Rule Suggestion - 'Same Task' Condition

[tlsi2000]

Rule Condition Suggestion
'Same Task' Condition

I have set up a rule that allows me to suspend tasks that do not complete an initialization phase (these are SETI tasks).  The init phase is (should be) 30sec to 1min long, and drops the GPU temp (on my system) to less than 37C.  Thus when the init gets 'stuck', the temp stays low, and the rule fires.

The rule is set to be based upon the temp (GPU temp) being less than 45C, with the rule timer set to 10 minutes to take into account any false positives.   I have the rule set to suspend the task on this condition.

What seems to be happening is when the tasks are short (10-20 min), the rule condition often becomes true because the each task drops the temp below the set point, and the rule timer happens to fire during the init phase of a *different* following task, resulting in suspending every 3/4/5th task.

I see and agree with the basis of the rule being system (non-task) specific, and there is a condition to test for a project or a specific application.

But...

Would there be a way to in indicate that the rule be same-task-specific or rather, keep the 'same task' as a condition ?   Maybe a checkbox condition that the rule also include the last task's identifier (id/name/whatever) as part of the next-check condition.

Of course one would need to determine if this is applicable to just my situation or is possibly applicable to a larger group of general cases before going to a lot of coding.

Thanks.

[fred]

Rule Condition Suggestion
'Same Task' Condition

Did you try to include a rule like Elapsed time > 10:00. This should give the task time to run and warm up the card.
When the problems happen you don't see any changes int the CPU % column. Like a high CPU usage?
I try to keep the rules general and not too specific.
Try analyzing the problem.  Run TThrottle to make a graph when things running fine.

[tlsi2000]

Ok, thanks for the insight on this.

That should work, as that then would be task-specific on the other condition.  I'll work in that direction.

Yes, I do agree on the general nature of the rules.

They really are great !

Thanks.

btw -- on v.064 when editing a rule, the 'enabled' checkmark goes away and has to be re-enabled on the rule just edited   (maybe intentional ?)

[fred]

btw -- on v.064 when editing a rule, the 'enabled' checkmark goes away and has to be re-enabled on the rule just edited   (maybe intentional ?)

No its a small bug.