Analysis of testing ranges for k's < 300

[gd_barnes]

MooooMoo asked for the Excel spreadsheet that I keep that shows where the testing of various k's < 300 are at vs. where they should be as well as their reservation status. It is attached.

I think it points out some things that need to be followed up on. What happened recently with Benson finding a prime on k=9 barely above n=1M is a good example. k=9 is one that has fallen well behind where it should be because it is one of several reserved by Jeffrey that have been dormant for many months. Note that all k's < 31 are at least at 1M in testing with the exception of k=5 that will be shortly. All k's < 17 are at 1.4M with these two exceptions. k=5 has been jump-started big time and k=9 needs the same.

Curtis and Anon, this might be something to use as a tool (not the holy grail) for deciding which k's to include in the 6th drive. Obviously there are many factors in deciding which ones to include in any drive.


Gary

[Cruelty]

Gary,
k=9 has been taken care of recently.

[gd_barnes]

That's what I'm saying, Cruelty. Benson is not associated with RPS and you see no 'Rieselprime' on his code. He coordinates with no one. He found the prime because Jeffrey's reservations have gone dormant for many months and haven't been released.

We have no idea what ranges that Benson has searched so we have to come up behind him and double-check everything. I checked an entire primeless gap from n=90K to n=500K on k=289 because Bo did the same thing to us by finding 2 primes between n=500K & 520K. I don't mind double-check work. I do it all the time. What I do mind is doing it when it shouldn't be necessary like is this case here. That is what stinks.

We need to get kickin' on following up on these dormant reservations instead of letting them languish, otherwise some of them like k=9 and k=61 will turn into another k=5.


G

[kar_bon]

yeah that's right, i mentioned it here yesterday: http://www.mersenneforum.org/showpos...&postcount=851
but nobody knows which range was tested because Benson don't comunicate with anybody!

PS: oh, gary was one minute faster!

[kar_bon]

what about another quick graphical overview of k<300:
some weeks ago i made a html-page with graphs of all k<300 with their ranges done.
this example is made by hand (i updated the data from today), but it could be done per script so every change will instantly available in this graph. there are some more information of a special k available. the picture shows as example k=29 (mouse hold over the green range). other infos are possible like last update or something else. here the k, the contributor, the range done and number of primes so far are displayed. also special primes could be shown by a red pixel with info like the prime n,...
perhaps a new idea for the 15k-page.
some ideas/opinions?
karsten

[lsoule]

That's a neat chart!

Though this statement

He found the prime because Jeffrey's reservations have gone dormant for many months and haven't been released.

is attributing reasons to Benson's motives that we don't know.

[gd_barnes]

Originally Posted by lsoule

That's a neat chart!

Though this statement


is attributing reasons to Benson's motives that we don't know.

No, no. You've attributed more to my statement than I intended. I in no way think Benson had any sinister ulterior motive at all to sabotage us in any way. What I attempted to imply is that he just picks k's that look good and that haven't been tested as far as they should, probably sees the range that has been tested here or on Prof. Keller's site, and starts testing. I was also attempting to say that had we been on the ball on k=9, it would not have been as easy for him to 'pick off' a juicy prime like that on a nice, low k.

If I were him now, I'd be taking serious looks at k=61, 5, 139, 67, and 75; the five k's that are the furthest behind in testing and all reserved by someone here at RPS. k=9 is further down the list but is still ripe because it's such a low k. It's behind even if he's tested it to n=1.1M, although not as much as the others.

Even though we're up past n=1M on k=5 now, it's still well behind and somebody may just beat us to the first prime for n>1M. But if we stay diligent and get it tested thru n=1.5M by the end of Feb. or sometime in March, then that chance is minimized greatly. Like I've alluded to in the k=5 reservation thread, I would bet that him or someone else already tried to pick us off on that one but we've been fortunate enough that there are most likely no primes left for n<1M. (I hope I'm wrong and someone finds one from the few ranges left to test.)

Larry, I think we need to follow up with Jeffrey and get him to release all or some of his k's. Would you agree? Let's make it much more difficult for the Benson's and Bo's of the world to pick us off like this on a reserved k. I'll be glad to post messages here, send him PMs, and/or attempt to find an Email for him at other sites. I'm on it if you guys say the word.

If we can get k=9 released, we can turn it into another k=5 and sick Sheep and his sieving, Curtis and his coordination and sieving calculations, and Carlos and his multiple fast cores, as well as others on it and quickly get it up to where it should be. Low k's are cool to test!


Thanks,
Gary

[gd_barnes]

Originally Posted by kar_bon

what about another quick graphical overview of k<300:
some weeks ago i made a html-page with graphs of all k<300 with their ranges done.
this example is made by hand (i updated the data from today), but it could be done per script so every change will instantly available in this graph. there are some more information of a special k available. the picture shows as example k=29 (mouse hold over the green range). other infos are possible like last update or something else. here the k, the contributor, the range done and number of primes so far are displayed. also special primes could be shown by a red pixel with info like the prime n,...
perhaps a new idea for the 15k-page.
some ideas/opinions?
karsten

Outstanding idea, Karsten! I'm all for it. I'd like to see a trend line added like I have on my graph, although I'm guessing that it wouldn't be too 'politically popular'.


Gary