Means your machine did a factor 40 longer to complete the task than was originally estimated, and yes many of the 333 are known to run long. After 15 hours failing, the original estimated must then have been less than 22.5 minutes. The project average last few days has been well over one hour, so it's hard to tell how this could have happened.

Does the device have a variable CPU speed? If the task came at for instance client benchmark 15000 but then was really running at a benchmark speed of 7500, this could occur sooner. At any rate, the enormous variability may be a reason for the technicians to up the </rsc_fpops_bound> to a factor 50.

The explanation is in my previous post.... the exceptions are covered through the upper bound factor of 40. Other projects have a factor 5 to 10. The unit must have been -originally- assigned right when a larger set of shorts were influencing the average run time for new work to be that low. This is not expressed in the day averages of which lowest was 0.69 hours on Feb.23, hence the suggestion made to the technicians.

Of course the client benchmark could be heavily optimistic. What are the Whetstone/Dhrystone values and was it Linux or Windows? If a device is claiming to be a Ferrari but really runs as VW, the allowed maximum runtime starts to play a roll.

BTW, I've seen tasks with max exceed being credited, just don't know if this is policy.

Of course the other question, seeing WinXP, what client version?

The OET1_ *-OM batches run quite a bit longer than other OET1 batches which causes problems for the estimations. A few weeks ago we decided to hold the OET1_*-OM batches until the other batches have run (and OET1_*-OM batches won't be run on android). A few of the OET1_*-OM batches had already been sent out though. The vast majority of those completed without problems, but a few hit the 'maximum elapsed time exceeded'. Those will just be re-run when we've completed other batches and are only running OET1_*-OM batches.

Seippel

Does it mean that android devices will be kicked out of this project in the near future?