Re: Mersenne: Williamette
Hi, At 08:40 PM 3/1/00 -0800, John R Pierce wrote: Hmm. Microprocessor Reports has released some interesting tidbits about the new Williamette processor which will probably be the Pentium-4... This is the chip Intel recently demonstrated running at 1.5GHz. Intel has info on this processor at http://developer.intel.com/design/processor/wmtsdg.htm The chip has good potential. The new SIMD2 instructions have the potential of doubling throughput. It also looks like FPU operations can be done in parallel with SIMD2 - that's triple the throughput of a PIII. Not to mention running at 1.5GHz! Of course, time will tell as to how good this processor really is. There could be other bottlenecks, it may not be easy to recode prime95 to use the SIMD2 instructions. The latencies for FPU operations are higher than in the P-III. The FXCH instruction is no longer free. Mispredicted branch penalties are higher, etc. etc. I've not heard any rumors as to a release date, but it looks like I'll have to buy two new computers in the next 12 months. An IA-64 and a Willamette! Regards, George _ Unsubscribe list info -- http://www.scruz.net/~luke/signup.htm Mersenne Prime FAQ -- http://www.tasam.com/~lrwiman/FAQ-mers
Mersenne: Security and Prime95/mprime
Hi, I'll likely be moving to a cable modem soon and intend to install a machine to act as a firewall, likely a Linux box. Since it will be sitting there all day doing nothing other than screening stuff between my LAN and the 'Net, I thought I'd run mprime (if Linux) on it. Of course, all of the security gurus say to run nothing beyond the programs actually needed on the firewall box. Am I creating any security risks by running mprime on the firewall box? I'm sure some of you must be doing that--noticed any problems? Thanks! Kel Utendorf _ Unsubscribe list info -- http://www.scruz.net/~luke/signup.htm Mersenne Prime FAQ -- http://www.tasam.com/~lrwiman/FAQ-mers
Mersenne: Re: Security and Prime95/mprime
On Thu, Mar 02, 2000 at 02:01:24PM -0500, St. Dee wrote: Am I creating any security risks by running mprime on the firewall box? You shouldn't, since mprime doesn't deal with server sockets (only the occasional HTTP traffic to PrimeNet) at all. The only problem I can think of, is that it eats a chunk of your RAM, so a DoS attack would probably be slightly easier (at least if somebody can connect 1 times to your FTP socket, and inetd fires up a new ftpd dæmon for every new socket). I run it on a 486sx/16 (24MB RAM, though) just fine, and that machine serves 3 proxy requests (HTTP) a day :-) /* 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
Re: Mersenne: Security and Prime95/mprime
On 2 Mar 00, at 14:01, St. Dee wrote: I'll likely be moving to a cable modem soon and intend to install a machine to act as a firewall, likely a Linux box. linux is a Good Move ... ceratinly, in its default state, it's at least as secure (when used as a firewall) as anything emanating from a certain purveyor of operating systems based near Seattle. It's cheaper, too! Since it will be sitting there all day doing nothing other than screening stuff between my LAN and the 'Net, I thought I'd run mprime (if Linux) on it. This sounds eminently sensible. Of course, all of the security gurus say to run nothing beyond the programs actually needed on the firewall box. Hey, I'm a security guru of a sort ... the idea is not to run anything which gives crackers a toehold, or causes unacceptable throttling of the firewall throughput. Am I creating any security risks by running mprime on the firewall box? I'm sure some of you must be doing that--noticed any problems? Few of us know what code George has embedded in the code which computes the tag which PrimeNet uses to check that incoming results are genuine. However, this does not seem to present a major risk! Apart from that, what mprime does is very network friendly seems to present an insignificant risk to operation of a firewall. I've run mprime on an anonymous FTP server for almost 18 months haven't had any incidents (yet). The basic rules are (a) always run mprime using "nice -n20" to give other processes all the CPU time they need; (b) never run mprime as root; (c) make it harder for any cracker who does get onto your system to exploit any weakness that may be in mprime by running it in a directory with no access to anyone except a user set up specially to run mprime. And make sure shadow passwords are enabled. Recent linux distributions do this by default. All this is virtually paranoia since I believe the risk posed by running mprime is practically nil - but it's good practise, anyway. Regards Brian Beesley _ Unsubscribe list info -- http://www.scruz.net/~luke/signup.htm Mersenne Prime FAQ -- http://www.tasam.com/~lrwiman/FAQ-mers
Mersenne: How much are the 10 M gamblers contributing?
Something I've been wondering about since entering this project - How much are the people who are trying to find a 10 million digit prime contributing to the search? They stand less than a tenth the ordinary chance per unit time, and any prime they find will take at least four or five years to definately place in sequence. I do understand that GIMPS' share of the prize money, if any, will go a long way towards maintaining the PrimeNet server, paying developers/porters of the client and so forth, but I just can't help wondering whether the people (It's only about five percent of the userbase, is stats is any guide) who are doing this are contributing. I'm not trying to flame anyone. Currently I am myself about a fifth of the way through two exponents in the 9.3 M range (with a p3-600 win98, if anyone cares), but I do respect that others may have different goals. If I succeed in finding a prime, I will have set a record that will likely stand for a year or two, and it would look great on my resume going into college :-) (I have taken some advanced math courses at the Rochester Institute of Technology while a high school student, if anyone is confused) I would certainly be more than happy if I found a prime, and even if I never do (which is quite likely) I will be content to have helped this project and to have sped up somewhat the efforts of whomever does win. After all, without each failure, the winner would have been delayed by one exponent. Another question, on a slightly different note: How likely is it that the cutting edge, so to speak, of the main first-time LL effort will check all the exponents up to 10 million before a 10 million digit prime is found? We're about a third of the way there now in terms of exponents, about halfway or more when you rule out composite exponents and I have no clue how close in terms of actual P90 time. Please, feel free to correct me if I made some factual errors - I have been, after all, only here eleven days. __ 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
Mersenne: Re: Williamette
On Thu, Mar 02, 2000 at 01:40:55PM -0500, George Woltman wrote: The new SIMD2 instructions have the potential of doubling throughput. But why would Intel market these instructions as `multimedia' instructions? Surely no normal MM tasks would need double precision. Of course, I shouldn't complain :-) The FXCH instruction is no longer free. Is SIMD2 (or SSE2, or whatever Intel likes to call it) _still_ stack-based? I thought Intel should have learned by now? By the way, if the registers are not aliased upon the FP registers, what will Intel do with the task switch problem? Back when MMX was new, I heard the reason for aliasing the MMX registers upon the FP registers was that no OS change would be neccessary (to save/restore the registers). Mispredicted branch penalties are higher, etc. etc. Any idea why? BTW, branching when there is two different sets of code (running in parallel) to take care of will be quite interesting :-) /* Steinar */ -- Homepage: http://members.xoom.com/sneeze/ _ Unsubscribe list info -- http://www.scruz.net/~luke/signup.htm Mersenne Prime FAQ -- http://www.tasam.com/~lrwiman/FAQ-mers
Mersenne: searching the biggies
Nathan Russell asked: How much are the people who are trying to find a 10 million digit prime contributing to the search? Any machine that is running GIMPS is contributing to mapping the great universal math-space. If one is factoring, double checking, finding a slew of Mersenne composites (I know 30 now), searching for 10E7s, regardless, any way you look at it, you are on the team. If we use the orderly system George and Scott have created, we are uncovering information that cannot be found any other way, information which will be our gift to humanity for all time. spike
Mersenne: searching the biggies 2
Nathan asked: How much are the people who are trying to find a 10 million digit prime contributing to the search? Now that you asked, I have been pondering this point. The payoff structure offered by the EFF has in some ways been a hindrance to GIMPS as well as a motivator. It is aesthetically unpleasing to have a big gap in the searched base, however, the "leading edge" will eventually catch up. It should be a straightforward task to estimate when that will be. When the 10E7 Mersenne is found and the 100K EFF prize is awarded, there is no reason to think some yahoos will jump up and try for a 10E8, since it would take over a century on a PII/400. Then, if a gap still exists, it would fill in and those who were trying for the 10E7 will have contributed just as much as anyone else. On the other hand, the existence of the EFF prize is a useful tool in convincing companies to run GIMPS, for most IT managers are quick to remind you that this *is* a business, and it is here to make money. (Then you hope that same IT manager hasnt the sophistication to calculate the actual odds of bagging the $100K, realizing that the mathematical expectation of the prize would not pay the extra electricity use...) spike
RE: Mersenne: searching the biggies 2
On the other hand, the existence of the EFF prize is a useful tool in convincing companies to run GIMPS, for most IT managers are quick to remind you that this *is* a business, and it is here to make money. (Then you hope that same IT manager hasnt the sophistication to calculate the actual odds of bagging the $100K, realizing that the mathematical expectation of the prize would not pay the extra electricity use...) Actually, the cost of electricity isn't a factor for most large companies. I can only rely on my knowledge of the various places I've worked. We all know I worked at US WEST, and I can say that they leave their computers on all the time. The other big telco I just spent the last 1.5 years at also leaves their computers on 24x7, as does the company I work for now. The reasons for doing so are pretty simple: software deployment. When you want to push out a new version of Software Application version x.xx, you do it at night when the employees won't be affected by a reboot. So you tell people to leave their machines on, but just logout. Obviously, Wake-On-Lan is a great idea, but with legacy hardware, you either leave computers on all the time or you can't be as friendly with your software deployment. That's why I hope that some brilliant person will be able to go to US WEST and say "Hey, let's setup a Primenet Proxy, get these 30,000+ NT machines looking for primes, and do some good research". That person won't be me, by the way. Or maybe someone can go to a company like MCI WorldCom and say "Hey, let's get your 80,000+ NT machines looking for primes." Of course, I wasn't about to make that suggestion myself...what can I say? I'm a little skittish about such things. :-) Aaron _ Unsubscribe list info -- http://www.scruz.net/~luke/signup.htm Mersenne Prime FAQ -- http://www.tasam.com/~lrwiman/FAQ-mers
Re: Mersenne: searching the biggies
From: Spike Jones [EMAIL PROTECTED] Reply-To: [EMAIL PROTECTED] To: [EMAIL PROTECTED] Subject: Mersenne: searching the biggies Date: Thu, 02 Mar 2000 19:49:36 -0800 Nathan Russell asked: How much are the people who are trying to find a 10 million digit prime contributing to the search? Any machine that is running GIMPS is contributing to mapping the great universal math-space. If one is factoring, double checking, finding a slew of Mersenne composites (I know 30 now), searching for 10E7s, regardless, any way you look at it, you are on the team. If we use the orderly system George and Scott have created, we are uncovering information that cannot be found any other way, information which will be our gift to humanity for all time. spike I couldn't have said it better. That is precisely why I switched to GIMPS from another project whose sole purpose was to, at the cost of something like ten times our computing power for triple our expected time to next prime, simply recover a cute little, probably political, saying that someone had hidden. And, did I mention, provide PR for the large sponsoring corporation at the same time. Our work remaining is infinite, theirs might as well be. Two of their other projects lost GIMPS-years of work each because of stupid bugs and I got fed up and left. I will not identify my previous project; it's had enough press already because some find that its rather repetitive results further their political goals. I happen to agree with those goals, but I will no longer support the effort because it is, to be blunt, accomplishing nothing. This may seem like flamebait, and I do not intend it that way. If anyone wishes to disagree with me, on- or off-list, feel free to do so. Nathan __ 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
