In an effort to raise the double check level for the Prime Sierpinski Project, an n=2M-3M range has been inserted. These WU's are substantially lower than the leading edge work that we are doing at the n=9M level.

The WU's were completed a couple years ago and already have a residue. PrimeGrid's goal will be to see if those residues are valid, identify which ones are not, and retest. A "dummy" task has already been created. Therefore, only a single test will be sent out. If the residue matches, then the WU will be complete. If it does not match, PG will continue to send out tasks until matching residues are returned.

It is our hope that there's a missed prime or two at the lower ranges. If a prime is found, it would definitely facilitate our work at the 9M level.

We'll continue first pass tests at the 9M level...just be prepared to see DC work every now and then as a little quality control is conducted.

Good Luck!
____________

The first 1000 WU's out of 14,000 have gone out. However, we'd like to move a little faster than that. If you have a core or two to spare, please consider switching to PSP LLR for a day or so.

The DC WU's are only 2-3 hours long compared to the 60-80 hours of "normal" PSP LLR WU's. Therefore, we should be able to drain the queue pretty fast.

Thank you for your consideration.

John
____________

So there is a hope to find a mega prime, am i right?...if a prime will be found, who'll be the initial finder?

So there is a hope to find a mega prime, am i right?...if a prime will be found, who'll be the initial finder?

The DC work for PSP LLR has ramped up rather nicely. Let's see if we can empty the queue to 3M before the end of the year. :)
____________

The DC work for PSP LLR has ramped up rather nicely. Let's see if we can empty the queue to 3M before the end of the year. :)

Seems like those residue-checks are necessary: http://www.primegrid.com/workunit.php?wuid=98910499

Yes - I was going to post about this as well... http://www.primegrid.com/workunit.php?wuid=99139024
____________

And others:

http://www.primegrid.com/workunit.php?wuid=98912416
http://www.primegrid.com/workunit.php?wuid=99132574

http://www.primegrid.com/workunit.php?wuid=98911920

will there be any stats on how many residues were wrong from the first pass?
____________

We passed the 1/2 way point several hours ago. Not too bad for starting the DC range less than 4 days ago. :) Well done everyone. Now let's see if we can finish the second half faster than the first.

As for stats, we are currently monitoring all tests that require a third check. I'm sure that once all DC work is complete, we can provide stats for how many of the original residues were mismatched.
____________

New Year's Eve Blue Moon

If you needed a reason to help out the DC effort before the New Year, here's the Blue Moon 24 hour mini Challenge lasting from now until the New Year.

A Blue Moon happens infrequently...and a New Year's Eve blue moon happens even less. Last one occurred 20 years ago in 1990.

So, head over to your PrimeGrid preferences page and select PSP LLR. Then leave your computers and put on your party hats because it's time to celebrate the end of 2009 and the beginning of 2010. :)

p.s. The goal??? To get "unprepared tasks" for the 2M-3M range to below 4000 by the end of the year.
____________

I put my E6550 and my W3520 on this project too. That must be around 50 GFLOPS from me now.
As soon as the Q9550 is through with the manual ppse-sieve (1040 on 2009-12-31) he is on too (another 10.5 GFLOPS)...
I think the 4k until 2010-01-01 00:00 UTC is achievable.
____________

Unfortunately, I got many of those long resends instead of short doublechecks, so I don't have enough free cores. :|
____________

Happy New Year!

Based on Meebo postings, it looks like 4000 "unprepared tasks" was reached before the new year. Thanks to everyone for making this possible. As of this post, we are already down to 3111. Looks like another day or two and the 2M-3M queue will be empty.

An unintended side effect has been the high completion rate of 8M-9M resends. :) Thank you for tolerating this.
____________

Yes, happy new year!

Only one long resend so far.
All others are short double checks!
And after finishing the Cullen-llr-tasks there are two more cores crunching the resends ...


____________
Sysadm@Nbg
my current lucky number: 3749*2^1555697+1
PSA-PRPNet-Stats-URL: http://u-g-f.de/PRPNet/

http://www.primegrid.com/workunit.php?wuid=99132386

what about this?
my wingman and me are conclusive and the formerly test is "marked as invalid"
was it a false prime before ??
can you explain ???

thx!
____________
Sysadm@Nbg
my current lucky number: 3749*2^1555697+1
PSA-PRPNet-Stats-URL: http://u-g-f.de/PRPNet/

The 2M-3M is almost complete. However, we must now wait for all the results to be returned. Once the 2M-3M range has been analyzed, we'll post the results and make a decision on how best to move forward. Until then, we will release some double check work in the 3M-4M range.

Thanks to everyone who participated in the 2M-3M double check effort. It's a great accomplishment. As for the 8M-9M side effect, we are over 94% complete now. That's an excellent side effect to have. :)

You can continue to follow progress here: PSP Range Statistics
____________

Double Check Update

All was going well until we added the new 3.0M-3.2M range. Then an oversight allowed about about 1400 WU's to be released with no "dummy" tasks. It has now been corrected. We apologize for the extra work.
____________

Double Check Update

It appears as though the DC effort for 2M-3M will take some time to complete all the WU's. With a 3 week deadline, there's not much we can do about what's been released. However, we did make a change recently to the new DC WU's going out. Deadline has been dropped to 7 days. This should help us close these newer ranges faster.

3.2M-3.5M was released a few days ago. Today, 3.5M-3.8M has been released. With the delay of the 2M-3M range, we'll at least go to 4M and probably up to 5M if we continue on this fast pace. This puts a completion date at around the end of January. Quite a remarkable accomplishment!

You can continue to follow progress here: PSP Range Statistics
____________

Double Check Update

The 2M-3M DC range continues to drag its feet. Looks like the 3 week deadline will definitely take its toll. We may actually complete the 3M-4M range before the 2M-3M since it was released with a 1 week deadline. :)

3.8M-4M has been uploaded. Progress is still moving quite nicely. We are still averaging about 1000 WU's a day and on pace with completing up to 5M by the end of January (11K WU's remain in the 4M-5M range).

You can continue to follow progress here: PSP Range Statistics

Once again, thank you to everyone participating in this effort.
____________

Double Check Update

The 2M-3M DC range remains open. The 3M-4M range is catching up quick with its 1 week deadline. As of this post, there are just over 1200 tests remaining in the 3M-4M range.

4M-5M has been uploaded. This concludes the DC work we'll be doing right now. Once that is complete, we'll return to first pass which will be in the 9M+ range.

Progress has slowed considerably. We are now averaging about 500 WU's a day (1/2 of what we were doing). The end of January goal will definitely be exceeded unless a farm or two or three (or a team or two) move over for a week. hint, hint! :)

Nevertheless, a tremendous amount of work is being accomplished. You can continue to follow progress here: PSP Range Statistics

Once again, thank you to everyone participating in this effort.
____________

Question here when the WU says psp_llr_42103977_4 does that mean that its a 4M double check?
____________


John M. Johnson "Novex"

No - the 4 means that it's the 4th WU generated to get that prime tested.

In terms of the PSP DC effort, that means the 1st one is the residue imported from PSP, 2 PG WUs that have errored or timed out, and you've got the 4th (alternatively, one or more of the prior PG WUs was completed but the residue didn't match, so another got generated).

____________

Double Check Update

Excellent progress was made today. Over 1K WU's were completed. There's still an outside chance that we could achieve the 31 Jan goal of having all "unprepared tasks" sent out.

Thank you to everyone participating in this effort.

You can continue to follow progress here: PSP Range Statistics
____________

Double Check Update

The higher n is taking its toll on progress. Not to mention another FFT crossover. Nevertheless, we're still moving forward.

The bittersweet news is that no primes have been found, yet. :) While the complete error rates have not been determined, this holds well for the integrity of first pass.

It has been determined to advance the DC effort to 6M (EDIT: 7775 tests). The reason for this is that 0.5M of the 5M-6M range has already been completed and double checked. Therefore, only the other half remains. This will bring the DC to within 3M of the first pass.

This is definitely the last extension for now. After the DC effort is complete to 6M, we will return to first pass work in the 9M range.

As always, thank you to everyone participating in this effort.

You can continue to follow progress here: PSP Range Statistics
____________

Double Check Update

The DC focus took a month off, but now that the Tour de Primes is over, we'd like to return to this effort and clear out the remaining tests. Since the last update, the 5M-6M range (7775 candidates) has been inserted. Even with the Tour, good progress was made. There are currently just over 1800 candidates left. It is our goal to zero out the "unprepared tasks" buffer by the end of the week for this range and return to first pass work.

On a good note, the 3M-4M range has been completed and we're only 8 tests away from finishing all work below 5M.

As always, thank you to everyone participating in this effort.

Double Check Update

It took a day longer than our goal but the good news, ALL "unprepared tasks" for the DC effort have been sent out. We are now back to first pass in the 9M-10M range.

Thank you to everyone who participated in this incredible DC effort. While no missed primes have been found yet, the project can confidently move forward to the higher ranges.

We're getting DC WUs again.
Were the lowest 2 of the k's not DC'd last time around?

____________

What proportion of first-pass WUs are being done by PG (given that they are already DC'd) versus the main PSP project?
____________

What proportion of first-pass WUs are being done by PG (given that they are already DC'd) versus the main PSP project?

I had one of these previously in the previous bout of DCing...

Given the low (ie 0) error rate on my host, I guess this WU has a bad residue (or was computed with a different program):

http://www.primegrid.com/workunit.php?wuid=178271359
____________

As ltd mentions, the first round of DC went up to 6M. The goal is to bring the second round up to at least 8M. Here's what's remaining for DC:

n < 7M : 4500 candidates
7M < n < 8M: 5280 candidates
8M < n < 9M: 7080 candidates
9M < n < 10M: 2416 candidates