Well unfortunately it still leaves the fan running and people can still hear it. I am insulted that you would provide such a short and justifiable argument for something that I ought to know about already, as your entire answer! I hope that subsequent replies will either address the actual problem I presented, or do a better job of convincing everyone that we don't need checkpoints.

This isn't the place to announce your elitism.

Nuzz604,

....lets not argue and show who's got most hair on chest....i win anyway ;>)

Look at what U and Dagorath and Me know about hibernation and your statement that hibernation still leaves the fan running....I've not ever come across a computer like that, but i'm not omniscient. On mine it trundles along for about 1 minutes until the box/cpu temp is down / the hibernation process is complete, then it closes.....it's like the hot car u park in the garage, where the fan should continue for a wee bit before it shuts down.....its protection as heat build up on an acutely stopped process can do odd things....well u know of course.

The DavidHobbs manual in forum-post tells all and more on saving your work without loosing a second and only use ZERO watts in hibernation as the whole memory state is written to disk!!!!!! Proof it to yourself and pull the cable from socket

We're here to help u along in the dedication to crunch decease to death.

Signed
The Soother

Nuzz604

Your post puts everyone that could help you on the defensive at the start so you could reasonably expect some sharp responses.

In my experience, checkpointing is a non-trivial problem. In most processes, there are natural places to checkpoing..upon completing some portion of the problem. Trying to checkpoint at other places gets extraordinarily complicated, error prone and time consuming..to the point of making the process impractical. Note I said impractical...not impossible.

I do not work on the code in any of these projects so I do not know the particulars. I would suggest however that the folks that actually do the coding would love to checkpoint more often for the reasons you have given but deemed it impractical for one reason or another. I trust their judgment since I do not have either the time or understanding to suggest alternatives.

One can imagine that a point exists where enough processing time is lost by everyone checkpointing that it exceeds the time lost by some not completing work units and being reset to zero. Such points exist along the continuum.

Addressing your problem, I don't really see a good solution for you because a faster computer will still have the same problem...just less wall-clock time would be lost each day. If you cannot tolerate the resets and cannot let the machine run 24 hours, perhaps you can run some one of the other worthy distributed computing projects.

Kind Regards and Thank You for "giving it a go".

Edit (spelling ( ) )

Dagorath,

took me a while to recover from your risposta......as i understand / as is inferred about state of your chest being a trodden lane, your wife / spouse / partner / girlfriend walks all over u (and u love it)