Mersenne Digest Sunday, 21 February 1999 Volume 01 : Number 512
----------------------------------------------------------------------
From: "Aaron Blosser" <[EMAIL PROTECTED]>
Date: Fri, 19 Feb 1999 08:15:37 -0700
Subject: RE: Mersenne: Problem Running NTPrime
> Curtis Cooper
> Sent: Thursday, February 18, 1999 7:04 PM
> To: [EMAIL PROTECTED]
> Subject: Mersenne: Problem Running NTPrime
> I am trying to run NTPrime Version 17.1 as a service
> under Windows NT 4.0. I have been having trouble
> keeping it running; it seems like the System Idle
> Process will take control of the CPU. Things
> will run fine for a day or two; NTPrime will
> start automatically and when I check the Task Manager
> it is using 99% of the CPU. But then I will come back a day or
> so later and sometime during that time period the
> System Idle Process took control and is running
> at 99% CPU usage and NTPrime has 0%.
<snip>
> Is there an explanation why the System Idle Process is taking
> control of the CPU?
The NT "idle" process is not so much a process as just a measurement of how
much your computer is *not* doing. The fact that NTPrime didn't seem to be
doing anything means that perhaps it hung, or maybe it didn't have any work
to do. Next time this happens, check the "worktodo.ini" file and see if
there's anything in there maybe.
It's quite normal to see the "idle" process with lots of CPU "usage" when no
programs are doing anything.
________________________________________________________________
Unsubscribe & list info -- http://www.scruz.net/~luke/signup.htm
------------------------------
From: [EMAIL PROTECTED]
Date: Fri, 19 Feb 1999 18:25:11 EST
Subject: Mersenne: Re: fascist symbolism
In a message dated 19/02/99 09:34:44, you write:
<< Subject: RE: Mersenne: Link from Knuth's Home Page
Note that the Fascist party in Italy was called this because its
symbol had a bundle of something (wheat? twigs?) in the middle, and
the Italian for bundle is "fascio".
>>
In ancient Roman times, it was a bundle of sticks carried by an official
called the "lictor" to symbolise his power to beat miscreants. The
modern Italian fascists added an axe in the centre of the bundle, to
symbolise capital punishment by beheading. This was not present
in the ancient Roman version. Or so I learnt in Latin classes, anyway.
Stanford University at one time had an outstation near Florence, Italy,
where there was a swimming pool with this fascist emblem on the
bottom. While there I took a bath in a tub (large) which had been
used by Mussolini, in addition to swimming in the pool.
I hope this puts an end to this off-topic thread.
George.
________________________________________________________________
Unsubscribe & list info -- http://www.scruz.net/~luke/signup.htm
------------------------------
From: Bryan Fullerton <[EMAIL PROTECTED]>
Date: Fri, 19 Feb 1999 19:55:25 -0500
Subject: Mersenne: Factoring...
Howdy,
I'm running PrimeNet on my 486sx/25 Linux router box (yeah, I know...). Right
now it's factoring M8956237, and it seems to do about one pass per day. It's
going to take a while. :)
Would adding a 487 math co-processor speed this up? If so, by how much? I
have no idea about the algorythm used, so I don't know if this is primarily
doing integer or floating point operations.
Thanks,
Bryan
- --
Bryan Fullerton http://www.samurai.com/
Owner, Lead Consultant http://www.feh.net/
Samurai Consulting http://www.icomm.ca/
"No, we don't do seppuku." Can you feel the Ohmu call?
________________________________________________________________
Unsubscribe & list info -- http://www.scruz.net/~luke/signup.htm
------------------------------
From: "John R Pierce" <[EMAIL PROTECTED]>
Date: Sat, 20 Feb 1999 01:11:24 -0800
Subject: Mersenne: phew. bigger and bigger
Just got my first primenet assignment over 7 million yesterday. phew.
estimated 65 days on a pentium 166mmx system (ok, so this is one of my junk
boxes that happened to get it).
- -jrp
________________________________________________________________
Unsubscribe & list info -- http://www.scruz.net/~luke/signup.htm
------------------------------
From: "Aaron Blosser" <[EMAIL PROTECTED]>
Date: Sat, 20 Feb 1999 08:58:35 -0700
Subject: RE: Mersenne: phew. bigger and bigger
> Just got my first primenet assignment over 7 million yesterday. phew.
> estimated 65 days on a pentium 166mmx system (ok, so this is one
> of my junk
> boxes that happened to get it).
I've got a couple of those myself.
Scott Kurowski: Didn't you say that you had planned to increase the
requirements for a machine to get an LL test assignment?
For the 7M+ exponents, would it be wise to raise the bar again to maybe a
200MHz machine or even higher?
Meanwhile, in other news, anyone spot my name in the March PC Magazine's
"Abort, Retry, Fail?" section (last page)? The previous mention some months
ago prompted me to write the author, asking him to correct some things. He
said he would and he did. He was actually pretty cool about the whole
thing, so I'm pleased.
Also, if anyone notices a bunch of new machines joining GIMPS in the next 6
months, don't hit the FBI speed dial button on your phone. I'll be getting
about 70 machines that I can do whatever I want with for a short time, plus
a few RISC boxes.
Speaking of RISC, I haven't checked, but is there good software for LL
testing on Sun Sparcstations and IBM RS/6000's?
Aaron
________________________________________________________________
Unsubscribe & list info -- http://www.scruz.net/~luke/signup.htm
------------------------------
From: Marc Getty <[EMAIL PROTECTED]>
Date: Sat, 20 Feb 1999 12:17:58 -0500
Subject: Mersenne: Factoring on a 486SX
> I'm running PrimeNet on my 486sx/25 Linux router box (yeah, I know...). Right
> now it's factoring M8956237, and it seems to do about one pass per day. It's
> going to take a while. :)
> Would adding a 487 math co-processor speed this up? If so, by how much? I
> have no idea about the algorythm used, so I don't know if this is primarily
> doing integer or floating point operations.
The "487 math co-processer" is actually a 486DX chip. Unlike previous
Intel CPUs where you just add an external math co-processer, the 486
SX line was an afterthought. I have never seen a 487, all I have ever
seen is 486DX's. They may exist, but in all likelihood they are just
a re-marked 486DX processor.
Assuming your CPU is not a soldered on surface mount CPU, meaning
an early socket design or a ZIF socket, you should swap it out with
something a little more speedy. If you want a free 486DX33 chip, I'll
send you one if you send me a SASE, but I would recommend going out
and paying a good $15-20 for a 486DX2-66 as long as it's an old Intel
5.0 volt model.
Marc Getty - [EMAIL PROTECTED] - [EMAIL PROTECTED] - ICQ: 12916278
http://www.getty.net - http://www.vwthing.org Work: 215-204-3291
http://etc.temple.edu/ Home: 215-322-8363
________________________________________________________________
Unsubscribe & list info -- http://www.scruz.net/~luke/signup.htm
------------------------------
From: Marc Getty <[EMAIL PROTECTED]>
Date: Sat, 20 Feb 1999 12:26:47 -0500
Subject: Mersenne: Exponents > 7 Million
> Just got my first primenet assignment over 7 million yesterday. phew.
> estimated 65 days on a pentium 166mmx system (ok, so this is one of my junk
> boxes that happened to get it).
I got my first exponent over 7 million on Tuesday, and have received
25 more since then. They appear to take 30 days on a PII 300 MHz, 40
days on a PII 233 MHz and 50 days on a PPro 200 MHz.
My smallest # is 7000313 and my largest to date is 7030579.
Marc Getty - [EMAIL PROTECTED] - [EMAIL PROTECTED] - ICQ: 12916278
http://www.getty.net - http://www.vwthing.org Work: 215-204-3291
http://etc.temple.edu/ Home: 215-322-8363
________________________________________________________________
Unsubscribe & list info -- http://www.scruz.net/~luke/signup.htm
------------------------------
From: [EMAIL PROTECTED]
Date: Sat, 20 Feb 1999 20:13:47 +0100
Subject: Mersenne: all these primes...
i like the idea of finding larger and larger prime numbers but i still
dont know what this will do for modern mathematics. Since numbers are
infinite, these primes must also be infinite but with higher counts
between each one. what's happening here; will future mathematicians
really need this data or is it more applicable to cryptography? thanks
all!
- --
Oliver Bonham-Carter
- -- By Order of the Fat Monkeys --
___________________________________________________________________
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
------------------------------
From: Bryan Fullerton <[EMAIL PROTECTED]>
Date: Sat, 20 Feb 1999 15:06:34 -0500
Subject: Re: Mersenne: Factoring on a 486SX
On Sat, Feb 20, 1999 at 12:17:58PM -0500, Marc Getty <[EMAIL PROTECTED]> wrote:
>
> The "487 math co-processer" is actually a 486DX chip. Unlike previous
> Intel CPUs where you just add an external math co-processer, the 486
> SX line was an afterthought.
Actually, 486sx's were 486dx's that didn't pass the math co-processor QA
testing. I was selling them when they came out.
> I have never seen a 487, all I have ever
> seen is 486DX's. They may exist, but in all likelihood they are just
> a re-marked 486DX processor.
I've sold 487 processors before - my question was not whether they exist,
but whether they'd speed up factoring. I've so far gotten 4 replies that
say there will be no increase because the math is all integer operations,
and one reply saying it will increase by 10-100x (which I assume is in
error). Thanks for all replies!
> Assuming your CPU is not a soldered on surface mount CPU
The system is an Ambra Enterprise, and the CPU is integrated onto the
motherboard.
Bryan
- --
Bryan Fullerton http://www.samurai.com/
Owner, Lead Consultant http://www.feh.net/
Samurai Consulting http://www.icomm.ca/
"No, we don't do seppuku." Can you feel the Ohmu call?
________________________________________________________________
Unsubscribe & list info -- http://www.scruz.net/~luke/signup.htm
------------------------------
From: Paul Derbyshire <[EMAIL PROTECTED]>
Date: Sat, 20 Feb 1999 17:13:22 -0500
Subject: Re: Mersenne: all these primes...
At 08:13 PM 2/20/99 +0100, you wrote:
>i like the idea of finding larger and larger prime numbers but i still
>dont know what this will do for modern mathematics. Since numbers are
>infinite, these primes must also be infinite but with higher counts
>between each one...
Actually, it is unknown whether there are infinitely many Mersenne primes,
although it is suspected that there are. To prove that there are requires
more than just finding lots of very big ones, and to prove that there are
not requirwes more than just finding a huge gap while pushing the exponent
envelope. A solid mathematical proof of either must be general... and
nobody has come up with one.
- --
.*. "Clouds are not spheres, mountains are not cones, coastlines are not
- -() < circles, and bark is not smooth, nor does lightning travel in a
`*' straight line." -------------------------------------------------
-- B. Mandelbrot |http://surf.to/pgd.net
_____________________ ____|________ Paul Derbyshire [EMAIL PROTECTED]
Programmer & Humanist|ICQ: 10423848|
________________________________________________________________
Unsubscribe & list info -- http://www.scruz.net/~luke/signup.htm
------------------------------
From: Spike Jones <[EMAIL PROTECTED]>
Date: Sat, 20 Feb 1999 17:01:45 -0800
Subject: Re: Mersenne: all these primes...
[EMAIL PROTECTED] wrote:
> i like the idea of finding larger and larger prime numbers but i still
> dont know what this will do for modern mathematics....
i thought of a good reason to have enormous prime numbers:
we would use them to impress extraterrestrial intelligences, should
we ever manage to contact them. any alien civilization capable
of building receivers must understand mathematics, and since
primes are prime in every base, they would understand mersenne
primes too. they may call them something else, such as
freemblookum primes, but they would get it. if there is no
other way, in principle, to discover freemblookums other than
massive brute force computation, we might make a little game
of it: we send them the first 37, they send us the 38th, we send them
39th etc, and so the whole universe will watch the outcome of our
first interstellar intelligence contest... {8^D spike
________________________________________________________________
Unsubscribe & list info -- http://www.scruz.net/~luke/signup.htm
------------------------------
From: Richard McDonald <[EMAIL PROTECTED]>
Date: Sun, 21 Feb 1999 11:00:18 -0500
Subject: Mersenne: MacGIMPS -- Benchmark info needed
(To Mersenne mailing list and known MacGIMPS users)
I could use some benchmark information on Mac computers running MacGIMPS to
help calibrate the kind of work to allocate. Anyone who is running
MacGIMPS on a PPC Mac who would be willing to run a timing test for several
hours, please read on. Anyone else, thanks for listening.
To run a timing test:
Please dedicate several hours to this test when your Mac is not doing
anything else -- say overnight.
Restart with no extensions running, and start up no other software.
Set your work-in-progress files aside, and start a MacGIMPS test on the
exponent 7,654,321 (that's not a magic number, I just want all timings to
be comparable and they depend on the exponent. )
Configure the following MacGIMPS options:
- - No checkpoints (set checkpoints to a large number like 1,000,000)
- - Create a new log file and request logging every 200 iterations
- - Responsiveness 0 (fastest compute)
- - No scheduled responsiveness changes
Let the test run for several hours -- overnight would be great.
Then, mail the resulting log file to mailto:[EMAIL PROTECTED] --
attach it or cut&paste into the email.
IMPORTANT: Include your CPU type (604, G3), clock rate, and the version
number of your copy of MacGIMPS in the email. Just the model number of
your computer will do if you don't know the CPU and clock rate. The test
is not useful without this information and if you forget to provide it I'll
ask you.
It's important that this test run undisturbed (no extensions and no other
software) and at priority 0 for consistency. If you can't do that because
of other demands on your computer time, thanks for your interest but please
don't send in tests.
You can throw the checkpoint file and document file away -- only the timing
information in the log file is needed.
Thanks & regards,
Richard
________________________________________________________________
Unsubscribe & list info -- http://www.scruz.net/~luke/signup.htm
------------------------------
From: =?iso-8859-1?Q?Andr=E9?= de Boer <[EMAIL PROTECTED]>
Date: Sun, 21 Feb 1999 17:56:27 +0100
Subject: Mersenne: linux
Hello,
Recently i installed linux on my pc and i'm thinking of
switching to this platform.
What's the best way to transform my prime configuration to
this platform.
Can i download the files for linux and copying the p* and q*
files to linux?
Thanks in advance for the reaction,
Andre de Boer
________________________________________________________________
Unsubscribe & list info -- http://www.scruz.net/~luke/signup.htm
------------------------------
From: "Floris Looyesteyn" <[EMAIL PROTECTED]>
Date: Sun, 21 Feb 1999 13:43:00 +0100
Subject: Re: Mersenne: all these primes...
>i thought of a good reason to have enormous prime numbers:
>we would use them to impress extraterrestrial intelligences, should
That is indeed a great idea!
> they may call them something else, such as
>freemblookum primes,
actually they call them du=fjxxqr'eowcs-ss, although
I never heard them say it.
No serious, don't you think that in a few years some
mathematic will come up with a way to calculate
which primes are mersenne primes? I mean without the
need to test all primes. And what will happen if he
hears about all those thousands of suckers who spent
lots of computers and power searching for them?
Floris.
________________________________________________________________
Unsubscribe & list info -- http://www.scruz.net/~luke/signup.htm
------------------------------
From: Spike Jones <[EMAIL PROTECTED]>
Date: Sun, 21 Feb 1999 22:06:10 -0800
Subject: Mersenne: Re: Mersenne Digest V1 #510
i see the new top producer is TempleU-CIS. i assume
this is temple university in pittsburg pennsylvania. i conjecture
that cis is computer information security? congratulations
templeU-CIS. how did you do it? spike
________________________________________________________________
Unsubscribe & list info -- http://www.scruz.net/~luke/signup.htm
------------------------------
End of Mersenne Digest V1 #512
******************************
