Re: [Gems-users] memory accesses


Date: Sat, 21 Mar 2009 20:11:29 -0400
From: Polina Dudnik <pdudnik@xxxxxxxxx>
Subject: Re: [Gems-users] memory accesses
Are these the parameters from .stats file or from the script or from the default configurations?

Are you sure that out of the three conditions

 if (IS_DEV_MEM_OP(mem_trans->s.ini_type) ||
      IS_OTH_MEM_OP(mem_trans->s.ini_type) ||
      mem_trans->s.physical_address > uinteger_t(RubyConfig::memorySizeBytes())

It is the third one that is true?

The reason I asked you what is inside your transaction is because I though that maybe you had some IO. But it seems like you have a very simple transaction. Are you sure the rest of your setup is right: like you are binding the threads and everything?

Polina

On Sat, Mar 21, 2009 at 7:49 PM, Shakeel Butt <shakeel.butt@xxxxxxxxx> wrote:
Yup, I wrote my own workload. Its really simple, I'm just incrementing
a global variable. It looks like:

-------------------------
for(i = 0 ; i < num_ops ; ++i)
{
  BEGIN_TRANSACTION(0)
  temp = counter;
  temp = temp + 1;
  counter = temp;
  COMMIT_TRANSACTION(0)
}
------------------------
counter is global variable and temp is local.

I'm attaching the ruby parameter list.

Shakeel



On Fri, Mar 20, 2009 at 11:12 PM, Polina Dudnik <pdudnik@xxxxxxxxx> wrote:
> Did you make your own transactional workload? If so, what do you have inside
> the transaction?
>
> Also, did you change any of the ruby parameters?
>
> On Fri, Mar 20, 2009 at 5:28 PM, Shakeel Butt <shakeel.butt@xxxxxxxxx>
> wrote:
>>
>> Hi,
>>
>> I am having problem with simics memory accesses during transactional
>> workload. The physical address generated by simics are larger than the
>> actual memory presented in the system. In "SimicsDriver::makeRequest"
>> all the such accesses are ignored by ruby. So, the memory accesses
>> during transactions are ignored. I am running 8 cpu system with
>> (256*8) MB of RAM but the physical addresses generated are like 40
>> bits. I have checked the logical address and that address is correct
>> (same as in program). Am I missing something?
>>
>> One more thing has anyone tried Tourmaline on simics 3.0 or tried to
>> migrate it to simics 3.0?
>>
>> Shakeel
>> _______________________________________________
>> 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.
>>
>
>
> _______________________________________________
> 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.
>
>
>

_______________________________________________
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.



[← Prev in Thread] Current Thread [Next in Thread→]