Hi
That has to be a mess to sort out better you then me. What I have seen is strange. I do a work unit that comes up as a error and every one (going on 6 now) after that also does. But then a work unit that ran in 1.00hr and 0.52hr and came up error on two different computers I ran in 1.70hr and it comes up valid.

I may have a older 2000 Lenovo at 1.67GHz but it seems to be doing very well.

So I have to ask; 1) how much of this do you think is over clocking? 2) Could some computers made for speed and media performance be inherently unstable for real world number crunching?

I'll just say I'm not in this for a prize just happy to help.

Thanks for your time.

I would say that was just dumb luck.
Your CPU should be fine. Unless there is some specific instruction that your CPU doesnt support because of its age but I doubt that otherwise the application would fail to even start.

Thanks for sorting this out so quickly guys, we really appreciate it - and thanks to our great crunchers for sticking with us!

Your Harvard CEP Team

I'm not sure if this is the kind of thing you are referring to, but I have these E224xxx tasks that are showing as errors:

E224286_243_J.65.C60H32N2S3.00032514.4.set1d06
E224255_008_I.68.C53H27N7O8.00091433.0.set1d06
E224232_137_I.66.C52H28N8O6.00379166.4.set1d06
E224287_565_J.65.C62H34S3.00034874.1.set1d06
E224261_755_I.68.C58H28N4O6.00288515.4.set1d06
E224280_326_J.66.C52H22N4O8S2.00072648.4.set1d06
E224272_475_J.61.C54H28N4S3.00098585.0.set1d06
E224237_307_I.65.C53H32N4O8.00090263.0.set1d06

MarkR

okay I tried to copy in what I was responding to, but that failed somehow.... so here goes.

I think you missed my point. As CPU speeds go up near the top end of everything else there has to come a point when some things just start to fail, be it tied to the complexity of the problem, or just because of errors from data transfer within the unit. In all my years working Quality Control I seen it happen time and again.... can't be just dumb luck.

Also I never said I was having a problem but ran work units as valid before any problem was corrected that ran on other systems and came up in error. There is always a upper limit for any tool dependent on the task. My computer as far as I know has no issues... I on the other hand has a few.

But as long as I am here again..... how many times can one work unit fail before it's no longer passed around?? I have two that failed and it seems they are going on the 7th try. Also what happens with the points/time everyone donated to show it's a DOA WU??

What the heck one more.... as this seems to be the 2nd validator problem in the past short march of weeks, why does it seem everyones reliability goes in the drain over a issue that had nothing to do with data crunchers?? No list of reliable units to start back up with or is there and I'm not seeing it's effects as everything now requires double validation?? Yes I'm sure everything is subject to validation checks anyway, but this current pattern isn't consistent with what happened just before the vaiadaters dropping out.

Just asking...