Some discussion here about using shorter deadlines at climateprediction.net

My E5400 @ 2.70GHz desktop, one of the slowest machines on the block now will complete the longest task in 5 months if only crunching when I am sat at the machine. Deadline of tasks when sent was 11 months.

My own view is that any scheduling problems caused by shortening the deadlines significantly to reduce the number of tasks sitting in caches would be a price worth paying. With more computers crunching the tasks straight away researchers would get their data back more quickly.

Am I missing anything?

Interesting that I was unable to log into the climate prediction forum. The deadline is over a year and the estimated time of completion is twenty-two days. I have eight of their processes running.

Does not seem to be any way to include a screenshot.

IN GOD WE TRUST

Interesting that I was unable to log into the climate prediction forum. The deadline is eight days, but the estimated time of completion is twenty-two days. I am going to let it run, but I am worried that the results may not be accepted past the deadline. I have eight of their processes running.

IN GOD WE TRUST

Strange, most deadlines are in the order of 11 months or longer. could you post a link to one of the tasks with an 8 day deadline? I wonder if it is an old task that has been reissued after either failing or timing out with previous cruncher(s) indicated by _1 or greater at the end of the task.

With CPDN you will still get credit even if past the deadline. however with the normal deadlines despite credit being granted, the results may not be worth much to the researchers. - A lot of this is historical from the days when most tasks took 9 months or longer even on what were then fast machines.

My fault for not checking the year. Was expecting a deadline in the same year and did not bother to look until I read another article about deadlines of over a year. I corrected the original post.

IN GOD WE TRUST