PrimeGrid
Please visit donation page to help the project cover running costs for this month

Toggle Menu

Join PrimeGrid

Returning Participants

Community

Leader Boards

Results

Other

drummers-lowrise
1) Message boards : General discussion : Problem loading compatible applications in Windows XP x64 (Message 140594)
Posted 2 days ago by Profile Michael GoetzProject donor
Different applications require different build procedures. We did in fact stop supporting XP-64 because we could not build one of our applications in such a way that it would run on both XP-64 and newer versions of Windows with the latest compilers. And before that, we had to jump through a lot of hoops to make XP-64 work. It was always the “problem child” of Windows.

I’m surprised that any of the apps still work on XP-64.
2) Message boards : Seventeen or Bust : Proposal: A new Sierpiński problem (Message 140589)
Posted 2 days ago by Profile Michael GoetzProject donor
Will these numbers be tested for Fermat divisibility?


Since the values of k are fairly high here, I think the chance of finding a Fermat divisor is quite low. I'll start a test on the new k = 68633 prime anyway and see if anything pops up.


I agree, the chances are minuscule. But we might as well test them, since it is only twelve numbers. /JeppeSN


Absolutely.
3) Message boards : General discussion : Problem loading compatible applications in Windows XP x64 (Message 140583)
Posted 2 days ago by Profile Michael GoetzProject donor
Please see this forum thread: https://www.primegrid.com/forum_thread.php?id=5578

Due to the inability to build software that would work on XP-64, we stopped supporting that operating system six years ago.

There's nothing we can do on our side. If you really want to use that computer here, I can think of few ways of circumventing the problem:

1) Using VBox or VMWare Player (both free), run a Linux VM under Windows XP, and run BOINC inside that VM. It will be able to run PrimeGrid tasks. I recommend using a 64 bit Linux distribution, but 32 bit will work too.

2) Modify the source code for the BOINC client, and build your own version that reports itself as running under 32 bit XP instead of 64 bit XP.

3) Use app_info.xml to supply your own LLR executable, and run the 32 bit version.

EDIT: I'm not sure whether or not the 32 bit code will work under XP-64. If it doesn't, then #2 and #3 above won't work. The only option that is sure to work is #1.
4) Message boards : Seventeen or Bust : Proposal: A new Sierpiński problem (Message 140572)
Posted 3 days ago by Profile Michael GoetzProject donor
Almost all pending tests has been returned, but 53 are still in progress.


7 of those are mine. They'll be completed in a few hours.
5) Message boards : Number crunching : many PPSsieve units are pending? (Message 140571)
Posted 3 days ago by Profile Michael GoetzProject donor
In case you weren't aware, the number of pending tasks is entirely dependent upon what other users are doing. High or low pending numbers have nothing to do with what's happening on the server.


Yes, but 25 just today (UTC) waiting for validation (35 total) is much higher than normal. So, how is the server?


You're waiting for your wingmen to return their tasks.

Likely reason for having a wingman problem right now:

There's a user who, possibly because of incorrect settings on BOINCStats BAM!, has about 60 thousand in-progress PPS Sieve tasks. To put this in perspective, there's normally about 20 to 30 thousand in progress PPS Sieve tasks TOTAL across the entire system. He's got three times as many tasks as everyone else combined. You (and everyone else) is waiting either for this guy's computers to complete the tasks, or for them to time out and and be sent to someone else.

Nothing is wrong, but you will need to wait a bit longer for these tasks to be validated.
6) Message boards : Number crunching : many PPSsieve units are pending? (Message 140558)
Posted 3 days ago by Profile Michael GoetzProject donor
In case you weren't aware, the number of pending tasks is entirely dependent upon what other users are doing. High or low pending numbers have nothing to do with what's happening on the server.
7) Message boards : General discussion : Please dont "Shut up" :) (Message 140549)
Posted 3 days ago by Profile Michael GoetzProject donor
When you say a ways to go, do you have any rough estimates on how long taking into consideration how many GFN-17 low are completed on average....also could we speed it up with a concentrated effort from a lot of us and would this get us moving onto new things faster or would it still be a very long way off? I like the thoughts Michael put forward actually and it sounds quite interesting


Finishing GFN-17 (or, really, any other project) has nothing to do with the work necessary to start other projects, so no, that won't get you moving any faster on new stuff.

GFN-17 was started almost exactly 5 years ago.

It will end at b=42,597,774.

We're currently at b=19,967,566, or 47%.

Based on that, you could say we have about 5 years to go, except that computers keep getting faster. My guess would therefore be somewhere around 2 to 3 years.

17-Low isn't as popular as 17-MEGA, for obvious reasons. The MEGA version has done about 50% more than the LOW version.
8) Message boards : Number crunching : Strange Behaviour (Message 140509)
Posted 6 days ago by Profile Michael GoetzProject donor
I think someone (possibly Mike) somewhere(possibly on this forum) addressed this problem before too, but why does the BOINC Client/Manager abort tasks according to their initial deadline instead of PG's or Stream's server's extended deadline? About 30 of Jeppe's problem's tasks were aborted, and at 7:26PM (11:26AM UTC, 1hr8min later) another around 20 GFN14 DC tasks will be aborted. Can there be any changes made to this?

Because these tasks were unstarted. Nothing will be lost aborting them, so both client and server a free to re-issue them.

Deadline extension works only work for in-progress tasks (they will send a special keep-alive message to server).


Really? That's strange. On my task page on your server it says "1st Jun 9:42:55 UTC" for those aborted LLR2 tasks, but on the BOINC Manager they say 30th May 5:42:55PM local time, which is 30th May 9:42:55 UTC. How does that happen?


The tasks have a 7-day deadline -- that's May 30th, and that's the ONLY date your computer knows.

On the server side, there's a global "grace period" setting which stream apparently also set to 7 days. The grace period is the amount of time, *after* the deadline, that the server waits before sending out a replacement task. The purpose is to prevent the server from needlessly sending out duplicate tasks in case the host computer is just a little bit late. (FYI, the PrimeGrid server is set to have a 25 hour grace period.)

What happened is that your tasks passed the deadline on May 30th (7 days), and the BOINC client did what it's supposed to do and aborted any expired tasks that had not yet been started. The date you see on the website isn't really the deadline; it's the date at which the server will send out a replacement task. It's intentionally later than the deadline.
9) Message boards : Number crunching : Strange Behaviour (Message 140503)
Posted 6 days ago by Profile Michael GoetzProject donor
I think someone (possibly Mike) somewhere(possibly on this forum) addressed this problem before too, but why does the BOINC Client/Manager abort tasks according to their initial deadline instead of PG's or Stream's server's extended deadline? About 30 of Jeppe's problem's tasks were aborted, and at 7:26PM (11:26AM UTC, 1hr8min later) another around 20 GFN14 DC tasks will be aborted. Can there be any changes made to this?


Any tasks that have not been started will be automatically aborted by the BOINC client on your computer when they exceed the deadline. The client has no knowledge of the deadline on the server. It only knows about the original deadline it was sent when it received the task.
10) Message boards : Seventeen or Bust : Proposal: A new Sierpiński problem (Message 140449)
Posted 12 days ago by Profile Michael GoetzProject donor
Also, it seems you can still get new tasks. You just can't upload the results to the server.


Next 10 posts
[Return to PrimeGrid main page]
DNS Powered by DNSEXIT.COM
Copyright © 2005 - 2020 Rytis Slatkevičius (contact) and PrimeGrid community. Server load 1.28, 1.54, 1.82
Generated 6 Jun 2020 | 10:45:01 UTC