Why does it take 30 to 40 minutes before Progress begins.


Message boards : Number crunching : Why does it take 30 to 40 minutes before Progress begins.

Message board moderation

To post messages, you must log in.
AuthorMessage
Kenny Frew

Send message
Joined: 16 Feb 13
Posts: 3
Credit: 326,040
RAC: 0
Message 909 - Posted: 22 Feb 2013, 21:27:11 UTC
That leaves 20 minutes for crunching. And if for any reason it is interupted, by my using my computer (heavy ram use) or restarting boinc it starts all over again and will take another 30 min. to begin crunching.
ID: 909 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
Dagorath

Send message
Joined: 16 Aug 12
Posts: 293
Credit: 1,116,280
RAC: 0
Message 921 - Posted: 26 Feb 2013, 0:26:23 UTC - in response to Message 909.  
I was hoping Kyong or one of the devs would answer your question but I guess they're too busy. I don't know exactly why it takes that long for Progress to begin but I like to think they are trying to fix that. Do realize it may not be fixable. In the meantime all you can do is try your best to avoid interrupting Asteroid tasks.
ID: 921 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
Kenny Frew

Send message
Joined: 16 Feb 13
Posts: 3
Credit: 326,040
RAC: 0
Message 923 - Posted: 26 Feb 2013, 5:44:20 UTC - in response to Message 921.  

Last modified: 26 Feb 2013, 5:46:19 UTC
Yeah, I was hoping to hear something but i'm glad to know i'm not the only one. I have to avoid high cpu and ram use during tasks and i have changed my interval between tasks to 90 minutes, so i get about an hour in at a time.
ID: 923 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
Dagorath

Send message
Joined: 16 Aug 12
Posts: 293
Credit: 1,116,280
RAC: 0
Message 924 - Posted: 26 Feb 2013, 7:23:58 UTC - in response to Message 923.  
If you highlight a task then click Properties it tells you the CPU time at the last checkpoint. I've never really paid much attention to that number but I presume it stays 0:00 until the task checkpoints. You could check and verify that by watching a task or two. If that's the way it works then you have an easy way of telling whether or not you can do whatever is you do that causes the problems. If the number is still 0:00 then you need to wait, if it's not 0:00 then it has checkpointed.

Another option is to setup a couple of scripts that cause your Asteroids tasks to run only at night while you sleep and or while you're at work. Suppose you get home from work at 5:00 PM, roughly. And suppose an Asteroids task requires 1.5 hours to complete on your system. You could setup a script to run automatically every day at 3:30 PM that would prevent further Asteroids tasks from downloading or starting while allowing any started ones to proceed. If you go to bed at midnight then you could have another script run automatically at that time to suspend all other projects which should force any Asteroids tasks to start, then unsuspend the other projects. Then Asteroids could/would run until 3:30 PM. This assumes whatever happens that causes the interruptions and problems does not happen automatically on its own while you sleep or work. You could setup a similar schedule for weekends, perhaps allow Asteroids to run only between 1 AM and 9 AM, for example.
ID: 924 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
Josef Durech
Project administrator
Project developer
Project tester
Project scientist

Send message
Joined: 18 Jun 12
Posts: 29
Credit: 5,504,007
RAC: 0
Message 925 - Posted: 26 Feb 2013, 11:19:29 UTC - in response to Message 909.  
The reason is that every task starts with an initialization when several control parameters are tested and then set up for the rest of the computing. During this process, there is no output and no checkpointing. As it takes less than 1/100 of the total time, I thought it wouldn't cause any practical problems.
ID: 925 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
Dagorath

Send message
Joined: 16 Aug 12
Posts: 293
Credit: 1,116,280
RAC: 0
Message 926 - Posted: 27 Feb 2013, 4:14:40 UTC - in response to Message 925.  
As it takes less than 1/100 of the total time


Kenny Frew has 4 completed tasks in his stats and the longest one required ~80,000 secs. 1/100 of that is 800 secs or ~13 minutes. I wonder why he claims 30 to 40 minutes of no progress? Sorry, I haven't checked the tasks carefully on my own hosts yet, too busy, but I do know they run for a few minutes with no progress report.

Checkpointing and progress estimates can be very difficult. If it's not possible during the initialization phase you mention then it's not possible. But if you can manage some sort of checkpointing it would be nice. Progress reports aren't critical but checkpointing is very important for some volunteers as their systems are up, down, on and off many times a day.
ID: 926 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
Profile BilBg
Avatar

Send message
Joined: 19 Jun 12
Posts: 221
Credit: 623,640
RAC: 0
Message 939 - Posted: 2 Mar 2013, 6:56:13 UTC - in response to Message 925.  

Last modified: 2 Mar 2013, 7:01:18 UTC
As it takes less than 1/100 of the total time, I thought it wouldn't cause any practical problems.


For me it takes more like 1/16 (15000/900) as I posted here:
http://asteroidsathome.net/boinc/forum_thread.php?id=68

And it happens not only on (first) start but also after restart of tasks from checkpoint (e.g. after BOINC exit/start)
(so with many restarts this may take >50% of the time (especially for people using 'Suspend work while computer is in use'))

Or like this case (two of Kenny Frew's tasks):
http://asteroidsathome.net/boinc/result.php?resultid=1452452
http://asteroidsathome.net/boinc/result.php?resultid=1452348

Stderr output
<core_client_version>7.0.28</core_client_version>
<![CDATA[
<stderr_txt>
20:36:37 (3624): No heartbeat from core client for 30 sec - exiting
20:39:46 (3564): No heartbeat from core client for 30 sec - exiting
21:10:10 (788): No heartbeat from core client for 30 sec - exiting
21:13:17 (3688): No heartbeat from core client for 30 sec - exiting
21:25:34 (704): No heartbeat from core client for 30 sec - exiting
21:46:44 (1796): No heartbeat from core client for 30 sec - exiting
11:04:53 (3804): called boinc_finish

</stderr_txt>





- ALF - "Find out what you don't do well ..... then don't do it!" :)
ID: 939 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
Paul and kirsty yates
Avatar

Send message
Joined: 25 Feb 13
Posts: 2
Credit: 1,363,596
RAC: 0
Message 942 - Posted: 4 Mar 2013, 20:24:37 UTC
hope someone can help
my pc last checkpointed at 2hr 51min
now running 7hrs 30 mins
only 12% done
also the stderr.txt file is empty ,is this right ?
should i dump this w/u so someone can try to compleat it ?
i have returned a compleat w/u so i know i can finish them
ID: 942 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
Paul and kirsty yates
Avatar

Send message
Joined: 25 Feb 13
Posts: 2
Credit: 1,363,596
RAC: 0
Message 943 - Posted: 4 Mar 2013, 22:05:24 UTC
seems to be checkpointing now every hour
weird will leave it to compleat
ID: 943 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
Profile GALAXY-VOYAGER

Send message
Joined: 11 Jan 13
Posts: 6
Credit: 163,920
RAC: 0
Message 1577 - Posted: 25 Aug 2013, 13:04:26 UTC - in response to Message 921.  
This is only a possibility with your situation, but it could be possible that it is because your Computer is Running Sheduled Updates and/or Virus Scan. Or even something like Ram Usage. You may not have sufficient RAM to Operate at Full Capacity when Multiple Things are Happening.
I used to have only 2GB RAM in my HP Notebook (Standard when I purchased it), which I only Run SETI on: All my Other Projects are on my Desktop. This was barely enought to Run Windows without anything else in a Comfortable Way. I Upgraded the RAM to 8GB and instead of Running at Average 60%-80% RAM Usage, with only Windows & SETI Running, it now Runs around 30%-40% RAM Usage. Which leaves plenty of room to Run Other Programs, even with SETI Still Running 2 Tasks at once.
So it may be worth checking out. You could also check you STARTUP Menu and Check which Programs ans Apps are Set To Start on Login. Remove any that are not needed to Run The System.
Another thing is ..... It may be a FRAGMENTATION Problem. Set your System to DEFRAG at least once a week. This could also be interferring if it happens to be Heavily Fragmented and takes excessive time to Defrag, and could be happening at the times you are experiencing problems.
Good Luck :-)
ID: 1577 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
Profile GALAXY-VOYAGER

Send message
Joined: 11 Jan 13
Posts: 6
Credit: 163,920
RAC: 0
Message 1912 - Posted: 10 Oct 2013, 4:40:16 UTC - in response to Message 921.  
Leaving Only 20 Minutes To Cruch ... Hmmmnn .. That doesn't make any sense. How long have the Tasks been sitting Idle after being Downloaded?
I currently have 2 Asteroids Tasks (1 Running and 1 Ready To Start). The Running Task only has about 32 Minutes Remaining. Both Tasks are what remains of 11 Asteroids Tasks which I was Allocated around 1:50am This Morning (AEST): The other 9 have been Completed and Uploaded.
Both Remaining Tasks have Deadlines of 20th October, so I can't understand why you only have 20 Minutes Crunching Time.
It could be that You are Crunching Newer Tasks First. Make sure that The Deadline Tab Column is Listed In Descending Date Order (ie Current or earliest Date at The Top). To change it, Click on the Tab.
I have found that if the List id running the other way, The Tasks are Crunched out of Order.

Good Luck.
ID: 1912 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote

Message boards : Number crunching : Why does it take 30 to 40 minutes before Progress begins.