The voltage reduction tests archae86 did on his E6600 and Q6600 intrigued me to try it on my own Q6600.
Back in May when I overclocked my quad to 3.2GHz I found it to run 8h stable (Prime95/Orthos) at 1.400V, but when I ran E@H a few WU's errored out within the first few days. I ran Orthos again and it failed just under 9h, I bumped the voltage up to 1.4125V and haven't had a single stability issue since in both Linux and XP. I regularly run 4 WU's and play Battlefield 2 at the same time and she's rock solid.
Just after the last 4 WU's started (in Linux) I rebooted at 1.400V to see if they'd complete at a lower voltage. So far so good. They've now completed and the next 4 have begun without issue.
I'll run a 1.400V for the next few days to test the XP and Linux app's and report on my findings.
Either the app's aren't as 'hard' on the CPU's, or somehow my hardware has 'settled' ad requires less voltage. I'm hoping for the latter. ;-)
Sorry to dig up this "old" thread but I just wanted to report that I'm still running at 1.400V and have had no problems whatsoever.
I guess I should try 1.3875V now, I'm liking this :-)
There are 10^11 stars in the galaxy. That used to be a huge number. But it's only a hundred billion. It's less than the national deficit! We used to call them astronomical numbers. Now we should call them economical numbers. - Richard Feynman
It features even more code to track some of the remaining problems. For now we are also distributing the PDB file (containing debugging information) again with the beta package.
This App is the first that has been built with VS2005. I hope that this helps with some of the library problems we see. I don't yet know how this affects the performance.
Please test.
BM
Hi,
This WU could not finish, reproducibly, on 2 different computers.
In task manager window, on "Processes" tab, in "Page Faults" column, for both "einstein_S5R2_4.38_windows_intelx86.exe" i have a value about 3000.
If I understand right, "page fault" occurs when required memory page located in swap file (or not?).
This value - good, bad, or norm?
Configuration:
ASUS P5K / E6550 (2.33 overclocked to 3.15) / 2x(1GB DDR2-6400)
Thank you!
i dont have an answer for ya, but you got me curious so i lookd at mine, e@h has been working on a wu for 90 hours. and my "page fault" count is. get this. 7,271,323 and climbing by about 1k every two secconds. so i sure hope its nothing lol
seeing without seeing is something the blind learn to do, and seeing beyond vision can be a gift.
i dont have an answer for ya, but you got me curious so i lookd at mine, e@h has been working on a wu for 90 hours. and my "page fault" count is. get this. 7,271,323 and climbing by about 1k every two secconds. so i sure hope its nothing lol
I think about einstein_*.exe perfomance. How many resources is lost on this "page faults"? Tiny, small, or more? :)
I'm ~140M page fualts on a pair of 635 credit WUs that are about hafl done, growning by several k per second. System is at 750/2000mb ram used, XpPro.
My vista laptop has ~270k page faults on each of the 666 credit WUs running. Those numbers are not growing and since the WUs are at 92/51% done I'm wondering if the app isn't being restarted between runs. It's at 620/1000mb physical ram used.
hmm. same wu its now at 56,038,282 the wu in question is http://einsteinathome.org/workunit/34671277 (sorry, dont know how to shrink url's on here) its been going for 108 hours and looks to have 3 more hours to go. the system has 1GB physical ram and a 1GB page file.
seeing without seeing is something the blind learn to do, and seeing beyond vision can be a gift.
Don't know how much help this is now that we've moved into S5R3, but one of my hosts faulted out on a 4.38 with a 105 while cleaning up it's R2 datapaks. This is the first compute error I've seen on this host in a long time which wasn't my fault, it's normally very reliable.
I looked over the whole WU, and another host had a problem with the WU. Also, unfortunately my host didn't hang in the dogfight quite long enough before getting shot down. One of the other wingmen was running late, and a new task got issued before he reported in, so this one going to be hanging in the DB until the new guy finishes or gets shot down. ;-)
If mine had hung in there for about another 4 1/2 hours this one one would have been history. :-(
Don't know how much help this is now that we've moved into S5R3, but one of my hosts faulted out on a 4.38 with a 105 while cleaning up it's R2 datapaks. This is the first compute error I've seen on this host in a long time which wasn't my fault, it's normally very reliable.
The "NULL pointer" message doesn't look good. However two machines finished this WU without error, so this doesn't look like a programming error (which the failed sanity check was meant to catch). Watch out for memory problems. Might have some transient problem, thiugh.
Sorry, the Wu can't be found anymore. If it had left the active database, it means that a canonical result was found, i.e. at least two machines finished this WU successfully and their results agreed.
There might be something wrong with the URL you gave, though.
RE: The voltage reduction
)
Sorry to dig up this "old" thread but I just wanted to report that I'm still running at 1.400V and have had no problems whatsoever.
I guess I should try 1.3875V now, I'm liking this :-)
There are 10^11 stars in the galaxy. That used to be a huge number. But it's only a hundred billion. It's less than the national deficit! We used to call them astronomical numbers. Now we should call them economical numbers. - Richard Feynman
RE: A new Windows App is
)
Hi,
This WU could not finish, reproducibly, on 2 different computers.
http://einsteinathome.org/workunit/34696485
Cheers,
Igor
Hello! In task manager
)
Hello!
In task manager window, on "Processes" tab, in "Page Faults" column, for both "einstein_S5R2_4.38_windows_intelx86.exe" i have a value about 3000.
If I understand right, "page fault" occurs when required memory page located in swap file (or not?).
This value - good, bad, or norm?
Configuration:
ASUS P5K / E6550 (2.33 overclocked to 3.15) / 2x(1GB DDR2-6400)
Thank you!
RE: Hello! In task manager
)
i dont have an answer for ya, but you got me curious so i lookd at mine, e@h has been working on a wu for 90 hours. and my "page fault" count is. get this. 7,271,323 and climbing by about 1k every two secconds. so i sure hope its nothing lol
seeing without seeing is something the blind learn to do, and seeing beyond vision can be a gift.
RE: i dont have an answer
)
I think about einstein_*.exe perfomance. How many resources is lost on this "page faults"? Tiny, small, or more? :)
I'm ~140M page fualts on a
)
I'm ~140M page fualts on a pair of 635 credit WUs that are about hafl done, growning by several k per second. System is at 750/2000mb ram used, XpPro.
My vista laptop has ~270k page faults on each of the 666 credit WUs running. Those numbers are not growing and since the WUs are at 92/51% done I'm wondering if the app isn't being restarted between runs. It's at 620/1000mb physical ram used.
hmm. same wu its now at
)
hmm. same wu its now at 56,038,282 the wu in question is http://einsteinathome.org/workunit/34671277 (sorry, dont know how to shrink url's on here) its been going for 108 hours and looks to have 3 more hours to go. the system has 1GB physical ram and a 1GB page file.
seeing without seeing is something the blind learn to do, and seeing beyond vision can be a gift.
Don't know how much help this
)
Don't know how much help this is now that we've moved into S5R3, but one of my hosts faulted out on a 4.38 with a 105 while cleaning up it's R2 datapaks. This is the first compute error I've seen on this host in a long time which wasn't my fault, it's normally very reliable.
87391100
Hmmm...
I looked over the whole WU, and another host had a problem with the WU. Also, unfortunately my host didn't hang in the dogfight quite long enough before getting shot down. One of the other wingmen was running late, and a new task got issued before he reported in, so this one going to be hanging in the DB until the new guy finishes or gets shot down. ;-)
If mine had hung in there for about another 4 1/2 hours this one one would have been history. :-(
Alinator
RE: Don't know how much
)
The "NULL pointer" message doesn't look good. However two machines finished this WU without error, so this doesn't look like a programming error (which the failed sanity check was meant to catch). Watch out for memory problems. Might have some transient problem, thiugh.
BM
BM
RE: This WU could not
)
Sorry, the Wu can't be found anymore. If it had left the active database, it means that a canonical result was found, i.e. at least two machines finished this WU successfully and their results agreed.
There might be something wrong with the URL you gave, though.
BM
BM