An offer...
 

After reading the "reading the status chart" thread in the Info and Answers section of the forum, I gave the following post- I'd like to repost it here so that more people read it...please post your comments. Thanks.


To the veterans on this forum... I have been a member of GIMPS for about a year and a half now, and a viewer of the project longer than that. Maybe surprisingly, I'm a high school student- actually a senior this year. This isn't the usual demogrpahic you might expect to participate in a distributed computing project, let alone one that deals with mathematics. However, I must say that although math may not be my strongest subject, it is one I mind the most interesting.

Some of you may know of an academic program by the name of "International Baccalaureate" or I.B. for short. If not, it is a structured and rigorous curriculum that is internationally organized. Getting to the point, for this program and more specifically for one of its classes, I need to adopt and carry out a project that is beneficial to someone, or a group. The project is due around mid-November and I have not yet decided what to do. I have not found anything locally that I can work on, though if I dug a little deeper I'm sure I could find something.

My question is this...if it was decided that the "status page" could be redone in a new format, the table, presentation format, etc; could I adopt that as my project for my class? I.B. stresses the project matches the client's needs, or yours in this case. I'd design it to your specifications. Some of the things already pointed out where "having a more linear logarthmic distribution of the ranges" and how the data on the table is presented- "work completed" vs. "work to be done," among other things, possibly linking the table to the server so that it updates automatically once the desired format has been achieved so as to eliminate manual errors. Its whatever the "client" wants. I'm just putting this offer on the table, whether George or anyone else here wants it or can use it. If so, then I will check with my instructor for the class and see if the project is appropriate, and I cannot see why it wouldn't be. I'm offering because this is something that I'm deeply interested in, and doing the reports and write-ups are always higher quality when you are working with something you are passionate about. Thanks for the consideration.

Kyle,

A fellow cruncher.

Kyle,

If you have been following the various discussions presented on this forum, you might notice that the "NFSNet" group has been pursuing its goal of factoring large numbers for a considerable number of years. Back before the turn of the Century (does that sound better than "a decade ago"?) we were setting records in the factoring of large numbers. Since then, a few of us have been "plodding along". We could sorely benefit from some "new blood" and enthusiasm. I am certain that, if you are willing, you, or others, would receive significant assistance if you desire to enhance our distributed effort in any manner.

If you have any interest in contributing to our effort, please contact me, Xilman, or Dr. Bob Silverman, either publicly or privately. (Sorry, Paul, "Dr. Xilman" just doesn't sound right).

Richard "Wacky" Wackerbarth

I took IB too. Let me guess; your project is for ITGS, right?

I would like to second Wacky's suggestion. The NFSNET webpage has been in sore need of a redesign for a couple of years, and NFS has been the most advanced state-of-the-art factoring algorithm for some time. You really don't need to even to know much about NFS to take on this project, but you should think about what sort of a presentation might help attract new participants. Browse the pages, and I am sure that Richard, Paul (Leyland), and Bob would be more than willing to answer your questions.

[QUOTE]If you have any interest in contributing to our effort, please contact me, Xilman, or Dr. Bob Silverman, either publicly or privately. (Sorry, Paul, "Dr. Xilman" just doesn't sound right).[/QUOTE]

I found the page, but what exactly needs to be done? I couldn't find the status page that was mentioned on one of the links, and the form of the number seems ambiguous- or at least seems that it would be much harder to show progress status because not every test is represented as an LL or factoring of 2^P -1. But then again, maybe I just need an idea of what you want.

[QUOTE]I took IB too. Let me guess; your project is for ITGS, right?[/QUOTE]

Yes, it is for ITGS. How was your IB experience and what did you do for your project?

[QUOTE=Primeinator;87291]I found the page, but what exactly needs to be done? I couldn't find the status page that was mentioned on one of the links, and the form of the number seems ambiguous- or at least seems that it would be much harder to show progress status because not every test is represented as an LL or factoring of 2^P -1. But then again, maybe I just need an idea of what you want.[/QUOTE]

Here is a link to some of our previous "stats".
[url]http://www.nfsnet.org/stats2/statsreporter.cgi?project=2_1466L&template=index.html[/url]

Basically, we "measure" each machine's participation (we have a metric that seems to work across multiple projects). The (multiple) servers can spit out "Machine MMM contributed UUU units of work on project PPP on date DDD". In some manner, we need to collect, total, and present this participation across multiple projects in a way that encourages additional participation. I leave it to the discression of the designer as to how this can best be accomplished.

There are a lot of "details" that I will be happy to further explain if someone is really interested in spending some significant effort in producing some "finished product". (at least a prototype :)

We also have some resource problems. But I feel that we can address them if we actually have someone wanting to "do something" rather than just lamenting the lack of our presentation.

[QUOTE=Primeinator;87291]Yes, it is for ITGS. How was your IB experience and what did you do for your project?[/QUOTE]

I made an online calendar converter. But my school gave me months to do the project, not weeks. In fact, I started the project during my junior year and finished in my senior year.

Then again, I didn't do very well in the course, I finished with a 5, if I remember correctly.

OK, this is really addressed to Kyle, rather than "jinydu".

[QUOTE=jinydu;87295]But my school gave me months to do the project, not weeks[/QUOTE]

I'm not familiar with this program. However, based on my decades of experience in software and hardware development, it is really difficult to COMPLETE anything MEANINGFUL in such a short period of time. If your advisor/mentor has already set up some project for you, then it is reasonable that you might complete "his" project within the alloted time. However, absent someone "giving you the project ready to be done", I have real reservations about how much you can accomplish on ANY project in such a short period. Design of "the project" is, in itself, a formidable task.

But, you should also understand that I came from an earlier era. While in Grad School, I received a grade of "Incomplete" on my "Graduate Research" course each semester for two years before I finally was ready to make my presentation and receive credit for those many semesters of effort.

Still, if you are interested and feel that any of us can help, there are at least a few of us who are quite willing to do whatever we can to assist in your educational process.

Here is the calendar converter website:

[url]http://www.freewebs.com/jincalendarconverter/[/url]

Wacky, just in case it wasn't clear, this is a high school project, not a graduate school project. The amount of time given is actually very high compared to the time given for most assignments. Currently, I am an undergraduate and I routinely have to complete 20+ page math assignments in one week. My record is 50 pages in 1 week.

[QUOTE]Wacky, just in case it wasn't clear, this is a high school project, not a graduate school project.[/QUOTE]

I don't dispute that.

When I was at that (high school) academic level, my advisors were there to make suggestions about projects that were already sufficiently defined that I could reasonably choose one of them and be reasonably assured that I could complete the requirements within the time allotted. My "choices" were far easier than the scope of Kyle's initial query.

As much as I would like to have Kyle make some contribution to "my" (NFSNet) project, or any other, I really am more concerned that he not be drawn into something that is, in scope, "over his head", given the time constraints of his service.

In that respect, I worry that he really needs to look to his advisor for a project that is "better defined".

At the same time, I would welcome his "introduction" to his advisor in the hope that he and I could propose an appropriate project for some future student.

Richard "Wacky" Wackerbarth

[QUOTE]But my school gave me months to do the project, not weeks. In fact, I started the project during my junior year and finished in my senior year.[/QUOTE]

Actually its just that we have to decide what we are doing for our project soon. The actual project isn't due until early 2007. However, we need to have a start or a good idea of what we are going to do long before then.

[QUOTE]As much as I would like to have Kyle make some contribution to "my" (NFSNet) project, or any other, I really am more concerned that he not be drawn into something that is, in scope, "over his head", given the time constraints of his service.[/QUOTE]

As mentioned above, I do have more time than was originally interpreted from my posts, it is just the deadline as to a "basic" idea or what I am going to do for the project that I need to decide upon in the next few weeks or so. I will look at the page you gave tomorrow to see what's what.

[QUOTE]When I was at that (high school) academic level, my advisors were there to make suggestions about projects that were already sufficiently defined that I could reasonably choose one of them and be reasonably assured that I could complete the requirements within the time allotted.[/QUOTE]

That is part of the I.B. program, to develop work independence and thinking. I have to admit that having such an open-ended assignment can be frustrating at times, however the I.B. program is highly regarded (if you do well on the exams at the end of the courses) among colleges- especially American ones, whose respective high schools are much more lax than high schools in Europe. Other parts of the criteria are strict and defined- as in covering so "requried material" and then choosing the enrichment materials. Its a good program, if you want their home site is [URL="http://www.ibo.org"]http://www.ibo.org[/URL]

[QUOTE]I made an online calendar converter. But my school gave me months to do the project, not weeks. In fact, I started the project during my junior year and finished in my senior year.[/QUOTE]

[QUOTE][url]http://www.freewebs.com/jincalendarconverter/[/url]
[/QUOTE]

That's pretty cool, how much mathematics did it require on your part?

[QUOTE]Still, if you are interested and feel that any of us can help, there are at least a few of us who are quite willing to do whatever we can to assist in your educational process.[/QUOTE]

I appreciate this a lot. Even if I don't end up using NFSNet as my project, it is still something I would like to help with in the future, time allowing. My schedule this year is pretty crazy. And, if I end up using either the main project status or yours, or any project I do just for the fun of it in the future, I will probably need your help for some of the technical aspects (which is allowed for the school project). I know there are a number of very knowledgable people on these forums.

Thanks for the support,

Kyle

[QUOTE=Primeinator;87312]That's pretty cool, how much mathematics did it require on your part?[/QUOTE]

Hardly any at all. The "algorithm" is used is brute force. I used other calendar converters to build an Excel file, then linked the website to the Excel file. You will also find that my converter is quite limited; I think it only extends to the end of 2006 or something (I don't remember exactly).

Wacky, unfortunately, it is unlikely that Primeinator will be given a more specific assignment. One of the frustrating aspects of parts of the IB diploma is that some of the work is very open-ended, which can be bewildering for some students (like me).

Oh well, at least I got 7's in Higher Math and Higher Physics.

[QUOTE=Primeinator;87134]
To the veterans on this forum... I have been a member of GIMPS for about a year and a half now, and a viewer of the project longer than that. Maybe surprisingly, I'm a high school student- actually a senior this year. This isn't the usual demogrpahic you might expect to participate in a distributed computing project, let alone one that deals with mathematics. However, I must say that although math may not be my strongest subject, it is one I mind the most interesting.
[/QUOTE]
You're not alone. A few years ago, I discovered my first top 5000 prime as a 15 year old :smile:

If you're curious, my school didn't offer IB's. There were only honors and AP classes for people who wanted to challenge themselves.

[QUOTE=Primeinator;87312]Actually its just that we have to decide what we are doing for our project soon. The actual project isn't due until early 2007. However, we need to have a start or a good idea of what we are going to do long before then.

As mentioned above, I do have more time than was originally interpreted from my posts, it is just the deadline as to a "basic" idea or what I am going to do for the project that I need to decide upon in the next few weeks or so.[/QUOTE]

Thank you for this clarification. I am pleased to learn that you have an adequate period of time to actually develop a project of reasonable complexity.

[QUOTE]You're not alone. A few years ago, I discovered my first top 5000 prime as a 15 year old

If you're curious, my school didn't offer IB's. There were only honors and AP classes for people who wanted to challenge themselves.[/QUOTE]

Good to meet you, I wonder how many other teenagers to age 21 or so are members of these forums?


[QUOTE]Hardly any at all. The "algorithm" is used is brute force. I used other calendar converters to build an Excel file, then linked the website to the Excel file. You will also find that my converter is quite limited; I think it only extends to the end of 2006 or something (I don't remember exactly).[/QUOTE]

You're correct, it says through 2006.

[QUOTE]Oh well, at least I got 7's in Higher Math and Higher Physics.[/QUOTE]

Wow, congratulations- that is very good.

[QUOTE]Thank you for this clarification. I am pleased to learn that you have an adequate period of time to actually develop a project of reasonable complexity.[/QUOTE]

Heh, no problem. I'm going to take a further look at your site now to see what I might be able to do for it. I think I will also pm George- because I have a number of ideas for the main GIMPS status page.

-Kyle

[QUOTE=Primeinator;87347]I think I will also pm George- because I have a number of ideas for the main GIMPS status page.[/QUOTE]

I've not responded until now because there are so many other deserving projects that could use qualified help. Depending on your abilities and available time you could have a major positive impact on these projects.

You could learn how to run a BOINC project and search for Cunningham factors with GMP-ECM, or search for Wieferich primes ([url]http://mathworld.wolfram.com/WieferichPrime.html)[/url], or create an NFSNet client, etc.

You could add professional touches to NFSNet's web site or client.

Or you could fall back on your idea to spiff up the GIMPS' status (and other?) pages.

[QUOTE]I've not responded until now because there are so many other deserving projects that could use qualified help. Depending on your abilities and available time you could have a major positive impact on these projects.

Or you could fall back on your idea to spiff up the GIMPS' status (and other?) pages.[/QUOTE]

This is the idea that I'm most interested in at the moment- creating a status table that includes data for all ranges of exponents from searches in the form of 2^P -1 because a number of people and large amount of processing time is being given to projects such as "100-Million Digit Prefactor" and "Operation Billion Digits." Some of the ideas I mentioned include a different layout of columns and rows, a different color scheme for different related regions of the table, small icons for each different region of the table that when clicked bring up a descriptive graphic of what that part of the table means (because some new users have a hard time understanding the table based from the post by Clyde in the Info and Answers forum), and several others. I am still brainstorming. Another idea I thought of was to have the table automatically update from the PrimeNet server say once daily. Though there are several problems with implmenting this. First, none of the exponents for OBD or 100M Digit Prefactor are tied to the server. Secondly, not all users working in largely-active ranges right now are using PrimeNet. If this idea was implemented, the table could be split for each OBD and 100M Digit Prefactor with one combined table and more complex table seperate that includes additional statistics (on the seperate tables only a few statistics, such as the present ones given, are shown). Other ideas were influenced by what other users have said. In the "reading status chart" thread in the Info and Answers forum, several users expressed a desire to have a more explicit set of ranges- either linear or logarithmic. In addition, I have a few other ideas I've come up with and I'm bound to come up with more if I actually play around with designing it. However, the focus of the this I.B. and ITGS project is to create the product that matches the needs/wants of the person or group you decided to do your project for, whatever those may be. With your permission, this is something that I would like to do- not only because I need to have some sort of related project for my class, but also because this is a subject that I have a lot of interest and passion in. Thanks for the consideration,

Kyle

You could update all those outdated Primenet banners.

[QUOTE]You could update all those outdated Primenet banners.[/QUOTE]

After reading the M44 is Here thread, I was actually going to add that to my table reconstruction.

[QUOTE]Some other banners (that need to be updated) say:

GIMPS / Primenet = 38 Cray T916 Supercomputers. Whoa.
GIMPS does 79 years of computing- every day. Help out.
GIMPS / Primenet. Join 15,000 fellow computer users.
38 of them are known in the universe. Find one more.
27,000 computers. Working as one. GIMPS / Primenet.
It exists. Hiding. Cloaked. Find Mersenne Prime #39.
GIMPS. Because 2,098,960 digits just aren't enough.
GIMPS. Because 2^6,972,593-1 just isn't big enough.
950,000,000,000 floating point operations / sec. Nifty.


These banners should say instead:

GIMPS / Primenet = 400 Cray T932 Supercomputers. Whoa.
GIMPS does 1800 years of computing- every day. Help out.
GIMPS / Primenet. Join 47,000 fellow computer users.
44 of them are known in the universe. Find one more.
72,000 computers. Working as one. GIMPS / Primenet.
It exists. Hiding. Cloaked. Find Mersenne Prime #45.
GIMPS. Because 9,808,358 digits just aren't enough.
GIMPS. Because 2^32,582,657-1 just isn't big enough.
22,000,000,000,000 floating point operations / sec. Nifty.[/QUOTE]

I can update them if you want. Right now though I need to get a little sleep- had a Youth Group Lock-In and I just got back and I'm pretty tired and have to go into work later this afternoon.

-Kyle

Kyle, on the one hand I like your idea of including 100MDPP and OBD in your tables. On the other hand, I don't quite like the idea of making the table rows linear or logarithmic. I think in at least one table the thing should be kept the way it is, because it shows us the FFT sizes used - this relates to the time needed to test an exponent. If I read the table correctly, and if we assume we're testing exponents around 35M, when we're testing exponents around 70M the time needed won't be just twice as now, it'll be larger because of the larger FFT (please notice I'm considering the same machine doing both tests).

What I'd suggest then would be, firstly, one "overall" table showing that there are at least five different Mersenne exponent areas:
(1) <14M, tested and double-checked, only 39 Mersennes here
(2) 14 - 79.3M, GIMPS/PrimeNet current work area
(3) 79.3 - ~332M, Lone Mersenne Hunters >79.3M area
(4) ~332M - ~3.3G, 100M Digits Prefactor Project area
(5) >~3.3G, Operation: Billion Digits area

And then, for areas (1) and (2), which are currently represented [URL="http://www.mersenne.org/status.htm"]here[/URL], instead of displaying it linear or logarithmic, why don't you change the row height? It's a way of keeping the FFT sizes and show the ranges covered are different in size.

Cheers,
Bruno

[QUOTE=brunoparga;87424]Kyle, on the one hand I like your idea of including 100MDPP and OBD in your tables. On the other hand, I don't quite like the idea of making the table rows linear or logarithmic. I think in at least one table the thing should be kept the way it is, because it shows us the FFT sizes used - this relates to the time needed to test an exponent. If I read the table correctly, and if we assume we're testing exponents around 35M, when we're testing exponents around 70M the time needed won't be just twice as now, it'll be larger because of the larger FFT (please notice I'm considering the same machine doing both tests).

What I'd suggest then would be, firstly, one "overall" table showing that there are at least five different Mersenne exponent areas:
(1) <14M, tested and double-checked, only 39 Mersennes here
(2) 14 - 79.3M, GIMPS/PrimeNet current work area
(3) 79.3 - ~332M, Lone Mersenne Hunters >79.3M area
(4) ~332M - ~3.3G, 100M Digits Prefactor Project area
(5) >~3.3G, Operation: Billion Digits area

And then, for areas (1) and (2), which are currently represented [URL="http://www.mersenne.org/status.htm"]here[/URL], instead of displaying it linear or logarithmic, why don't you change the row height? It's a way of keeping the FFT sizes and show the ranges covered are different in size.

Cheers,
Bruno[/QUOTE]

I don't see a problem. FFT size is related to the logarithm. If you'll notice on the table you linked...the rows are separated by a factor of about 2[sup]1/4[/sup]. This is delimiting the base 2 log by fourths.

The Primenet status page uses a linear scale, and I believe this may be the one Primeinator was referring to. Of course, there are many more rows in this table, so they would have to be divided further than by FFT size.

Drew

Then I guess the thing to do now is decide what all information will be put on the table. My current idea, or the one I believe will be the most practical to implement, is to have several different tables (representing different ranges similar to what brunoparga mentioned) showing just basic information and statistics (range, number of exponents, testing status, FFT sizes, etc) and then having 1 larger and more complex table showing more information and combining all the ranges.

The question now becomes, "what all do you want on each table, and how do you want it depicted on the table?"

-Kyle

Another thing to keep in mind is that the FFT sizes vary depending on the version of Prime95 and the processor for each range.

[QUOTE]I don't see a problem. FFT size is related to the logarithm. If you'll notice on the table you linked...the rows are separated by a factor of about 21/4. This is delimiting the base 2 log by fourths.[/QUOTE]

How do you figure? Perhaps a lack of sleep is causing me to calculate this incorrectly...you are speaking of the exponent ranges right?

12,830,000 x 2^1/4 yields 15,257,527...which is close to the range cap 15.3M

But raising multiply this by 2^1/4 results in 18194868- and the difference continues to get better from there. However, I have a feeling that I am going about your explanation the correct way. Could you explainw what you mean by "the rows are separated by a factor of 2^1/4?" (which to me indicated multiplication).

Also, is there a reason that these particular ranges where chosen if they FFT size is a direct correlation of the logarthim anyway?

Thanks,

Kyle