IOW's you can't have it both ways. You're implying if you ran them one at a time the 7970's would magically start cranking them out in 3000 seconds.
I can tell you for a FACT, they won't.
i'm not implying, i've performed some test.
single WU on 7970 runs for 5600 sec.
2 WUs in parallel completed in 7800 seconds.
4 WUs in 13200 sec.
and so on.
OK, I'll believe that when I start seeing your host post some completed tasks that have RUN TIMES of 5600 seconds.
Until then, oh look at those pigs flying upside down over there!
OK, I'll believe that when I start seeing your host post some completed tasks that have RUN TIMES of 5600 seconds.
Until then, oh look at those pigs flying upside down over there!
Alinator, you're way out of line!
Try taking a look at this workunit. His hostID is 7416776. That's close enough to 5600s for me. Obviously he wouldn't leave it doing tasks 1x when it can be so much more productive at higher values so you're not likely to find too many of these.
I think an apology is in order! Astrocrab is to be congratulated for contributing valuable performance data about HD7970 GPUs.
I can confirm the results that Astrocrab is seeing running a single task. I am running the same model card and Linux OS and a single task runs between 5555 and 5600 seconds with the few tasks that I ran in this configuration.
Ok, so this is figured out now, but we still have to solve the problem of the OP, right?
I think we are close, let's see what we've got:
- There seems to be a problem reading the init_data.xml in the slot, as is evident from the log output, forcing the app to run in standalone mode which is bad.
- As Ageless pointed out, this file also is responsible to keep track of the processing time, which explains why the CPU time reported is always 0 sec.
It would be useful to have a look at the init_data.xml file , BUT you should never, ever post this file in the forum, as it contains the authenticator, which would allow anybody to log into your account.
So robl, if you know a thing or two about xml files you might want to look at this file, e.g. with tools like xmllint that can check the syntax of xml files.
If you are ok with this, you can also send me such an init_data.xml file (feel free to replace the authenticator with some random stuff) from a slot that has a BRP5 job in it to heinz-bernd.eggenstein[ät]aei.mpg.de for inspection. It would be helpful for us to know whether the BOINC client or the boinc_api code (which parses the XML) is to blame, in the latter case we would want to compile new app versions.
Ok, so this is figured out now, but we still have to solve the problem of the OP, right?
I think we are close, let's see what we've got:
- There seems to be a problem reading the init_data.xml in the slot, as is evident from the log output, forcing the app to run in standalone mode which is bad.
- As Ageless pointed out, this file also is responsible to keep track of the processing time, which explains why the CPU time reported is always 0 sec.
It would be useful to have a look at the init_data.xml file , BUT you should never, ever post this file in the forum, as it contains the authenticator, which would allow anybody to log into your account.
So robl, if you know a thing or two about xml files you might want to look at this file, e.g. with tools like xmllint that can check the syntax of xml files.
If you are ok with this, you can also send me such an init_data.xml file (feel free to replace the authenticator with some random stuff) from a slot that has a BRP5 job in it to heinz-bernd.eggenstein[ät]aei.mpg.de for inspection. It would be helpful for us to know whether the BOINC client or the boinc_api code (which parses the XML) is to blame, in the latter case we would want to compile new app versions.
Cheers
HB
Bikeman,
I have emailed you the init_data.xml file as you suggested after slight modification to "field data". The structure is unchanged. This is a BRP4cuda32nv270 file. I ran xmllint and yes the file is corrupted. Not being xml experienced I could go no further.
I and others I'm sure will be interested in your response. Knowing who/how this file is generated would certainly be educational.
RE: RE: IOW's you can't
)
OK, I'll believe that when I start seeing your host post some completed tasks that have RUN TIMES of 5600 seconds.
Until then, oh look at those pigs flying upside down over there!
RE: OK, I'll believe that
)
Alinator, you're way out of line!
Try taking a look at this workunit. His hostID is 7416776. That's close enough to 5600s for me. Obviously he wouldn't leave it doing tasks 1x when it can be so much more productive at higher values so you're not likely to find too many of these.
I think an apology is in order! Astrocrab is to be congratulated for contributing valuable performance data about HD7970 GPUs.
Cheers,
Gary.
RE: OK, I'll believe that
)
man, i don't see any reason why you should not believe me.
as you wish:
one and two
I can confirm the results
)
I can confirm the results that Astrocrab is seeing running a single task. I am running the same model card and Linux OS and a single task runs between 5555 and 5600 seconds with the few tasks that I ran in this configuration.
RE: Until then, oh look at
)
Alinator, you're not understanding a simple concept. Please spend some time studying it. You don't want to look silly.
RE: RE: OK, I'll believe
)
My apologies, seeing is believing.
However, the way you were talking made it seem like you were doing this right now rather than the beginning of the month. ;-)
I was already impressed with my 7970's and all I can say now is WOW! :-D
RE: However, the way you
)
oh, time conjugation in english verbs is my huge problem (
LOL... Don't worry about
)
LOL...
Don't worry about it. I'm sure your English is far better than my Russian! :-)
Hi! Ok, so this is figured
)
Hi!
Ok, so this is figured out now, but we still have to solve the problem of the OP, right?
I think we are close, let's see what we've got:
- There seems to be a problem reading the init_data.xml in the slot, as is evident from the log output, forcing the app to run in standalone mode which is bad.
- As Ageless pointed out, this file also is responsible to keep track of the processing time, which explains why the CPU time reported is always 0 sec.
It would be useful to have a look at the init_data.xml file , BUT you should never, ever post this file in the forum, as it contains the authenticator, which would allow anybody to log into your account.
So robl, if you know a thing or two about xml files you might want to look at this file, e.g. with tools like xmllint that can check the syntax of xml files.
If you are ok with this, you can also send me such an init_data.xml file (feel free to replace the authenticator with some random stuff) from a slot that has a BRP5 job in it to heinz-bernd.eggenstein[ät]aei.mpg.de for inspection. It would be helpful for us to know whether the BOINC client or the boinc_api code (which parses the XML) is to blame, in the latter case we would want to compile new app versions.
Cheers
HB
RE: Hi! Ok, so this is
)
Bikeman,
I have emailed you the init_data.xml file as you suggested after slight modification to "field data". The structure is unchanged. This is a BRP4cuda32nv270 file. I ran xmllint and yes the file is corrupted. Not being xml experienced I could go no further.
I and others I'm sure will be interested in your response. Knowing who/how this file is generated would certainly be educational.