Hi there,

I was wondering if anybody could help me clarify which kind of cache
mechanism adopted for gem5's ARM L2 Cache. Is it strictly inclusive,
exclusive or neither?

According to the trace I generated from gem5's ARM simulator, I found:

(1) for one address, if there is a "ReadReq" and hit a cache line in L2,
the next few accesses will continue to hit that line. That seems like it is
not totally exclusive.

(2) for one address, if there is a "ReadExReq" which incurred by a write
miss, the next accesses will become "miss". It seems that there are some
lines in L1 but not in L2, which means that it is not strictly inclusive.

If all of my observation is true, the L2 Cache in gem5's ARM simulator is
neither inclusive nor exclusive. Is it really what ARM chips do now? Why
the L2 Cache is simulated like this?

Thanks very much.

Best regards,
_______________________________________________
gem5-users mailing list
[email protected]
http://m5sim.org/cgi-bin/mailman/listinfo/gem5-users

Reply via email to