Random Quote

Site Links

Welcome to p196.org!

Lychrel Numbers and Palindromes

11/06/06 I've redone the 300 dataset, and updated the info above just for posterity...

11/2/06 WOW!! Where have I been?!?!? Lost in other worlds and busy.

I'm not even going to say that I'm back in any normal sense of the word, but I have finally gotten around to updating the Milestones page to the point that I stopped at back on May 1, 2006. When I reached 300 million digits, I was just getting ready to move to North Florida. So I turned the machine off, to pack it up. Then I moved, and it just never seemed to get put back together... The summer is always busy with activities, and this one was no exception. Then it's kids back in school and whatever, and now, well... I still don't have the time to work this site, but I did want to at least put out a message that it's on long-term hold.

So if you're new to the site, I still thank you for coming by, and hope that your not disappointed that I have stopped the quest. If you're one of my old friends, I apologize to you all for being absent for so long.

I do have the site on a new server, and a new email address on the Contact Me page. With any luck, these changes will last quite a while.

I do currently have Vaughn's app re-running the 299 dataset, since I seem to have deleted the 300. Bummer! At a minimum, I want to be able to fill in the table above accurately for iterations and calculations. MAYBE I'll let it keep running, maybe not... Regardless, I've enjoyed this quest immensely, and all the friends I've made and people I've met. I just became mildly bored with it, and when that collided with all the other events... Well... You understand.



2/3/06 For better or worse, I have made changes to the way my "Table of Contents" is displayed..

I never liked the fact that if someone linked to a subpage, or if a subpage showed up in a search engine, when the user clicked on the link, it brought them over here, and kicked them all the way to the intro page. I figured if they wanted to see something on the Milestones page, they should get the Milestones page! At the time, that I made the switch to using Javascript to display the T.O.C., it was because I was frequently adding entire sections to the site, and had to keep updating ALL the pages when I added a new section. Well, I've only added a single new section in quite a while, so I've gone back to a set of hard links on each page...

I'm sure someone knows a better way, than what I've done, and I'm willing to listen to suggestions if there are any... (I know Jason had an idea a number of years ago for me, but I never implemented it at the time, and now I can't find the email. :-( I should have listened better then...)

If anyone really HATES it, let me know, and I may change it back...

1/27/06 Changes and updates all over the place. Most of them minor...

Matt and Pierre have both sent me their latest results and I have updated the Lychrel Records page and the appropriate subpages.

I've made some minor corrections and updates to the Terms and Definitions page.

I've broken out a separate page for testing of "Network Capable" 196 apps. Right now, it is available here, and can be clicked on from the Software Comparisons page. I'm not sure I like it linked that way, but it is what it is for now...

I've updated the Software Comparisons page itself with the latest times from Pierre and Vaughn. Vaughn has moved smoothly into number 1 slot on that page. I haven't asked him if it's stable enough yet to be used as working app, so right now, I'm staying with what I've been using, even at the risk of losing some time. Of course, with Pierre's app hanging out there in space, losing a few hours or days just doesn't seem relevant any longer!! :-)

I was happy to add Afghanistan and Monaco to the Where Were You page, for a new total of 114 visiting countries.

1/26/06 I have crossed 100 Quadrillion digit pairs added!!!

Pierre has released v1.01 to me, and I have been testing it in various configurations. It is indeed faster still. Turns out that in this case, the more CPU's I add to the mix, the better the results. (Not that I had any doubt about that, but it is nice to verify that the load is being spread across other CPU's and not getting bogged down with network traffic!!) I have been mixing and matching P4's and P3's and wireless and Cat5, trying to come up with the best overall configuration. My notes are scattered and random, but I will try to get some data up this weekend.

One thing that came across in e-mail this week among the group, is that Pierre managed to process the Lychrel 10577 out to 73 million digits in only 2 1/2 DAYS!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!

WOW!!! We have no word yet on how many processors were involved, or what the configuration was, but we were all understandably shocked by the news!

Vaughn summed it up like this: Okay, it's time I threw in the towel. 73 million digits in two days? The deep test in under 1 minute? I mean come on!!! I didn't understand when you said monster!!! I stand corrected, less than a year to reach 1 billion digits. You can verify all the work Wade did in the last 6 years in less than a month!!! What would be the point of Internet networking? To reach several quadrillion in 1 minute?

Wade, I guess that's it for the quest. You can close the page and we can do something else.

I'm currently trying to decipher all of the email and traffic that has been sent recently, so maybe I can explain in general, the method Pierre used to achieve such amazing work. But it's a lot to sort through...

I'm going to add a new page similar to the Software Comparisons page, just in case anyone else wants to create an app similar to Pierre's... It just doesn't make any sense to compare it to the stand-alone apps that already exist on that page. I hope to get it up this weekend as well...

1/5/06 Just an quick update on Pierre's progress... v0.9 managed to shave some more time off itself... Tonight's best run clocked in at 9:43 for my deep test. I'm trying to figure out how to "standardize" that testing, since it should be different with different numbers of machines or network setups... I'll formalize it when someone else sends an app that runs across a network. For now, I'm just reporting the fastest time I've managed to get for the deep test from any configuration.

Vaughn also got inspired to add Hyper-threading to one of his app versions... The version I tested tonight ran an impressive 0:51 for the shallow test and 12:12 for for the deep. WELL below the existing records on the Software Comparisons page. Congratulations to Vaughn!!! I will be out of town over the weekend, but I will do some more testing early next week, and when I feel comfortable with all the aspects of this new app, I will probably make the switch! (That is unless Eric steps up the the plate over the weekend and I come back to a new app from him as well! :-) )

The inspiration that seems to have hit everyone with the release of Pierre's announcement only helps me in the long run. I hope everyone continues to compete to improve!! Let the blood and sweat fly!!!


1/2/06 Happy New Year to all!! I sincerely hope that everyone had a safe and happy New Year, and that everyone prospers in 2006!!

I've archived 2005's info, so if you want to see it, or any of the other history, make sure you check out the Blackboard Archives page.

Of mild interest, there have been no new visiting countries for the last few months. Maybe everyone on the planet with an internet connection has already checked in with us! Of course, that would mean that only 41,804 people have internet connections, so I guess that's not it... :-)

Eric Goldstein sent me word that he has passed 210 million digits on 879. I haven't heard any updates recently from Matt or Vaughn on any of the other numbers that were under test, but what I do have has been updated up the Lychrel Records page.

I mentioned in November that Jason had finished the search of the 18 digit numbers for the longest delayed palindrome. There was some confusion on my part about whether Vaughn had verified Jason's work, or if the results were messed up. For Clarification, the following comments were sent by Vaughn on 11/27/06: I have confirmed Jason's results for the 18 digit set. Processing by distributed computing completed Sunday 20th November. Total processing time 129 days (cumulative over 5 machines). Also my single threaded non-distributed application verified all of Jason's data from the 1 digit to 16 digit sets several times, the latest being back on June 21st. (That latest run, using the latest optimized [at that time] Pentium 4 version, took just over 7 1/2 days. My latest Pentium 4 version would take 88% of that time and my Athlon version would take 61% of that time though on a slower processor. The tests up to 16 digits were run to a lower iteration limit than what Jason used however we both ran the 18 digit sets to the same iteration limit of 300 and we have agreed on a limit of 315 for the 19 digit set.) Both Jason and Vaughn have since moved on to the 19 digit numbers. The difference is that Jason expects around 5.5 years to complete the set, and Vaughn's new distributed app is expected to finish in only around 9 1/2 months!

Then, on November 30th, Jason announced he had discovered a new record for longest delayed 119 digit palindrome at 261 iterations. If I have the dates and info correct, Vaughn actually found a 259 iteration record before Jason found this one, but they were independently running over the dataset in a different method, so Vaughn found one and jason found a different one within a short period of time. Check out Jason's World Records web page for details.

On December 7th, 2005, I got a note from Kevin Bowman of the United States. He wrote about about his thoughts and adventures here: Scherzi . I'm really happy he wrote, since it introduced me to the following piece of electronic test equipment: The Palindrometer !! I've just got to get one!! Just in case that the linked original dies in the future, I have shamelessly copied the photo to my own site, and you can get it here. My thanks to Kevin for the article.

Jason made a very persuasive argument about crediting the people that worked on the 196 problem BEFORE John Walker's 1 million digits. As a result, I've added some information on their accomplishments to the Milestones and Credited People pages. Right now, all of that information comes directly from Jason's page, with his permission. If I find more references, I'll add them as well. My thanks to Jason!!

And I think finally for today...

I would like to introduce Pierre-Andre Laurent. Pierre has written the first app that I have ever seen that will run the 196 quest across a network!! The one item that has been longest on my Wish List has finally been offered to me.

Pierre and I have been talking most of the month of December, and I have been testing one version or another all month long. Some of them ran across the network but didn't output a file I could verify. Others output a file, but wouldn't run across my network. Still others did both, but wouldn't load my ISF files for deep testing. Earlier today, Pierre sent me a version that does all three!! This was the minimum functionality that I would accept for an announcement.

How does it stack up??? I should wait a month to report my results, so that everyone can fume with anticipation, but I can't do that to you guys... But I first need to say that Pierre has an entire list of "to do" items before he's satisfied with it enough for "real world use". The app I have is very much a work in progress.

Soooo????? How does it perform right now?? In a word: Amazing.

Consider the improvements to any of the apps over the last couple years have only been measured in a few seconds in deep testing. The fastest I have been able to get this app to perform is a full 3 minutes and 49 seconds faster in the deep test than Eric Goldstein's app!!! So you'll know I didn't mistype it, I'll repeat it: 3:49 faster than Eric's app.

As a stand alone app, Pierre has managed to get a time of 1:09 Shallow and 15:25 Deep (Compared to the record 1:00 and 13:55.) However, the more occurrences of the app I add to the system, the faster it seems to go. Taking advantage of the P4's Hyper-threading ability, I can get a host and two clients running on one machine and 2 additional clients running on a second machine talking to each other through my 100 mbs router. This gives me a total deep time of 10:27. When I add my 2.53 GHz laptop, and run 2 clients on it as well, the time for the deep test drops to 10:06. At this point, I'm out of fast, easy machines to hook up. I've got others around the house, but for tonight, I'm just not going to go through the headache. (But I'm sure I will tomorrow. :-) )

Pierre has offered me ownership of the app and I asked him to release it in the same manner Ben and others have with their apps under a GPL license. I'm sure he will when he's complete with it, and I'm hoping he'll send a copy to Ben Despres for inclusion on Ben's Site. That way, everyone can poke around in the guts of the thing, and maybe make even more improvements.

I've put a sort of screen shot for the monitoring program Here . It is showing the 6 clients running during the best test of the night. (Don't forget, that a 7th copy is running as the host.)

This really is the biggest news in my end of the quest in quite a number of years. I really hope everyone agrees with me that Pierre has done an amazing job, and deserves a huge congratulations!!

And for the record, I'm still using Eric Goldstein's app, right now, until there is a bit more polish on this new distributed app.