Relative to a 2GHz or 4GHz CPU? One thing you could do to make this  
work reasonably easily is change the global tick rate to however fast  
you want to tick your CPU. Then all the latencies could be in terms of  
ticks (e.g. '5t') which would be 5 cycles at 2 or 4GHz. The CPU in  
this case could be 1t.

Another options would be to have a global indicating the desired  
frequency/latency things should be relative to and add a new specified  
that multiplies by that frequency/latency.  You would need to modify  
src/python/m5/convert.py to understand your new system and add a new  
character to decode (e.g. '5r' would mean 5 clocks at whatever the  
global clock rate you've specified somewhere is).

For a while we tried to have a 'c' specifier that was relative to the  
CPU the object was part of. However, this quickly became very  
complicated when you have multiple CPUs that could operate at  
different frequencies.

Hope that helps,

Ali




On Oct 29, 2008, at 8:07 AM, Clint Smullen wrote:

> I've not seen any such function, since M5 does not use a global clock
> rate (only the global tick rate). In your script, however, you can
> easily add a function that does the conversion.
>
> On Oct 29, 2008, at 12:39 AM, Shoaib Akram wrote:
>
>> I just want them to be set to constant number of cycles. And change
>> the frequency. for instancem bridge should queue requets for 5
>> cycles, at 2GHz or at 4GHz...
>>
>> ---- Original message ----
>>> Date: Tue, 28 Oct 2008 21:22:47 -0700
>>> From: Ali Saidi <[EMAIL PROTECTED]>
>>> Subject: Re: [m5-users] specification of delay
>>> To: M5 users mailing list <[email protected]>
>>>
>>> You can specify them is Hz I believe (e.g. 2GHz == .5ns) or are you
>>> wanting it to be relative to the CPU frequency?
>>>
>>> Ali
>>>
>>>
>>>
>>> On Oct 28, 2008, at 9:15 PM, Shoaib Akram wrote:
>>>
>>>> is there a way to easily change the specification of delay for
>>>> caches, bridges, and memory from 'ns' etc to actual number of
>>>> cycles.
>>>>
>>>> This is useful when comparing across different frequencies since
>>>> number of cycles remain constant.
>>>> _______________________________________________
>>>> m5-users mailing list
>>>> [email protected]
>>>> http://m5sim.org/cgi-bin/mailman/listinfo/m5-users
>>>>
>>>
>>> _______________________________________________
>>> m5-users mailing list
>>> [email protected]
>>> http://m5sim.org/cgi-bin/mailman/listinfo/m5-users
>> _______________________________________________
>> m5-users mailing list
>> [email protected]
>> http://m5sim.org/cgi-bin/mailman/listinfo/m5-users
>
> _______________________________________________
> m5-users mailing list
> [email protected]
> http://m5sim.org/cgi-bin/mailman/listinfo/m5-users
>

_______________________________________________
m5-users mailing list
[email protected]
http://m5sim.org/cgi-bin/mailman/listinfo/m5-users

Reply via email to