Ooh...my bad again with this latest batch of results. Here's what it looks like has happened: I essentially popped the two straggling k/n pairs into the workfile.txt of an LLRnet client running on IB8000. Then, I let it run until those two stragglers were done. It finished them and submitted them to the server, and I presumed that it put them into that day's results file as it normally would. Yet, when I downloaded all the results to process them later that day, I forgot to download the "copy of current results.txt" file as well--and that's where the results would be (actually, as it turns out, that's where I *thought* they would be)--but, by the time I noticed all this, I was already almost done with processing everything. So, I figured...it's only two results, and I already have the results sitting around from the client that submitted the results earlier. So, I just took the results from the client's lresults.txt file and inserted them into the big results file from there--assuming that the server had accepted them correctly and inserted them into the 2/16 results file.

However, of course, we have now seen that they did not indeed go into the 2/16 results file. They're probably in the 2/17 results file--but, I guess the moral of the story is: never assume that a results is on the server unless you've actually downloaded it back from the server. :smile:

Gary, despite all this, the results that I sent you should still be complete and correct; I verified them in full against the original sieve file. All the results should be on the server, too, but possibly in different places than I had assumed. Feel free to verify them with the original sieve file to make sure, of course.

Regarding splitting up the ranges in n=20K pieces in the range table: ah, I should have realized there was a method to the madness. :rolleyes: Okay, then--in the future, I'll be sure to ask before making any changes like that. Should have done that in the first place. :smile:

Max :smile:

P.S.: Just to make sure that everything is indeed as I've projected in this post, later today I'll check all the recent results files to make sure those two results did indeed make it to the server. And yes, I'll write myself a note so I don't forget. :wink:

the 2 missing pairs are still missing!

the last pairs with n=448k in results_20090211_0000_IB_nplb_8000.txt.

so, please, put those 2 pairs in the I8000 server and let them processed there!

[quote=kar_bon;163304]the 2 missing pairs are still missing!

the last pairs with n=448k in results_20090211_0000_IB_nplb_8000.txt.

so, please, put those 2 pairs in the I8000 server and let them processed there![/quote]
Hmm...you're right. They do indeed seem to be missing. :unsure:

Given this: David, can you please stop the port 8000 server, insert these two lines immediately after the header line, and restart the server?
[I]1579 448383
1689 448383[/I]

Thanks,
Max :smile:

Done.

[quote=IronBits;163321]Done.[/quote]
Thanks! :smile:

Reserving n=475K-490K for port IB8000.

All results for n=400K-460K for this drive have now been processed. All primes were checked against the 1st post in this drive, the top-5000 site, the k=300-2000 page, and the 8th drive page. No missing or incorrect primes were found.

Excellent work everyone. This was a huge range with a lot of primes (127) to have no problems on!

Karsten, the only nitpick that I saw on your 8th drive page is that 1551*2^439313-1 is out of order from left to right on that line. You've usually been listing them from lowest to highest n-value from left to right so I just thought I'd point it out.


Gary

Reserving n=490K-500K for port IB8000...the final file for this drive! :smile:

Will this server be loaded again soon, or are you waiting for the sieve to complete?

It will be loaded up with n=500K-510K when my sieveing range up to P=15T is complete in ~2 days. At P=15T, I am breaking off the n=500K-600K range.

Once complete to n=500K, we will actually be completed with this drive, and I'll start a new 10th drive for the n=500K-1M range. At the current rate of processing, we have enough work in the server for 4-5 days so the timing should work out well.

When my sieving range is done and the new drive is started, I'm going to pour a bunch of cores on this drive to pick up our top-5000 primes output, although I must say that it has been quite excellent lately, especially for having no rallies.


Gary

Ian, I removed the n>500K prime from the 1st post of this drive. It is the 1st prime of the new 10th drive so I am showing it there.


Gary