On Monday 29 June 2009, khaled wrote:
> I'm using an AMD Opteron 4 cores, I Want to have the TLB and the L3 
> cache misses
> so I had used valgrind with callgrind option but I had just the L1 & L2 
> caches misses
> 
> is there any option with the valgrind tools to have the TLB and the L3 
> cache misses ???

The cache model of Cachegrind/Callgrind does not simulate TLBs.

To see the L3 behavior, you could specify the L3 parameters
(size/associativity/linesize) for Cachegrinds/Callgrind L2.

However:
* Cachegrinds/Callgrinds cache model is inclusive, while AMDs caches are
  exclusive (L2/L3 are victim caches). This will lead to some differences.
* L3 on AMDs 4-cores is shared, so the scenario for the simulation assumes
  that all other cores are idle.

You could add a wishlist as bug report.

Josef 

> 
> thank you so much
> 



------------------------------------------------------------------------------
_______________________________________________
Valgrind-users mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/valgrind-users

Reply via email to