[parsec-users] PARSEC simulation results
herbert1984106 at gmail.com
Wed May 19 18:14:55 EDT 2010
the block size is keep the same under different ways (64bytes), the number
of set of L2 cache will decrease while increasing ways per set. Will this
have huge impact on miss rate of Facesim and Streamcluster? the results show
the L2 overall accesses are the same under different ways-associative. And
the L1 private cache of each processors have no impact.
-------------- next part --------------
An HTML attachment was scrubbed...
More information about the parsec-users