I think I'll follow that one in a few days, see where it goes. As for now, I'm still content with the old 4.37. I hope something can help out those poor Xeons, though.
microcraft
"The arc of history is long, but it bends toward justice" - MLK
Thanks for the link. I have a couple of machines crunching EAH/Seti (90/10). I got the seti science apps that he created quite a while ago. Makes a brilliant difference to Seti times on P3 & P4 and a smaller improvement on AMD (AthlonXP).
I've never bothered with an optimised CC though. The optimised seti app however is great on P4 as even at 90/10 a nice amount of seti gets done without subtracting much from EAH.
Quote:
I think I'll follow that one in a few days, see where it goes. As for now, I'm still content with the old 4.37. I hope something can help out those poor Xeons, though.
If I were you I'd probably upgrade 4.37 to something like 5.1.5 or whatever is the latest so as to be able to help testing for the impending 5.2 release, if you've got the time and inclination, that is. A number of people seem to be successfully running at the bleeding edge and you've certainly got the skills to observe and give useful feedback.
Not sure what you mean by the reference to "poor Xeons". I only quickly skimmed the thread and didn't see any Xeons. Probably just couldn't see them. I try to stay away from Seti as much as possible ...
Just now at 4:22 pm here , EAH's progress is 25.12 % .
As for the 2 times :
- Time to completion still continue to fall : 73:34:46
- The CPU time being 24:43:30 , always in progression , because yesterday
it was 05:29:06 ( of course if i'm not wrong about the CPU time )
I'm sure now about the CPU time , it has reached 25 , so it s always in progression
Cheers
EAH has had a good long run now so it should change back to Seti in the not too distant future. When it does, BOINC should hopefully allow Seti to complete and just before it completes BOINC should attempt to get the next Seti work unit. Everything appears to be normal. We will just wait for BOINC to start Seti again.
If I were you I'd probably upgrade 4.37 to something like 5.1.5 or whatever is the latest so as to be able to help testing for the impending 5.2 release, if you've got the time and inclination, that is. A number of people seem to be successfully running at the bleeding edge and you've certainly got the skills to observe and give useful feedback.
Not sure what you mean by the reference to "poor Xeons". I only quickly skimmed the thread and didn't see any Xeons. Probably just couldn't see them. I try to stay away from Seti as much as possible ...
Hmmm...I gave the .18 Einstein a short (10WU) run 12 days ago or so, with mixed results - all 10 WUs validated fine, but computation times were 10% higher, so I discontinued, and I wasn't able to give it a thorough test for what I assume was the main fix because I can't find an option to display the graphics, although my rig has an Ati Radeon vidcard. Probably has something to do with my initial installation preferences back in March, if there was a "do not use graphic" option, I'm sure I selected it, remembering from SetiClassic how much of a hit WU times took. Besides, Einstein wouldn't be able to find it, because I've since deleted the .scr file from Windows. Any feedback on what effect the 5.1.5 (or whatever) Boinc beta has upon times and benchmarks?
As far as the reference to "poor Xeons", no, there was no reference to them in that thread. I was just hoping something could be done for them, for their benchmarks or whatever, because they have such horrible claimed credit, usually in the low 40s. While that usually doesn't affect me, when I see 2 of them partnered with me on a WU, it sets off alarm bells, "Here comes a low granted credit WU", but I can only imagine running one, taking so long on a WU, and then getting basically half-credit for all the time. That must suck. Good thing for them that there are usually others to pull the credit up, the system usually working out OK in the end.
microcraft
"The arc of history is long, but it bends toward justice" - MLK
Just now at 4:22 pm here , EAH's progress is 25.12 % .
As for the 2 times :
- Time to completion still continue to fall : 73:34:46
- The CPU time being 24:43:30 , always in progression , because yesterday
it was 05:29:06 ( of course if i'm not wrong about the CPU time )
Still ~98 hours for the total. Darn, I thought that the effect of all the tweaking we did to recover cycles for crunching would bring a continuing, though diminishing, reduction in total time. The substantial drop from 100 hours to 98 hours over the first 5% had me expecting that our modifications would bring the total down to somewhere around 90 hours eventually. Einstein crunches very linearly (%complete/CPU time) vs Seti, which has a quick section, a slow section, a moderate section, and a hurry-up at the end.
microcraft
"The arc of history is long, but it bends toward justice" - MLK
You just saved me a reply :). You managed to score the (usually) small interval between the quorum being formed and the validator noticing :). You'll also notice in that particular quorum that PowerMacs often make very low claims as well. Quite annoying when you get saddled with a couple of those :).
I'll reply to other messages later when I get some time.
...You'll also notice in that particular quorum that PowerMacs often make very low claims as well. Quite annoying when you get saddled with a couple of those :).
I'll reply to other messages later when I get some time.
Whew, I'll bet! Good thing the AltiVec optimize at least helps PowerMacs sail through the WUs, but still an unfairly low credit claim. On the very next WU quorum from the one you looked at, not 10 minutes later, I got sandwiched between a PowerMac and a Xeon. Yikes! Ouch! 43credits :-0
microcraft
"The arc of history is long, but it bends toward justice" - MLK
Gary, As I was just now
)
Gary,
As I was just now poking through the Seti MB, I saw that Maverick has put up 4.72 Windows and Linux CCs optimized for mmx, SSE, and SSE2.
http://setiathome.berkeley.edu/forum_thread.php?id=20782
I think I'll follow that one in a few days, see where it goes. As for now, I'm still content with the old 4.37. I hope something can help out those poor Xeons, though.
microcraft
"The arc of history is long, but it bends toward justice" - MLK
RE: As I was just now
)
Thanks for the link. I have a couple of machines crunching EAH/Seti (90/10). I got the seti science apps that he created quite a while ago. Makes a brilliant difference to Seti times on P3 & P4 and a smaller improvement on AMD (AthlonXP).
I've never bothered with an optimised CC though. The optimised seti app however is great on P4 as even at 90/10 a nice amount of seti gets done without subtracting much from EAH.
If I were you I'd probably upgrade 4.37 to something like 5.1.5 or whatever is the latest so as to be able to help testing for the impending 5.2 release, if you've got the time and inclination, that is. A number of people seem to be successfully running at the bleeding edge and you've certainly got the skills to observe and give useful feedback.
Not sure what you mean by the reference to "poor Xeons". I only quickly skimmed the thread and didn't see any Xeons. Probably just couldn't see them. I try to stay away from Seti as much as possible ...
Cheers,
Gary.
Dear Gary and Michael ,
)
Dear Gary and Michael ,
Just now at 4:22 pm here , EAH's progress is 25.12 % .
As for the 2 times :
- Time to completion still continue to fall : 73:34:46
- The CPU time being 24:43:30 , always in progression , because yesterday
it was 05:29:06 ( of course if i'm not wrong about the CPU time )
Ariane
RE: Dear Gary and
)
I'm sure now about the CPU time , it has reached 25 , so it s always in progression
Cheers
Ariane
RE: RE: Dear Gary and
)
EAH has had a good long run now so it should change back to Seti in the not too distant future. When it does, BOINC should hopefully allow Seti to complete and just before it completes BOINC should attempt to get the next Seti work unit. Everything appears to be normal. We will just wait for BOINC to start Seti again.
Cheers,
Gary.
RE: RE: If I were you I'd
)
Hmmm...I gave the .18 Einstein a short (10WU) run 12 days ago or so, with mixed results - all 10 WUs validated fine, but computation times were 10% higher, so I discontinued, and I wasn't able to give it a thorough test for what I assume was the main fix because I can't find an option to display the graphics, although my rig has an Ati Radeon vidcard. Probably has something to do with my initial installation preferences back in March, if there was a "do not use graphic" option, I'm sure I selected it, remembering from SetiClassic how much of a hit WU times took. Besides, Einstein wouldn't be able to find it, because I've since deleted the .scr file from Windows. Any feedback on what effect the 5.1.5 (or whatever) Boinc beta has upon times and benchmarks?
As far as the reference to "poor Xeons", no, there was no reference to them in that thread. I was just hoping something could be done for them, for their benchmarks or whatever, because they have such horrible claimed credit, usually in the low 40s. While that usually doesn't affect me, when I see 2 of them partnered with me on a WU, it sets off alarm bells, "Here comes a low granted credit WU", but I can only imagine running one, taking so long on a WU, and then getting basically half-credit for all the time. That must suck. Good thing for them that there are usually others to pull the credit up, the system usually working out OK in the end.
microcraft
"The arc of history is long, but it bends toward justice" - MLK
RE: Dear Gary and
)
Still ~98 hours for the total. Darn, I thought that the effect of all the tweaking we did to recover cycles for crunching would bring a continuing, though diminishing, reduction in total time. The substantial drop from 100 hours to 98 hours over the first 5% had me expecting that our modifications would bring the total down to somewhere around 90 hours eventually. Einstein crunches very linearly (%complete/CPU time) vs Seti, which has a quick section, a slow section, a moderate section, and a hurry-up at the end.
microcraft
"The arc of history is long, but it bends toward justice" - MLK
Off-topic, but I wander
)
Off-topic, but I wander :-)
Gary, this is something new to me - 3 valid on-time results on a WU, and no granted??? http://einsteinathome.org/workunit/2225766
Do you see this very often?
{edited moments later}
Disregard the above. Credit granted now. I guess I just caught the server in-process.
microcraft
"The arc of history is long, but it bends toward justice" - MLK
You just saved me a reply :).
)
You just saved me a reply :). You managed to score the (usually) small interval between the quorum being formed and the validator noticing :). You'll also notice in that particular quorum that PowerMacs often make very low claims as well. Quite annoying when you get saddled with a couple of those :).
I'll reply to other messages later when I get some time.
Cheers,
Gary.
RE: ...You'll also notice
)
Whew, I'll bet! Good thing the AltiVec optimize at least helps PowerMacs sail through the WUs, but still an unfairly low credit claim. On the very next WU quorum from the one you looked at, not 10 minutes later, I got sandwiched between a PowerMac and a Xeon. Yikes! Ouch! 43credits :-0
microcraft
"The arc of history is long, but it bends toward justice" - MLK