-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
http://reviews.gem5.org/r/2957/
-----------------------------------------------------------

Review request for Default.


Repository: gem5


Description
-------

Changeset 10933:184eac6db147
---------------------------
ruby: call setMRU from L1 controllers, not from sequencer
Currently the sequencer calls the function setMRU that updates the replacement
policy structures with the first level caches.  While functionally this is
correct, the problem is that this requires calling findTagInSet() which is an
expensive function.  This patch removes the calls to setMRU from the sequencer.
The first level controllers shoul now update the replacement policy.  If the
access is a cache hit, the function findTagInSet is called only once when the
block is first found (common case).  If the access is a miss, then the
controller calls setMRU() just before calling the sequencer for returning the
data.


Diffs
-----

  src/mem/protocol/MESI_Three_Level-L0cache.sm 5c76426fd9ee 
  src/mem/protocol/MESI_Three_Level-L1cache.sm 5c76426fd9ee 
  src/mem/protocol/MESI_Two_Level-L1cache.sm 5c76426fd9ee 
  src/mem/protocol/MESI_Two_Level-L2cache.sm 5c76426fd9ee 
  src/mem/protocol/MI_example-cache.sm 5c76426fd9ee 
  src/mem/protocol/MOESI_CMP_directory-L1cache.sm 5c76426fd9ee 
  src/mem/protocol/MOESI_CMP_directory-L2cache.sm 5c76426fd9ee 
  src/mem/protocol/MOESI_CMP_token-L1cache.sm 5c76426fd9ee 
  src/mem/protocol/MOESI_CMP_token-L2cache.sm 5c76426fd9ee 
  src/mem/protocol/MOESI_hammer-cache.sm 5c76426fd9ee 
  src/mem/protocol/MOESI_hammer-dir.sm 5c76426fd9ee 
  src/mem/protocol/RubySlicc_Types.sm 5c76426fd9ee 
  src/mem/ruby/structures/CacheMemory.hh 5c76426fd9ee 
  src/mem/ruby/structures/CacheMemory.cc 5c76426fd9ee 
  src/mem/ruby/system/Sequencer.cc 5c76426fd9ee 

Diff: http://reviews.gem5.org/r/2957/diff/


Testing
-------


Thanks,

Nilay Vaish

_______________________________________________
gem5-dev mailing list
[email protected]
http://m5sim.org/mailman/listinfo/gem5-dev

Reply via email to