Thank you for the link. Just missed it. What kind of job can I do on this host?
I cannot do FGRP3 because of huge memory consumtion and I have no usable GPU there. Any suggestions?
Thank you for the link. Just missed it. What kind of job can I do on this host?
I cannot do FGRP3 because of huge memory consumtion and I have no usable GPU there. Any suggestions?
Huge memory consumption?
On my Win 7 system each CPU FGRP3 task uses ~260-270 MB RAM and the Gravity Wave S6 tasks use ~300-310 MB.
It's the GPU version of the FGRP3 tasks that use a lot of video-RAM on the GPU, the CPU part is ~20 MB more than the CPU version.
Edit: CPU FGRP3 1.09 uses ~260-270 MB, CPU FGRP3 1.11 uses ~280-290 MB but still less than the Gravity Wave S6 tasks. Add to that the numbers varies a bit depending on what stage is processed, seems to be some 10s up or down.
Thank you for the link. Just missed it. What kind of job can I do on this host?
I cannot do FGRP3 because of huge memory consumtion and I have no usable GPU there. Any suggestions?
Huge memory consumption?
On my Win 7 system each CPU FGRP3 task uses ~260-270 MB RAM and the Gravity Wave S6 tasks use ~300-310 MB.
It's the GPU version of the FGRP3 tasks that use a lot of video-RAM on the GPU, the CPU part is ~20 MB more than the CPU version.
Edit: CPU FGRP3 1.09 uses ~260-270 MB, CPU FGRP3 1.11 uses ~280-290 MB but still less than the Gravity Wave S6 tasks. Add to that the numbers varies a bit depending on what stage is processed, seems to be some 10s up or down.
I far as I saw my 1-core Windows XP hosts consume about 650 Mb of RAM. Nothing more is running there. But I have only 512 Mb there, so I have to opt out of FGRP3. Are there any chances that BRP4 CPU tasks will be back again some day?
running this since 02 15 2014 project einstein@home application Gravitional
name h1_0938.15_56 it got to 25% after a few days
and its gone no further since.the deadline passed on 03 01 2014. what do I do with it? any suggestion welcome as long as there not to technical with computer jargon
... it got to 25% after a few days and its gone no further since.
If you ever see that progress has stopped for say 12-24hrs or more, you should try stopping and restarting BOINC. On older and slower machines, it is possible that progress may appear to have stopped for some time, even a few hours, but it's unlikely not to make some visible progress in 24 hours or more. How long does it normally take to crunch S6CasA tasks on your computer?
Sometimes, for odd reasons, crunching may stall and stopping and restarting BOINC may get things going again. How often do you reboot the machine? Does your machine only run when you are actually using it? Do you have BOINC suspend crunching when there is user activity? Do you allow tasks to be removed from memory when crunching is suspended? The latter two are preference settings that you can control and all of these things may have a negative effect on apparent crunching progress.
Quote:
the deadline passed on 03 01 2014. what do I do with it? any suggestion welcome as long as there not to technical with computer jargon
An extra task has already been sent to a third computer as a result of your task being past the deadline. You can see the three tasks that are part of the workunit here. Your task has the time shown in red, there is one already completed, and the third copy is still in progress (showing in green). By looking at the computer doing that third task (click on the computer ID) you can estimate that the third copy will probably be returned in the next day or two. At that time the workunit will probably be completed and you will get no credit for work done. If your task is still sitting at 25%, your best option would be to abort it and start on a new task.
Have had GWS S6 on three systems.Reach 100% but fails to finish.I suspend other projects and says it is running but it never finishes or reports.Reset project on one and things are going well.Should I abort on other systems?Will run for days and just take up disk space.
Looking over the runs of the einstein_S6CasA_1.08_windows_x86_64__GWopencl-nvidia-Beta.exe application still in my All tasks list, I found that all of those run on my Windows Vista computer gave Error while computing within the first 10 seconds, but all of those on my Windows 7 computer completed and validated.
The problem is detected in msvcrt.dll, inside vsnprintf.
Could this application be using some feature that works properly withing the Windows 7 version of msvcrt.dll but not the Windows Vista version?
Do you plan to fix this problem, or do I need to look for a way to tell the server to send my Windows Vista computer only workunits other than the gravity wave type?
Do you need to modify these workunits to include whatever is needed to make the call stack show what routine within this application makes the failing call to msvcrt.dll, or can you use the current call stacks plus files on the server to show this to you but not the users?
Could this application be using some feature that works properly withing the Windows 7 version of msvcrt.dll but not the Windows Vista version?
Not that I'm aware of. The code around "Set up communication with graphics process." is still the very same that we have been using for years for all our (GW) Windows applications. This code, however, calls some BOINC routines which implementation might have changed.
Quote:
Do you plan to fix this problem, or do I need to look for a way to tell the server to send my Windows Vista computer only workunits other than the gravity wave type?
The time I would be willing to invest and possibly fixing this problem depends on how severe / widespread this particular problem is. Are you aware of anyone else who sees something similar? Anyone else here reading this?
Do you need to modify these workunits to include whatever is needed to make the call stack show what routine within this application makes the failing call to msvcrt.dll, or can you use the current call stacks plus files on the server to show this to you but not the users?
Certainly not the workunits, possibly the application. It should, however, be possible to get a bit more out of the available information with the appropriate tools oflline; I just need to find the time.
Could this application be using some feature that works properly withing the Windows 7 version of msvcrt.dll but not the Windows Vista version?
Not that I'm aware of. The code around "Set up communication with graphics process." is still the very same that we have been using for years for all our (GW) Windows applications. This code, however, calls some BOINC routines which implementation might have changed.
Quote:
Do you plan to fix this problem, or do I need to look for a way to tell the server to send my Windows Vista computer only workunits other than the gravity wave type?
The time I would be willing to invest and possibly fixing this problem depends on how severe / widespread this particular problem is. Are you aware of anyone else who sees something similar? Anyone else here reading this?
BM
I have essentially no information on who else participating in this project is running Windows Vista, or who else is seeing this problem.
For now, I have set my Windows Vista computer so it will not download any gravity wave workunits at all. That should allow it to do more useful work.
Thank you for the link. Just
)
Thank you for the link. Just missed it. What kind of job can I do on this host?
I cannot do FGRP3 because of huge memory consumtion and I have no usable GPU there. Any suggestions?
RE: Thank you for the link.
)
Huge memory consumption?
On my Win 7 system each CPU FGRP3 task uses ~260-270 MB RAM and the Gravity Wave S6 tasks use ~300-310 MB.
It's the GPU version of the FGRP3 tasks that use a lot of video-RAM on the GPU, the CPU part is ~20 MB more than the CPU version.
Edit: CPU FGRP3 1.09 uses ~260-270 MB, CPU FGRP3 1.11 uses ~280-290 MB but still less than the Gravity Wave S6 tasks. Add to that the numbers varies a bit depending on what stage is processed, seems to be some 10s up or down.
RE: RE: Thank you for the
)
I far as I saw my 1-core Windows XP hosts consume about 650 Mb of RAM. Nothing more is running there. But I have only 512 Mb there, so I have to opt out of FGRP3. Are there any chances that BRP4 CPU tasks will be back again some day?
running this since 02 15 2014
)
running this since 02 15 2014 project einstein@home application Gravitional
name h1_0938.15_56 it got to 25% after a few days
and its gone no further since.the deadline passed on 03 01 2014. what do I do with it? any suggestion welcome as long as there not to technical with computer jargon
RE: ... it got to 25% after
)
If you ever see that progress has stopped for say 12-24hrs or more, you should try stopping and restarting BOINC. On older and slower machines, it is possible that progress may appear to have stopped for some time, even a few hours, but it's unlikely not to make some visible progress in 24 hours or more. How long does it normally take to crunch S6CasA tasks on your computer?
Sometimes, for odd reasons, crunching may stall and stopping and restarting BOINC may get things going again. How often do you reboot the machine? Does your machine only run when you are actually using it? Do you have BOINC suspend crunching when there is user activity? Do you allow tasks to be removed from memory when crunching is suspended? The latter two are preference settings that you can control and all of these things may have a negative effect on apparent crunching progress.
An extra task has already been sent to a third computer as a result of your task being past the deadline. You can see the three tasks that are part of the workunit here. Your task has the time shown in red, there is one already completed, and the third copy is still in progress (showing in green). By looking at the computer doing that third task (click on the computer ID) you can estimate that the third copy will probably be returned in the next day or two. At that time the workunit will probably be completed and you will get no credit for work done. If your task is still sitting at 25%, your best option would be to abort it and start on a new task.
Cheers,
Gary.
Have had GWS S6 on three
)
Have had GWS S6 on three systems.Reach 100% but fails to finish.I suspend other projects and says it is running but it never finishes or reports.Reset project on one and things are going well.Should I abort on other systems?Will run for days and just take up disk space.
Never mind.This machine just
)
Never mind.This machine just uploaded.
Looking over the runs of the
)
Looking over the runs of the einstein_S6CasA_1.08_windows_x86_64__GWopencl-nvidia-Beta.exe application still in my All tasks list, I found that all of those run on my Windows Vista computer gave Error while computing within the first 10 seconds, but all of those on my Windows 7 computer completed and validated.
The problem is detected in msvcrt.dll, inside vsnprintf.
Could this application be using some feature that works properly withing the Windows 7 version of msvcrt.dll but not the Windows Vista version?
Do you plan to fix this problem, or do I need to look for a way to tell the server to send my Windows Vista computer only workunits other than the gravity wave type?
Do you need to modify these workunits to include whatever is needed to make the call stack show what routine within this application makes the failing call to msvcrt.dll, or can you use the current call stacks plus files on the server to show this to you but not the users?
RE: RE: RE: Could this
)
Certainly not the workunits, possibly the application. It should, however, be possible to get a bit more out of the available information with the appropriate tools oflline; I just need to find the time.
BM
BM
RE: RE: Could this
)
I have essentially no information on who else participating in this project is running Windows Vista, or who else is seeing this problem.
For now, I have set my Windows Vista computer so it will not download any gravity wave workunits at all. That should allow it to do more useful work.