Btw at the moment all BRP servers are down and we get simply no task at all.
BRP4 work distribution for GPUs is working OK as I write this.
Maybe you saw that the BRP 4 work generators were not running, this is normal in situations when a sufficient server side buffer of work has been generated. The work generators then start again once the "watermark" of prepared workunits is dropping below a certain limit.
Is it possible for the GPU tasks like BRP4 to download say 3 days worth of tasks in cache. It works for CPU tasks but I feel not for GPU.
Btw at the moment all BRP servers are down and we get simply no task at all.
Even when they say not working I still get work. BM explains this well in his post.
Quote:
Maybe you saw that the BRP 4 work generators were not running, this is normal in situations when a sufficient server side buffer of work has been generated. The work generators then start again once the "watermark" of prepared workunits is dropping below a certain limit.
Thanks BM I just learned something new here. Though I kinda guessed as I have not had any problems getting work.
Now for your question Hypernova.
Quote:
Is it possible for the GPU tasks like BRP4 to download say 3 days worth of tasks in cache.
Yes. It is posted some where in these forums but I will explain it here for ya. You need to go to Computing preferences http://einstein.phys.uwm.edu/prefs.php?subset=global and change 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 (max 10 days). This is your base work load min. to have in your cache. The other setting Maintain enough work for an additional is like a top end buffer to the min. Change 1 or both of these should give you enough work for up to atleast 10 days. We use this mostly when we know the servers are going to be down so we can keep working while they fix/upgrade and so forth but you can run this all the time just make sure you don't get too much work and miss the 14 day dead line.
PC setup MSI-970A-G46 AMD FX-8350 8 core OC'd 4.45GHz 16GB ram PC3-10700 Geforce GTX 650Ti Windows 7 x64 Einstein@Home
Is it possible for the GPU tasks like BRP4 to download say 3 days worth of tasks in cache. It works for CPU tasks but I feel not for GPU.
Btw at the moment all BRP servers are down and we get simply no task at all.
Even when they say not working I still get work. BM explains this well in his post.
Quote:
Maybe you saw that the BRP 4 work generators were not running, this is normal in situations when a sufficient server side buffer of work has been generated. The work generators then start again once the "watermark" of prepared workunits is dropping below a certain limit.
Thanks BM I just learned something new here. Though I kinda guessed as I have not had any problems getting work.
Now for your question Hypernova.
Quote:
Is it possible for the GPU tasks like BRP4 to download say 3 days worth of tasks in cache.
Yes. It is posted some where in these forums but I will explain it here for ya. You need to go to Computing preferences http://einstein.phys.uwm.edu/prefs.php?subset=global and change 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 (max 10 days). This is your base work load min. to have in your cache. The other setting Maintain enough work for an additional is like a top end buffer to the min. Change 1 or both of these should give you enough work for up to atleast 10 days. We use this mostly when we know the servers are going to be down so we can keep working while they fix/upgrade and so forth but you can run this all the time just make sure you don't get too much work and miss the 14 day dead line.
My configuration is already like this. But at the moment I am getting no GPU tasks, and when it was the case I always got only very few at a time, except if I forced manually multiple updates in which case I would get some atsks at each call.
Running other projects has a big effect on this. To make a easy by pass stop all new work from every project except E@H. You might need to clear out some work first. After you get the work you want let the other projects get new work again.
Quote:
except if I forced manually multiple updates in which case I would get some atsks at each call.
Sometimes you have to force the update to happen.
Edit: With more than 1 project changing the settings will run for most to all projects to call for work. I have an idea and will send ya a pm about it.
PC setup MSI-970A-G46 AMD FX-8350 8 core OC'd 4.45GHz 16GB ram PC3-10700 Geforce GTX 650Ti Windows 7 x64 Einstein@Home
Is it possible for the GPU tasks like BRP4 to download say 3 days worth of tasks in cache. It works for CPU tasks but I feel not for GPU.
Btw at the moment all BRP servers are down and we get simply no task at all.
Even when they say not working I still get work. BM explains this well in his post.
Quote:
Maybe you saw that the BRP 4 work generators were not running, this is normal in situations when a sufficient server side buffer of work has been generated. The work generators then start again once the "watermark" of prepared workunits is dropping below a certain limit.
Thanks BM I just learned something new here. Though I kinda guessed as I have not had any problems getting work.
Now for your question Hypernova.
Quote:
Is it possible for the GPU tasks like BRP4 to download say 3 days worth of tasks in cache.
Yes. It is posted some where in these forums but I will explain it here for ya. You need to go to Computing preferences http://einstein.phys.uwm.edu/prefs.php?subset=global and change 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 (max 10 days). This is your base work load min. to have in your cache. The other setting Maintain enough work for an additional is like a top end buffer to the min. Change 1 or both of these should give you enough work for up to atleast 10 days. We use this mostly when we know the servers are going to be down so we can keep working while they fix/upgrade and so forth but you can run this all the time just make sure you don't get too much work and miss the 14 day dead line.
My configuration is already like this. But at the moment I am getting no GPU tasks, and when it was the case I always got only very few at a time, except if I forced manually multiple updates in which case I would get some atsks at each call.
What cache settings are you actually running?
Boinc 7 will cache to the Min plus the Max buffer sizes, then wait for the work available to drop below the Min buffer size before asking for work again, having a small value for the 2nd cache setting will allow Boinc 7 to ask more frequently,
Is it possible for the GPU tasks like BRP4 to download say 3 days worth of tasks in cache. It works for CPU tasks but I feel not for GPU.
Btw at the moment all BRP servers are down and we get simply no task at all.
Even when they say not working I still get work. BM explains this well in his post.
Quote:
Maybe you saw that the BRP 4 work generators were not running, this is normal in situations when a sufficient server side buffer of work has been generated. The work generators then start again once the "watermark" of prepared workunits is dropping below a certain limit.
Thanks BM I just learned something new here. Though I kinda guessed as I have not had any problems getting work.
Now for your question Hypernova.
Quote:
Is it possible for the GPU tasks like BRP4 to download say 3 days worth of tasks in cache.
Yes. It is posted some where in these forums but I will explain it here for ya. You need to go to Computing preferences http://einstein.phys.uwm.edu/prefs.php?subset=global and change 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 (max 10 days). This is your base work load min. to have in your cache. The other setting Maintain enough work for an additional is like a top end buffer to the min. Change 1 or both of these should give you enough work for up to atleast 10 days. We use this mostly when we know the servers are going to be down so we can keep working while they fix/upgrade and so forth but you can run this all the time just make sure you don't get too much work and miss the 14 day dead line.
My configuration is already like this. But at the moment I am getting no GPU tasks, and when it was the case I always got only very few at a time, except if I forced manually multiple updates in which case I would get some atsks at each call.
What cache settings are you actually running?
Boinc 7 will cache to the Min plus the Max buffer sizes, then wait for the work available to drop below the Min buffer size before asking for work again, having a small value for the 2nd cache setting will allow Boinc 7 to ask more frequently,
Claggy
I have 5 days cache and 1 day minimum setting. I run only one project on each machine. On this one runs only einstein@home. There should be no interference from the other projects. And I still get no more BRP4 GPU taks.
You are right indeed Gundolf.
Here under the log file after a fesh start:
****************************************
3/9/2013 9:10:21 AM | | No config file found - using defaults
3/9/2013 9:10:21 AM | | Starting BOINC client version 7.0.54 for windows_x86_64
3/9/2013 9:10:21 AM | | log flags: file_xfer, sched_ops, task
3/9/2013 9:10:21 AM | | Libraries: libcurl/7.25.0 OpenSSL/1.0.1 zlib/1.2.6
3/9/2013 9:10:21 AM | | Data directory: C:\ProgramData\BOINC
3/9/2013 9:10:21 AM | | Running under account Fabio Manzini
3/9/2013 9:10:21 AM | | Processor: 12 GenuineIntel Intel(R) Core(TM) i7 CPU X 990 @ 3.47GHz [Family 6 Model 44 Stepping 2]
3/9/2013 9:10:21 AM | | Processor: 256.00 KB cache
3/9/2013 9:10:21 AM | | Processor features: fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush dts acpi mmx fxsr sse sse2 ss htt tm pni ssse3 cx16 sse4_1 sse4_2 popcnt aes syscall nx lm vmx tm2 pbe
3/9/2013 9:10:21 AM | | OS: Microsoft Windows 7: Ultimate x64 Edition, Service Pack 1, (06.01.7601.00)
3/9/2013 9:10:21 AM | | Memory: 23.99 GB physical, 47.98 GB virtual
3/9/2013 9:10:21 AM | | Disk: 279.36 GB total, 118.26 GB free
3/9/2013 9:10:21 AM | | Local time is UTC +1 hours
3/9/2013 9:10:21 AM | | CAL: ATI GPU 0: AMD Radeon HD 7900 series (Tahiti) (CAL version 1.4.1741, 3072MB, 3032MB available, 10240 GFLOPS peak)
3/9/2013 9:10:21 AM | | OpenCL: AMD/ATI GPU 0: AMD Radeon HD 7900 series (Tahiti) (driver version 1084.4 (VM), device version OpenCL 1.2 AMD-APP (1084.4), 3072MB, 3032MB available, 10240 GFLOPS peak)
3/9/2013 9:10:21 AM | | app version refers to missing GPU type CUDA
3/9/2013 9:10:21 AM | GPUGRID | Application uses missing CUDA GPU
3/9/2013 9:10:21 AM | | app version refers to missing GPU type CUDA
3/9/2013 9:10:21 AM | GPUGRID | Application uses missing CUDA GPU
3/9/2013 9:10:21 AM | Einstein@Home | Found app_config.xml
3/9/2013 9:10:21 AM | Einstein@Home | URL http://einstein.phys.uwm.edu/; Computer ID 6744055; resource share 100
3/9/2013 9:10:21 AM | GPUGRID | URL http://www.gpugrid.net/; Computer ID 86429; resource share 100
3/9/2013 9:10:21 AM | World Community Grid | URL http://www.worldcommunitygrid.org/; Computer ID 1148763; resource share 100
3/9/2013 9:10:21 AM | World Community Grid | General prefs: from World Community Grid (last modified 15-Nov-2012 08:58:21)
3/9/2013 9:10:21 AM | World Community Grid | Computer location: work
3/9/2013 9:10:21 AM | | General prefs: using separate prefs for work
3/9/2013 9:10:21 AM | | Reading preferences override file
3/9/2013 9:10:21 AM | | Preferences:
3/9/2013 9:10:21 AM | | max memory usage when active: 18425.33MB
3/9/2013 9:10:21 AM | | max memory usage when idle: 22110.40MB
3/9/2013 9:10:21 AM | | max disk usage: 50.00GB
3/9/2013 9:10:21 AM | | (to change preferences, visit a project web site or select Preferences in the Manager)
3/9/2013 9:10:21 AM | | Not using a proxy
3/9/2013 9:10:22 AM | Einstein@Home | Restarting task LATeah0018U_1104.0_670900_0.0_0 using hsgamma_FGRP2 version 104 in slot 6
3/9/2013 9:10:22 AM | Einstein@Home | Restarting task LATeah0018U_1104.0_670780_0.0_0 using hsgamma_FGRP2 version 104 in slot 11
3/9/2013 9:10:22 AM | Einstein@Home | Restarting task LATeah0018U_1104.0_672300_0.0_0 using hsgamma_FGRP2 version 104 in slot 9
3/9/2013 9:10:22 AM | Einstein@Home | Restarting task LATeah0018U_48.0_340_-7.3e-11_2 using hsgamma_FGRP2 version 104 in slot 5
3/9/2013 9:10:22 AM | Einstein@Home | Restarting task LATeah0018U_1104.0_666760_0.0_1 using hsgamma_FGRP2 version 104 in slot 2
3/9/2013 9:10:22 AM | Einstein@Home | Restarting task LATeah0018U_1104.0_672060_0.0_0 using hsgamma_FGRP2 version 104 in slot 4
3/9/2013 9:10:22 AM | Einstein@Home | Restarting task LATeah0018U_1104.0_670680_0.0_1 using hsgamma_FGRP2 version 104 in slot 10
3/9/2013 9:10:22 AM | Einstein@Home | Restarting task LATeah0018U_1104.0_676780_0.0_1 using hsgamma_FGRP2 version 104 in slot 0
3/9/2013 9:10:22 AM | Einstein@Home | Restarting task LATeah0018U_1104.0_674760_0.0_0 using hsgamma_FGRP2 version 104 in slot 3
3/9/2013 9:10:22 AM | Einstein@Home | Restarting task LATeah0018U_1104.0_676560_0.0_0 using hsgamma_FGRP2 version 104 in slot 7
3/9/2013 9:10:22 AM | Einstein@Home | Restarting task LATeah0018U_1104.0_674240_0.0_1 using hsgamma_FGRP2 version 104 in slot 1
3/9/2013 9:10:22 AM | Einstein@Home | Restarting task LATeah0018U_1136.0_100_0.0_1 using hsgamma_FGRP2 version 104 in slot 8
3/9/2013 9:10:55 AM | Einstein@Home | update requested by user
3/9/2013 9:10:58 AM | Einstein@Home | Sending scheduler request: Requested by user.
3/9/2013 9:10:58 AM | Einstein@Home | Not requesting tasks: don't need
3/9/2013 9:11:01 AM | Einstein@Home | Scheduler request completed
*************************************
Every time a hit update for the project einstein then I get in the logfile Not requesting tasks : don't need.
This is very strange and have no idea why.
And indeed my cache of remaining CPU tasks is now nearly empty. In a few hours my machine will stop working on Einstein by lack of new tasks.
I checked many parameters and I do not find where is the issue.
Thanks for any help.
You haven't set the manager not to accept new tasks by any chance?
No the manager is correctly set for Einstein project. The project is active and will allow new tasks.
I have two other projects that are GPUGrid and WCG. Those two are suspended and do not allow new tasks to be received.
In two hours my cache will be empty and Einstein will stop.
I switched to the latest version of Boinc 7.0.54 to see if it made a difference, but no.
RE: Btw at the moment all
)
BRP4 work distribution for GPUs is working OK as I write this.
Maybe you saw that the BRP 4 work generators were not running, this is normal in situations when a sufficient server side buffer of work has been generated. The work generators then start again once the "watermark" of prepared workunits is dropping below a certain limit.
Cheers
HBE
RE: Is it possible for the
)
Even when they say not working I still get work. BM explains this well in his post.
Thanks BM I just learned something new here. Though I kinda guessed as I have not had any problems getting work.
Now for your question Hypernova.
Yes. It is posted some where in these forums but I will explain it here for ya. You need to go to Computing preferences http://einstein.phys.uwm.edu/prefs.php?subset=global and change 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 (max 10 days). This is your base work load min. to have in your cache. The other setting Maintain enough work for an additional is like a top end buffer to the min. Change 1 or both of these should give you enough work for up to atleast 10 days. We use this mostly when we know the servers are going to be down so we can keep working while they fix/upgrade and so forth but you can run this all the time just make sure you don't get too much work and miss the 14 day dead line.
PC setup MSI-970A-G46 AMD FX-8350 8 core OC'd 4.45GHz 16GB ram PC3-10700 Geforce GTX 650Ti Windows 7 x64 Einstein@Home
RE: RE: Is it possible
)
My configuration is already like this. But at the moment I am getting no GPU tasks, and when it was the case I always got only very few at a time, except if I forced manually multiple updates in which case I would get some atsks at each call.
Running other projects has a
)
Running other projects has a big effect on this. To make a easy by pass stop all new work from every project except E@H. You might need to clear out some work first. After you get the work you want let the other projects get new work again.
Sometimes you have to force the update to happen.
Edit: With more than 1 project changing the settings will run for most to all projects to call for work. I have an idea and will send ya a pm about it.
PC setup MSI-970A-G46 AMD FX-8350 8 core OC'd 4.45GHz 16GB ram PC3-10700 Geforce GTX 650Ti Windows 7 x64 Einstein@Home
RE: RE: RE: Is it
)
What cache settings are you actually running?
Boinc 7 will cache to the Min plus the Max buffer sizes, then wait for the work available to drop below the Min buffer size before asking for work again, having a small value for the 2nd cache setting will allow Boinc 7 to ask more frequently,
Claggy
RE: RE: RE: RE: Is it
)
I have 5 days cache and 1 day minimum setting. I run only one project on each machine. On this one runs only einstein@home. There should be no interference from the other projects. And I still get no more BRP4 GPU taks.
RE: ...And I still get no
)
That's because your host doesn't ask for any. From your last contact log:
Gruß,
Gundolf
Computer sind nicht alles im Leben. (Kleiner Scherz)
RE: RE: ...And I still
)
You are right indeed Gundolf.
Here under the log file after a fesh start:
****************************************
3/9/2013 9:10:21 AM | | No config file found - using defaults
3/9/2013 9:10:21 AM | | Starting BOINC client version 7.0.54 for windows_x86_64
3/9/2013 9:10:21 AM | | log flags: file_xfer, sched_ops, task
3/9/2013 9:10:21 AM | | Libraries: libcurl/7.25.0 OpenSSL/1.0.1 zlib/1.2.6
3/9/2013 9:10:21 AM | | Data directory: C:\ProgramData\BOINC
3/9/2013 9:10:21 AM | | Running under account Fabio Manzini
3/9/2013 9:10:21 AM | | Processor: 12 GenuineIntel Intel(R) Core(TM) i7 CPU X 990 @ 3.47GHz [Family 6 Model 44 Stepping 2]
3/9/2013 9:10:21 AM | | Processor: 256.00 KB cache
3/9/2013 9:10:21 AM | | Processor features: fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush dts acpi mmx fxsr sse sse2 ss htt tm pni ssse3 cx16 sse4_1 sse4_2 popcnt aes syscall nx lm vmx tm2 pbe
3/9/2013 9:10:21 AM | | OS: Microsoft Windows 7: Ultimate x64 Edition, Service Pack 1, (06.01.7601.00)
3/9/2013 9:10:21 AM | | Memory: 23.99 GB physical, 47.98 GB virtual
3/9/2013 9:10:21 AM | | Disk: 279.36 GB total, 118.26 GB free
3/9/2013 9:10:21 AM | | Local time is UTC +1 hours
3/9/2013 9:10:21 AM | | CAL: ATI GPU 0: AMD Radeon HD 7900 series (Tahiti) (CAL version 1.4.1741, 3072MB, 3032MB available, 10240 GFLOPS peak)
3/9/2013 9:10:21 AM | | OpenCL: AMD/ATI GPU 0: AMD Radeon HD 7900 series (Tahiti) (driver version 1084.4 (VM), device version OpenCL 1.2 AMD-APP (1084.4), 3072MB, 3032MB available, 10240 GFLOPS peak)
3/9/2013 9:10:21 AM | | app version refers to missing GPU type CUDA
3/9/2013 9:10:21 AM | GPUGRID | Application uses missing CUDA GPU
3/9/2013 9:10:21 AM | | app version refers to missing GPU type CUDA
3/9/2013 9:10:21 AM | GPUGRID | Application uses missing CUDA GPU
3/9/2013 9:10:21 AM | Einstein@Home | Found app_config.xml
3/9/2013 9:10:21 AM | Einstein@Home | URL http://einstein.phys.uwm.edu/; Computer ID 6744055; resource share 100
3/9/2013 9:10:21 AM | GPUGRID | URL http://www.gpugrid.net/; Computer ID 86429; resource share 100
3/9/2013 9:10:21 AM | World Community Grid | URL http://www.worldcommunitygrid.org/; Computer ID 1148763; resource share 100
3/9/2013 9:10:21 AM | World Community Grid | General prefs: from World Community Grid (last modified 15-Nov-2012 08:58:21)
3/9/2013 9:10:21 AM | World Community Grid | Computer location: work
3/9/2013 9:10:21 AM | | General prefs: using separate prefs for work
3/9/2013 9:10:21 AM | | Reading preferences override file
3/9/2013 9:10:21 AM | | Preferences:
3/9/2013 9:10:21 AM | | max memory usage when active: 18425.33MB
3/9/2013 9:10:21 AM | | max memory usage when idle: 22110.40MB
3/9/2013 9:10:21 AM | | max disk usage: 50.00GB
3/9/2013 9:10:21 AM | | (to change preferences, visit a project web site or select Preferences in the Manager)
3/9/2013 9:10:21 AM | | Not using a proxy
3/9/2013 9:10:22 AM | Einstein@Home | Restarting task LATeah0018U_1104.0_670900_0.0_0 using hsgamma_FGRP2 version 104 in slot 6
3/9/2013 9:10:22 AM | Einstein@Home | Restarting task LATeah0018U_1104.0_670780_0.0_0 using hsgamma_FGRP2 version 104 in slot 11
3/9/2013 9:10:22 AM | Einstein@Home | Restarting task LATeah0018U_1104.0_672300_0.0_0 using hsgamma_FGRP2 version 104 in slot 9
3/9/2013 9:10:22 AM | Einstein@Home | Restarting task LATeah0018U_48.0_340_-7.3e-11_2 using hsgamma_FGRP2 version 104 in slot 5
3/9/2013 9:10:22 AM | Einstein@Home | Restarting task LATeah0018U_1104.0_666760_0.0_1 using hsgamma_FGRP2 version 104 in slot 2
3/9/2013 9:10:22 AM | Einstein@Home | Restarting task LATeah0018U_1104.0_672060_0.0_0 using hsgamma_FGRP2 version 104 in slot 4
3/9/2013 9:10:22 AM | Einstein@Home | Restarting task LATeah0018U_1104.0_670680_0.0_1 using hsgamma_FGRP2 version 104 in slot 10
3/9/2013 9:10:22 AM | Einstein@Home | Restarting task LATeah0018U_1104.0_676780_0.0_1 using hsgamma_FGRP2 version 104 in slot 0
3/9/2013 9:10:22 AM | Einstein@Home | Restarting task LATeah0018U_1104.0_674760_0.0_0 using hsgamma_FGRP2 version 104 in slot 3
3/9/2013 9:10:22 AM | Einstein@Home | Restarting task LATeah0018U_1104.0_676560_0.0_0 using hsgamma_FGRP2 version 104 in slot 7
3/9/2013 9:10:22 AM | Einstein@Home | Restarting task LATeah0018U_1104.0_674240_0.0_1 using hsgamma_FGRP2 version 104 in slot 1
3/9/2013 9:10:22 AM | Einstein@Home | Restarting task LATeah0018U_1136.0_100_0.0_1 using hsgamma_FGRP2 version 104 in slot 8
3/9/2013 9:10:55 AM | Einstein@Home | update requested by user
3/9/2013 9:10:58 AM | Einstein@Home | Sending scheduler request: Requested by user.
3/9/2013 9:10:58 AM | Einstein@Home | Not requesting tasks: don't need
3/9/2013 9:11:01 AM | Einstein@Home | Scheduler request completed
*************************************
Every time a hit update for the project einstein then I get in the logfile Not requesting tasks : don't need.
This is very strange and have no idea why.
And indeed my cache of remaining CPU tasks is now nearly empty. In a few hours my machine will stop working on Einstein by lack of new tasks.
I checked many parameters and I do not find where is the issue.
Thanks for any help.
You haven't set the manager
)
You haven't set the manager not to accept new tasks by any chance?
RE: You haven't set the
)
No the manager is correctly set for Einstein project. The project is active and will allow new tasks.
I have two other projects that are GPUGrid and WCG. Those two are suspended and do not allow new tasks to be received.
In two hours my cache will be empty and Einstein will stop.
I switched to the latest version of Boinc 7.0.54 to see if it made a difference, but no.