Re: [Gems-users] L1 I cache statistic problem

["Mike Marty" <mikem@xxxxxxxxxxx>]

Unfortunately, for x86 targets, Virtutech Simics currently does not allow instructions to be stalled.

 

--Mike

 

----- Original Message -----

From: guo fei

To: gems-users@xxxxxxxxxxx

Sent: Monday, September 11, 2006 10:04 PM

Subject: [Gems-users] L1 I cache statistic problem

hi,

I need some help with regards to Ruby statistic.
I am trying to run Ruby(only,no opal) by using MSI_MOSI_CMP_directory protocol upon a X86 disk image . The statisitc of L1 D cache and L2 cache looks fine. But the L1 I cache statistic are all 0.

I check the source code and find the reason is that for all the memory_transactions ("mem_trans")come from simics, if the transaction type is "mem_trans->s.type ==Sim_Trans_Instr_Fetch", the "mem_trans->s.may_stall" is always 0.  Therefore in the
" SimicsDriver::isUnhandledTransaction(memory_transaction_t* mem_trans)"
function, all the instruction fetch transactions are filtered.

I did add -stall arguement when running simics, and the load/store transactions seem correct.
Can anyone help with this issue?
Any comment is highly appreciated.

Fei

---

雅虎免费邮箱-3.5G容量，20M附件

---

_______________________________________________
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.