Weekly Report (10/20)

@ Phoenix 

I attached the ppt file for presentation. 

A summary and the time taken by each simulation are added. 

@ Booksim 

with the fly topology, overall accepted throughput of voq is not better than voq. (similar to mesh and ring) 

but before that, I am wondering that the overall average latency of voq(vc=16 buf=16) is larger than novoq(vc=1, buf=16), especially for fly topology. This happens when packet size = 4 or using 25% hotspot traffic. 

I tested same thing for 64node-fly and the result is the same regardless of buf size. 

It might be a simulation error or not, I should figure it out. 

Thanks. 

minjeong