That would cause a lot of traffic, so if that's what you're looking for, I
would try it.
On Nov 24, 2014 11:33 AM, "Thom Popovici" <[email protected]> wrote:

> True, but that will make the data reside in the cache. I would like to see
> contention at the bus level, dram level (rank, bank, row level). I thought
> by integrating dramsim2 with marss and kicking out the caches I could do
> that.
>
> I could also try to make the l1 caches very small, lets say 64B =
> cache_line_size. This might help, right?
>
> Thanks,
> Thom
>
> On Mon, Nov 24, 2014 at 11:29 AM, Brendan Fitzgerald <
> [email protected]> wrote:
>
>> Possibly, but there would have to be significant changes.
>>
>> An easier solution would be to make L1 perfect, which is a simconfig
>> option and make the latency to L2 and main memory 0.
>>
>> I haven't tried this so I'm not sure what might happen.
>>
>> On Sun, Nov 23, 2014 at 8:40 PM, Thom Popovici <[email protected]>
>> wrote:
>>
>>> Hi, all! I am trying to do some experiments on a full system simulator
>>> that permits me to remove the caches completely from the system
>>> configuration. I noticed that in MARSS you have a config file. If I do
>>> not
>>> mention the caches will the simulator work?
>>>
>>> Thanks
>>>
>>>
>>> _______________________________________________
>>> http://www.marss86.org
>>> Marss86-Devel mailing list
>>> [email protected]
>>> https://www.cs.binghamton.edu/mailman/listinfo/marss86-devel
>>>
>>
>>
>
>
> --
> Thom Popovici
>
_______________________________________________
http://www.marss86.org
Marss86-Devel mailing list
[email protected]
https://www.cs.binghamton.edu/mailman/listinfo/marss86-devel

Reply via email to