Welcome to MilkyWay@home

cache = 0 WU

Message boards : Number crunching : cache = 0 WU
Message board moderation

To post messages, you must log in.

1 · 2 · Next

AuthorMessage
Alessandro Freda

Send message
Joined: 11 Apr 08
Posts: 11
Credit: 43,597,062
RAC: 0
Message 35537 - Posted: 12 Jan 2010, 11:55:44 UTC

Yesterday after weeks of normal behavior (cache = 24 WU) without any change my client side, the server start to send only 1 WU and with a waiting time of about 1 min. between result upload and WU download. This cause continue start and stop of the Collatz that cause also a slow down on its WU (double the GPU time per WU).

So I've reset the project leaving the optimized .20b 64bit intel application (running on a 4770), and try the standard application, but at this point no more WU.

I've detached the project waiting for a solution.
Any idea ?
ID: 35537 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
Profile The Gas Giant
Avatar

Send message
Joined: 24 Dec 07
Posts: 1947
Credit: 240,884,648
RAC: 0
Message 35569 - Posted: 12 Jan 2010, 19:58:33 UTC - in response to Message 35537.  

Yesterday after weeks of normal behavior (cache = 24 WU) without any change my client side, the server start to send only 1 WU and with a waiting time of about 1 min. between result upload and WU download. This cause continue start and stop of the Collatz that cause also a slow down on its WU (double the GPU time per WU).

So I've reset the project leaving the optimized .20b 64bit intel application (running on a 4770), and try the standard application, but at this point no more WU.

I've detached the project waiting for a solution.
Any idea ?

Why detach?

Have you set your preferences to use GPU?
ID: 35569 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
Alessandro Freda

Send message
Joined: 11 Apr 08
Posts: 11
Credit: 43,597,062
RAC: 0
Message 35570 - Posted: 12 Jan 2010, 20:24:49 UTC - in response to Message 35569.  
Last modified: 12 Jan 2010, 20:25:15 UTC

Yesterday after weeks of normal behavior (cache = 24 WU) without any change my client side, the server start to send only 1 WU and with a waiting time of about 1 min. between result upload and WU download. This cause continue start and stop of the Collatz that cause also a slow down on its WU (double the GPU time per WU).

So I've reset the project leaving the optimized .20b 64bit intel application (running on a 4770), and try the standard application, but at this point no more WU.

I've detached the project waiting for a solution.
Any idea ?

Why detach?

I've attach again without behaviour change.

Have you set your preferences to use GPU?

Yes. There was no change in the preferences when the server start to send the message "No work sent". Now I've try some mods without success. Is it possible continue to use the app_info.xml to send cmdline to the old 0.20b release ?
ID: 35570 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
Profile banditwolf
Avatar

Send message
Joined: 12 Nov 07
Posts: 2425
Credit: 524,164
RAC: 0
Message 35572 - Posted: 12 Jan 2010, 20:40:51 UTC - in response to Message 35570.  

Is it possible continue to use the app_info.xml to send cmdline to the old 0.20b release ?

You can still use the opti app instead of the dl. I'm sure changes need to be made to use the app file for the dl.
Doesn't expecting the unexpected make the unexpected the expected?
If it makes sense, DON'T do it.
ID: 35572 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
Alessandro Freda

Send message
Joined: 11 Apr 08
Posts: 11
Credit: 43,597,062
RAC: 0
Message 35573 - Posted: 12 Jan 2010, 21:29:07 UTC - in response to Message 35572.  

BTW these are the logs, now seems that no WUs are requested:

12/01/2010 22:05:40 ATI GPU 0: ATI Radeon HD 4700/4800 (RV740/RV770) (CAL version 1.4.515, 512MB, 1062 GFLOPS peak)
12/01/2010 22:05:40 Milkyway@home Found app_info.xml; using anonymous platform
12/01/2010 22:05:40 Milkyway@home URL http://milkyway.cs.rpi.edu/milkyway/; Computer ID 130238; resource share 100
12/01/2010 22:05:40 Milkyway@home General prefs: from Milkyway@home (last modified 12-Jan-2010 20:40:08)
12/01/2010 22:05:40 Milkyway@home Host location: none
12/01/2010 22:05:40 Milkyway@home General prefs: using your defaults
12/01/2010 22:05:40 Reading preferences override file
12/01/2010 22:05:40 Preferences limit memory usage when active to 1842.24MB
12/01/2010 22:05:40 Preferences limit memory usage when idle to 1842.24MB
12/01/2010 22:05:40 Preferences limit disk usage to 43.94GB
12/01/2010 22:05:40 file projects/milkyway.cs.rpi.edu_milkyway/parameter_222F5_3s_v2.txt not found
12/01/2010 22:05:40 file projects/milkyway.cs.rpi.edu_milkyway/stars222F5 not found
12/01/2010 22:05:42 Milkyway@home Started download of parameter_222F5_3s_v2.txt
12/01/2010 22:05:42 Milkyway@home Started download of stars222F5
12/01/2010 22:05:44 Milkyway@home Finished download of parameter_222F5_3s_v2.txt
12/01/2010 22:06:15 Milkyway@home Finished download of stars222F5
12/01/2010 22:07:52 Milkyway@home update requested by user
12/01/2010 22:07:56 Milkyway@home Sending scheduler request: Requested by user.
12/01/2010 22:07:56 Milkyway@home Not reporting or requesting tasks
12/01/2010 22:08:07 Milkyway@home Scheduler request completed
12/01/2010 22:08:07 Milkyway@home General prefs: from Milkyway@home (last modified 12-Jan-2010 22:04:26)
12/01/2010 22:08:07 Milkyway@home Host location: none
12/01/2010 22:08:07 Milkyway@home General prefs: using your defaults
12/01/2010 22:08:07 Reading preferences override file
12/01/2010 22:08:07 Preferences limit memory usage when active to 1842.24MB
12/01/2010 22:08:07 Preferences limit memory usage when idle to 1842.24MB
12/01/2010 22:08:07 Preferences limit disk usage to 43.94GB
ID: 35573 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
Profile The Gas Giant
Avatar

Send message
Joined: 24 Dec 07
Posts: 1947
Credit: 240,884,648
RAC: 0
Message 35591 - Posted: 13 Jan 2010, 10:00:17 UTC - in response to Message 35570.  

Yesterday after weeks of normal behavior (cache = 24 WU) without any change my client side, the server start to send only 1 WU and with a waiting time of about 1 min. between result upload and WU download. This cause continue start and stop of the Collatz that cause also a slow down on its WU (double the GPU time per WU).

So I've reset the project leaving the optimized .20b 64bit intel application (running on a 4770), and try the standard application, but at this point no more WU.

I've detached the project waiting for a solution.
Any idea ?

Why detach?

I've attach again without behaviour change.

Have you set your preferences to use GPU?

Yes. There was no change in the preferences when the server start to send the message "No work sent". Now I've try some mods without success. Is it possible continue to use the app_info.xml to send cmdline to the old 0.20b release ?

I hate to press about this, but since the server software upgrade have you checked your preferences on the web page?

In BOINC Manager have you set the preference to Use GPU while computer is in use?
ID: 35591 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
Alessandro Freda

Send message
Joined: 11 Apr 08
Posts: 11
Credit: 43,597,062
RAC: 0
Message 35592 - Posted: 13 Jan 2010, 10:35:04 UTC - in response to Message 35591.  

Ok, these are my Computing preferences:

Suspend work while computer is on battery power?
(matters only for portable computers) no
Suspend work while computer is in use? no
Suspend GPU work while computer is in use?
Enforced by version 6.7+ no
'In use' means mouse/keyboard activity in last 3 minutes
Suspend work if no mouse/keyboard activity in last
(Needed to enter low-power mode on some computers) --- minutes
Do work only between the hours of (no restriction)
Leave applications in memory while suspended?
(suspended applications will consume swap space if 'yes') yes
Switch between applications every
(recommended: 60 minutes) 60 minutes
On multiprocessors, use at most 16 processors
On multiprocessors, use at most
Enforced by version 6.1+ 100 % of the processors
Use at most
(Can be used to reduce CPU heat) 100 percent of CPU time


Use at most 100 GB disk space
Leave at least
(Values smaller than 0.001 are ignored) 0.001 GB disk space free
Use at most 50% of total disk space
Tasks checkpoint to disk at most every 60 seconds
Use at most 75% of page file (swap space)
Use at most 90% of memory when computer is in use
Use at most 90% of memory when computer is not in use


Computer is connected to the Internet about every
(Leave blank or 0 if always connected.
BOINC will try to maintain at least this much work.) 0 days
Maintain enough work for an additional 0 days
Confirm before connecting to Internet?
(matters only if you have a modem, ISDN or VPN connection) no
Disconnect when done?
(matters only if you have a modem, ISDN or VPN connection) no
Maximum download rate: no limit
Maximum upload rate: no limit
Use network only between the hours of (no restriction)
Skip image file verification?
Check this ONLY if your Internet provider modifies image files (UMTS does this, for example).
Skipping verification reduces the security of BOINC. no
ID: 35592 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
Alessandro Freda

Send message
Joined: 11 Apr 08
Posts: 11
Credit: 43,597,062
RAC: 0
Message 35593 - Posted: 13 Jan 2010, 10:37:48 UTC - in response to Message 35592.  

I've a Radeon 4770 on a quad 9400S.

And these are the MilkyWay@home preferences:

Resource share
If you participate in multiple BOINC projects, this is the proportion of your resources used by MilkyWay@home 100
Use CPU
(enforced by 6.10+ clients) no
Use NVIDIA GPU if present
(enforced by 6.10+ clients) no
Use ATI GPU if present
(enforced by 6.10+ clients) yes
Is it OK for MilkyWay@home and your team (if any) to email you? yes
Should MilkyWay@home show your computers on its web site? no
Default computer location ---
Maximum CPU % for graphics
0 ... 100
ID: 35593 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
Profile kashi

Send message
Joined: 30 Dec 07
Posts: 311
Credit: 149,490,184
RAC: 0
Message 35594 - Posted: 13 Jan 2010, 11:51:06 UTC - in response to Message 35593.  
Last modified: 13 Jan 2010, 12:03:03 UTC

Since the server change when using the app_info.xml ATI 0.20b version you may need to set MilkyWay preferences "Use CPU (enforced by 6.10+ clients)" to yes and "Use ATI GPU if present (enforced by 6.10+ clients)" to no.

You may also need to adjust resource share and cache size of all projects you are currently attached to because the MilkyWay ATI v0.20 b application is now treated as a CPU for BOINC debt and work fetch.

I've found it easiest to download a large cache of CPU project work (1 day for example) then set no new tasks on CPU projects. Then adjust cache to 2 days to enable BOINC to continue downloading MilkyWay ATI work without BOINC debt buildup stopping ATI work download. Needs to be repeated daily. Depending on your particular choice of CPU projects, resource share and whether you process 1 or 2 ATI tasks concurrently on your GPU you may also need to use <report_results_immediately>1</report_results_immediately> and <zero_debts>1</zero_debts> in a cc_config.xml file as well.

You should also be aware that if you are requesting assistance it is impossible to give you the most relevant advice relating to your configuration when your computer is hidden. There are some that choose not to answer at all when people ask for help with their computer details hidden.
ID: 35594 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
Alessandro Freda

Send message
Joined: 11 Apr 08
Posts: 11
Credit: 43,597,062
RAC: 0
Message 35598 - Posted: 13 Jan 2010, 15:15:03 UTC - in response to Message 35594.  

Since the server change when using the app_info.xml ATI 0.20b version you may need to set MilkyWay preferences "Use CPU (enforced by 6.10+ clients)" to yes and "Use ATI GPU if present (enforced by 6.10+ clients)" to no.


This do not seems works.
Logs before mod:
13/01/2010 16:05:35 Milkyway@home Sending scheduler request: To fetch work.
13/01/2010 16:05:35 Milkyway@home Requesting new tasks for GPU
13/01/2010 16:05:40 Milkyway@home Scheduler request completed: got 0 new tasks
13/01/2010 16:05:40 Milkyway@home Message from server: No work sent

Log after the mod:

13/01/2010 16:05:49 Milkyway@home update requested by user
13/01/2010 16:05:50 Milkyway@home Sending scheduler request: Requested by user.
13/01/2010 16:05:50 Milkyway@home Not reporting or requesting tasks
13/01/2010 16:05:55 Milkyway@home Scheduler request completed

Seems that the client stop to request work.
BTW dont like waste CPU time on this project that can use the (more powerful) GPUs.
ID: 35598 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
Profile kashi

Send message
Joined: 30 Dec 07
Posts: 311
Credit: 149,490,184
RAC: 0
Message 35600 - Posted: 13 Jan 2010, 15:41:45 UTC - in response to Message 35598.  

It will still be processed on your GPU if your app_info.xml is correctly configured, even though BOINC requests it as a CPU task.

Try suspending all your CPU projects, if BOINC then downloads MilkyWay tasks you know it is a BOINC debt problem that is currently preventing ATI tasks from being requested and downloaded.

If it is a BOINC debt problem then you will need to start adjusting resource share of attached projects and cache size and doing whatever else is necessary to get BOINC to continue to download MilkyWay ATI tasks.

If it isn't a BOINC debt problem then I don't know what is causing it.
ID: 35600 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
Alessandro Freda

Send message
Joined: 11 Apr 08
Posts: 11
Credit: 43,597,062
RAC: 0
Message 35602 - Posted: 13 Jan 2010, 15:53:55 UTC - in response to Message 35600.  

Solved! :)

I've deleted these files from BOINC data dir:

sched_reply_milkyway.cs.rpi.edu_milkyway.xml
sched_request_milkyway.cs.rpi.edu_milkyway.xml
statistics_milkyway.cs.rpi.edu_milkyway.xml


restart BOINC => 24 WU downloaded immediately.

Notice that before failed a project reset and even a detach-attach.
ID: 35602 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
Alessandro Freda

Send message
Joined: 11 Apr 08
Posts: 11
Credit: 43,597,062
RAC: 0
Message 35603 - Posted: 13 Jan 2010, 15:53:58 UTC - in response to Message 35600.  
Last modified: 13 Jan 2010, 15:54:16 UTC

DEL
ID: 35603 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
Profile JockMacMad TSBT
Avatar

Send message
Joined: 28 Jan 09
Posts: 31
Credit: 85,934,108
RAC: 0
Message 35649 - Posted: 14 Jan 2010, 13:25:40 UTC
Last modified: 14 Jan 2010, 13:26:24 UTC

I am also seeing an issue with one machine not being allocated tasks. Other machines are fine as was this till 2 days ago. Now it never get tasks always returns 0 tasks sent.

I have reset and failing that I detached and reinstalled opti app.

All looks okay and the stars file gets downloaded but it still will not send WUs.

I have 10 days set on the network prefs so I was expecting to get 24 units.

I have since copied the folder from a running installation and the same. NADA.

BOINC client is 6.10.24 and as I say till 48 hours ago was running perfectly.
ID: 35649 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
Profile banditwolf
Avatar

Send message
Joined: 12 Nov 07
Posts: 2425
Credit: 524,164
RAC: 0
Message 35650 - Posted: 14 Jan 2010, 18:18:07 UTC - in response to Message 35649.  


I have 10 days set on the network prefs so I was expecting to get 24 units.


Is that the 'connect every x days'? If it is you need to be under 3 days since the deadline is 3 days for mw. If the deadline is actually extended to 8 days then you would need to be less than that. Boinc thinks your computer won't connect in time to report the wus.(I know Travis says the deadline is now 8 days, I have not seen a wu that reflected that yet) The buffer does not matter.
Doesn't expecting the unexpected make the unexpected the expected?
If it makes sense, DON'T do it.
ID: 35650 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
Bigred
Avatar

Send message
Joined: 23 Nov 07
Posts: 33
Credit: 300,042,542
RAC: 0
Message 35652 - Posted: 14 Jan 2010, 18:57:34 UTC - in response to Message 35593.  

I've a Radeon 4770 on a quad 9400S.

And these are the MilkyWay@home preferences:



Maximum CPU % for graphics
0 ... 100


This was the culprit on my systems. I reset it to 5 and everything started working.
ID: 35652 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
Profile JockMacMad TSBT
Avatar

Send message
Joined: 28 Jan 09
Posts: 31
Credit: 85,934,108
RAC: 0
Message 35862 - Posted: 20 Jan 2010, 19:06:19 UTC

I find the OPTI app causes the starvation to occur and once I revert to the stock app it starts running fine again.

I now have a second machine go the same way on the OPTI and this also now has been reverted to stock and it has resumed crunching.

No changes to any config were made and all my machines share the same app_info.xml for the OPTI I install from a USB key.

All very weird as some machines have been running for 6+ months and have just stopped.

No harm done though as at least reverting to stock is so easy and it works.
ID: 35862 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
Profile skildude
Avatar

Send message
Joined: 13 Jan 08
Posts: 52
Credit: 59,823,507
RAC: 0
Message 35992 - Posted: 24 Jan 2010, 1:55:41 UTC

I've tried these tricks and I'm still only getting 1 WU at a time. this is really annoying that my GPU stops when it finishes a WU then has to wait on a new one to come in. can my cache be set higher?
Blah blah blah you know the rest
ID: 35992 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
Profile Labbie
Avatar

Send message
Joined: 29 Aug 07
Posts: 327
Credit: 116,463,193
RAC: 0
Message 35993 - Posted: 24 Jan 2010, 2:33:53 UTC - in response to Message 35992.  

I've tried these tricks and I'm still only getting 1 WU at a time. this is really annoying that my GPU stops when it finishes a WU then has to wait on a new one to come in. can my cache be set higher?


What is your cache set at? If it is longer than three days (might be 5 now that the deadlie has been extended), BOINC worries that you won't make the deadline and only downloads one WU at a time.

As a test, lower you cache to 1 day and see what happens.



Calm Chaos Forum...Join Calm Chaos Now
ID: 35993 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
Profile skildude
Avatar

Send message
Joined: 13 Jan 08
Posts: 52
Credit: 59,823,507
RAC: 0
Message 35997 - Posted: 24 Jan 2010, 5:36:18 UTC - in response to Message 35993.  

did that and still nothing

I am getting this message on boinc
1/23/2010 11:27:47 PM Milkyway@home Message from server: (won't finish in time) BOINC runs 93.5% of time, computation enabled 57.0% of that
the number seems to be slowing rising. I assume this is the problem. With any luck it will fix itself by morning

Blah blah blah you know the rest
ID: 35997 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
1 · 2 · Next

Message boards : Number crunching : cache = 0 WU

©2024 Astroinformatics Group