Re: [Gems-users] Gems 2.0 MESI_CMP_directory L1 + L2 miss profiling

[Polina Dudnik <pdudnik@xxxxxxxxx>]

What specific errors did you get?

On Tue, Jun 16, 2009 at 9:02 PM, Lyon Shinseky <lyon.shinseky@xxxxxxxxx> wrote:

Hi,

The miss profiling actions in the MESI_CMP_directory-L2cache.sm seem to be commented out (no uu_profilemiss action in L1 while it's there in MOESI_CMP_directory ? ) and  hence the default stats don't print L1 or L2 misses (I set REMOVE_SINGLE_CYCLE_DCACHE_FAST_PATH to true since it's a CMP protocol ). I tried to uncomment the uu_profilemiss action and the convertToGenericType slicc function in the .sm files but it ended up in a bunch of errors. The errors were given by the slicc parser. However MOESI_CMP_directory compiles fine when the uu_profilemiss actions are uncommented in the .sm files.

With the miss profiling commented out in MESI_CMP_directory .sm files , the compilation and working of the protocol. Has anyone had this problem ? The MESI_CMP_directory protocol is probably the most popular protocol in GEMS evaluations. Just wondering how this could be the case, it would be the most stable and tested of all.

Thanks
Lyon

_______________________________________________
Gems-users mailing list
Gems-users@xxxxxxxxxxx
https://lists.cs.wisc.edu/mailman/listinfo/gems-users
Use Google to search the GEMS Users mailing list by adding "site:https://lists.cs.wisc.edu/archive/gems-users/" to your search.