> integers be stored in "big-endian" format. Or, alternatively, that we
> devote _two_ bits of the file version number - both the MS and LS
> bits - to flags, set to 0 for little-endian format or 1 for big-
> endian format. Still leaves 30 bits for file version number info, and
> allows clients to write whatever format is easiest. Write operations
> outnumber read operations by a big margin!
This is sensible. I like the big/little endian flag idea. I've tried
to design the header so that it can be represented by a C struct and
just (f)write()-en straight to the file. Having big/little endian
flag is an excellent idea here.
Also it might just be better to store the actual numeric data as
int8's rather than int32s - this way the endian ness of the processors
doesn't matter.
- --
Nick Craig-Wood
[EMAIL PROTECTED]
http://www.axis.demon.co.uk/
_________________________________________________________________
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 #646
******************************
