>refCount is just counting the number of times that block has been
>referenced. The lru policy is maintained without making use of refCount.

i can not understand why....
Doesn't "refCount is just counting the number of times that block has
been referenced" represent a Most Recently Used (or Lease Recently
Used) block?


>Read the function accessBlock() in lru.cc
>From my understandings,

1) a block is found based on set and tag
BlkType *blk = sets[set].findBlk(tag);

2) If the block is found, it is placed on the MRU position:
sets[set].moveToHead(blk);

3) The next 'if" statement is about timing.

4) incrementing the refCount which shows this block has been accessed
blk->refCount += 1;

5) the block is returned to caller (cache_impl.hh::access())

So where is replacement?

On 1/3/12, Nilay <[email protected]> wrote:
> On Tue, January 3, 2012 1:14 am, Mahmood Naderan wrote:
>> Hi,
>> If you look at findVictim(), you can see that a block is selected
>> (with set number) and no matter what is the refCount (I think this is
>> LRU counter), the block is selected as victim.
>
> refCount is just counting the number of times that block has been
> referenced. The lru policy is maintained without making use of refCount.
>
>>
>>     BlkType *blk = sets[set].blks[assoc-1];
>>     if (blk->isValid()) {
>>         replacements[0]++;
>>         totalRefs += blk->refCount;
>>         ++sampledRefs;
>>         blk->refCount = 0;
>>         .....
>>         DPRINTF(CacheRepl, "set %x: selecting blk %x for replacement\n",
>>                 set, regenerateBlkAddr(blk->tag, set));
>>     }
>>     return blk;
>>
>> As far as I know, in LRU policy, after finding the set number, all
>> counters in the ways are compared to find the minimum one. That is
>> actually show the least referenced way.
>
> The block that has been referenced the least number of times is not
> necessarily the LRU block. Read the function accessBlock() in lru.cc to
> figure out how the LRU policy is being enforced.
>
>>
>> Can someone clarify that?
>> --
>> // Naderan *Mahmood;
>> _______________________________________________
>> gem5-users mailing list
>> [email protected]
>> http://m5sim.org/cgi-bin/mailman/listinfo/gem5-users
>>
>
>
> --
> Nilay
>
> _______________________________________________
> gem5-users mailing list
> [email protected]
> http://m5sim.org/cgi-bin/mailman/listinfo/gem5-users
>


-- 
--
// Naderan *Mahmood;
_______________________________________________
gem5-users mailing list
[email protected]
http://m5sim.org/cgi-bin/mailman/listinfo/gem5-users

Reply via email to