[Gems-users] DNUCA search

[Steve Barrus <sbarrus@xxxxxxxxxxxx>]

I've noticed some strange behavior with MOESI_CMP_NUCA when I set
PERFECT_DNUCA_SEARCH to false.  I seems like it does the search
correctly, but there are cases were a block exists in multiple
banksets at the same time (even with ENABLE_REPLICATION set to false).
I modified the findBlock function in DNUCAPredictor.C to count the
number of banksets that have the block that it's searching for and it
sometimes it returns numbers greater than two.  Should this behavior
be possible?  Perhaps I just don't understand how how this is supposed
to work, but could this be a bug?  It doesn't seem to have do this
with PERFECT_DNUCA_SEARCH set to true.

Thanks,

-Steve

ruby0.setparam g_NUM_PROCESSORS 8
ruby0.setparam g_MEMORY_SIZE_BYTES 4294967296
ruby0.setparam g_PROCS_PER_CHIP 8
ruby0.setparam g_NUM_L2_BANKS 256
ruby0.setparam g_NUM_MEMORIES 8
ruby0.setparam g_NUM_DNUCA_BANK_SETS 16
ruby0.setparam L1_CACHE_ASSOC 2
ruby0.setparam L2_CACHE_ASSOC 16
ruby0.setparam_str g_NETWORK_TOPOLOGY FILE_SPECIFIED
ruby0.setparam_str g_DYNAMIC_TIMEOUT_ENABLED false
ruby0.setparam_str REMOVE_SINGLE_CYCLE_DCACHE_FAST_PATH true
ruby0.setparam_str g_CACHE_DESIGN NUCACOL
ruby0.setparam_str g_adaptive_routing false
ruby0.setparam NUMBER_OF_VIRTUAL_NETWORKS 7
ruby0.setparam g_endpoint_bandwidth 1000
ruby0.setparam_str g_NUCA_PREDICTOR_CONFIG DNUCA
ruby0.setparam_str ENABLE_MIGRATION true
ruby0.setparam_str COLLECTOR_HANDLES_OFF_CHIP_REQUESTS true
ruby0.setparam_str PERFECT_DNUCA_SEARCH false