Mersenne Digest Friday, January 2 2004 Volume 01 : Number 1101
---------------------------------------------------------------------- Date: Mon, 29 Dec 2003 10:18:12 -0500 From: George Woltman <[EMAIL PROTECTED]> Subject: Re: Mersenne: Re: Where is M23494381? At 10:48 AM 12/29/2003 +0000, Doug Feather wrote: >I've had the same experience with some exponents assigned to me. My guess >is that someone has been doing trial factoring work semi-independent of >PrimeNet and has just reported all the work that they have done. This is probably a good explanation - especially in light of your point #2 >1. For a long time on this page: >http://mersenne.org/primenet/ >there where no exponents in the range listed 21.9M to 22.0M. This has >recently changed with them now all being listed as available for first time >Lucas Lemer testing. These exponents were manually reserved a year and a half ago for some trial factoring work. Yesterday, I cancelled that reservation and added the exponents to the server. _________________________________________________________________________ Unsubscribe & list info -- http://www.ndatech.com/mersenne/signup.htm Mersenne Prime FAQ -- http://www.tasam.com/~lrwiman/FAQ-mers ------------------------------ Date: Mon, 29 Dec 2003 19:39:17 +0100 From: =?iso-8859-1?Q?Ignacio_Larrosa_Ca=F1estro?= <[EMAIL PROTECTED]> Subject: Re: Mersenne: Re: Where is M23494381? Monday, December 29, 2003 11:48 AM [GMT+1=CET], Doug Feather <[EMAIL PROTECTED]> escribió: >> I has assigned that exponent to factor. But today it disappears from >> my Individual Account Report. > > I've had the same experience with some exponents assigned to me. My > guess is that someone has been doing trial factoring work > semi-independent of PrimeNet and has just reported all the work that > they have done. My reasons for this belief is 2 fold. > ... > 2. I had a couple of exponents booked out on one machine, these too > disappeared recently but that machine has since checked in to give > new end dates. They now appear there again but there are 2 > differences. The booking out date has changed to the check-in date > and the range the number had already been factored to has increased. > He is the listing as it was just before it disappeared and how it > appears now: > > 23492377 F 57 108.0 84.4 64.4 14-Nov-03 19:56 > 07-Sep-03 11:52 Dad 1800 v19/v20 > 23492377 F 58 1.2 134.8 102.8 > 27-Dec-03 13:59 Dad 1799 v19/v20 > > - Range already factored to has increased from 57 to 58. > - Days assigned to me reduced from 108.0 to 1.2 > - Date assigned to me changed from 7/9/03 to 27/12/03 (or in US date > format 9/7/03 to 12/27/03) Yes, it is the same situation. The exponente is now reasigned to me with a bit more already factored. Thanks. Best regards, Ignacio Larrosa Cañestro A Coruña (España) [EMAIL PROTECTED] _________________________________________________________________________ Unsubscribe & list info -- http://www.ndatech.com/mersenne/signup.htm Mersenne Prime FAQ -- http://www.tasam.com/~lrwiman/FAQ-mers ------------------------------ Date: Mon, 29 Dec 2003 21:46:21 +0100 From: Matthias Waldhauer <[EMAIL PROTECTED]> Subject: Mersenne: Different word sizes/accuracy for forward/inverse transform During my studies of FFT+NTT I got following idea: In case of FFT we have to provide enough bits per word to safely hold the squares of the values in frequency space without getting rounding errors during the inverse transform. But actually we don't need that many bits during the forward transform. That offers the possibility to use a floating point format with a smaller mantissa. I assume that unfortunately the 24bit mantissa of single precision numbers is just a few bits too short to fully utilize the double precision 53bit mantissa on the way back. However there is maybe still a possible gain by using SP for forward and DP for inverse transform thanks to roughly double throughput for single precision calculations using SIMD instructions on some architectures. The memory traffic wouldn't decrease much if at all because IMO it would be better to have the 32bit values at nearly the same locations as the doubles later to avoid additional TLB/cacheline issues, which could appear if some out-of-place transform style is being used. An enhancement is possible by grouping SP values of 2 cachelines together into one to reduce memory traffic. The unused space of cacheline size would be used for the doubles and the grouping of SP values is also necessary to be able to load them efficiently using SIMD instructions. Another story is, if this scheme could be applied to NTTs by using primes of different sizes. Unfortunately the values we get after a forward transform are repeatedly multiplied by powers of roots and added - - always modulo a chosen prime. So I imagine it will be difficult or even impossible to change these values in a way that a different prime (with more bits) can be used for the inverse transform. Any ideas? _________________________________________________________________________ Unsubscribe & list info -- http://www.ndatech.com/mersenne/signup.htm Mersenne Prime FAQ -- http://www.tasam.com/~lrwiman/FAQ-mers ------------------------------ Date: Fri, 02 Jan 2004 20:23:40 -0800 From: "Terry S. Arnold" <[EMAIL PROTECTED]> Subject: Mersenne: Hyperthreading & ABIT IS7 I just brought up a new box with P4 3.0 on an ABIT IS& MB. I only appear to be getting 50% of the cycles for Prime95. I am running XP Pro SP1. How do I get the full power available to Prime95? Terry _Terry S. Arnold 2975 B Street San Diego, CA 92102 USA [EMAIL PROTECTED] (619) 235-8181 (voice) (619) 235-0016 (fax) _________________________________________________________________________ Unsubscribe & list info -- http://www.ndatech.com/mersenne/signup.htm Mersenne Prime FAQ -- http://www.tasam.com/~lrwiman/FAQ-mers ------------------------------ End of Mersenne Digest V1 #1101 *******************************
