[cadynce] Model Checking results for GT4 w 50, 48 and 46 Blades.
Gautam Thaker
gthaker at atl.lmco.com
Tue May 29 14:38:14 CDT 2007
Adam Porter wrote:
> Gautam,
>
> Which configurations did you use for this?
>
> Adam
https://escher.isis.vanderbilt.edu/twiki/pub/CADYNCE/OneFiftyFiftyCase1/1000_allocs_at_50_48_and_46_nodes.tar.bz2
has the allocations from CMU that I have used.
I am presently trying to run these with rate monotonic priorities and preemption, but am
not sure why I don't see more improvements than I am seeing, so I have some work to do
here. I will continue to keep wiki updated.
GHT
>
>
>
> Gautam Thaker wrote:
>> Please see
>>
>> https://repo.isis.vanderbilt.edu/twiki/bin/view/CADYNCE/OneFiftyFiftyCase1#1000_Allocations_using_a_modifie
>>
>>
>> where I have uploaded results of some long simulation
>> runs with GT4 over the memorial day weekend. Comments
>> welcome. All SCIs were at same priority though the
>> scheduling model is run to completion and not time
>> shared (we don't know the quanta). I believe the
>> results would not be too different, but hard to say.
>> Also, not sure how different the results would be if
>> we employ rate monotically assigned priorities with
>> preemption. A bit of care is needed there since we
>> have merging of flows in GT4, i assume we can inherit
>> the priority of the highest flowrate that merges...
>>
>> I will attempt these priority based runs next in order
>> to gain more insights.
>>
>> Gautam
>>
>>
>>
>> ____________________________________________________________________________________You
>> snooze, you lose. Get messages ASAP with AutoCheck
>> in the all-new Yahoo! Mail Beta.
>> http://advision.webevents.yahoo.com/mailbeta/newmail_html.html
>> _______________________________________________
>> Cadynce mailing list
>> Cadynce at list.isis.vanderbilt.edu
>> http://list.isis.vanderbilt.edu/mailman/listinfo/cadynce
>>
>
> _______________________________________________
> Cadynce mailing list
> Cadynce at list.isis.vanderbilt.edu
> http://list.isis.vanderbilt.edu/mailman/listinfo/cadynce
More information about the Cadynce
mailing list