HI Erik, 

Following up on Amin's description, yes it's a issue,
maybe a bug, but it's hard to know what to do about it. Either way
someones configuration they've been using will change. 

Thanks, 

Ali


On 31.01.2013 10:09, Amin Farmahini wrote: 

> Hi Erik,
> 
> There is
such a problem and it caused me some problem in the past. I don't
> have
an easy solution for it. I pass all my parameters through the command
>
line to make sure Options.py doesn't override them.
> 
> Thanks,
>
Amin
> 
> On Thu, Jan 31, 2013 at 6:36 AM, Erik Tomusk
<[email protected]> wrote:
> 
>> Hello All, It looks like default
command line options override default values in configuration scripts.
I'm not sure if this would be considered a bug. For example:
O3_ARM_v7a.py defines dcache size as 32kB. Options.py defines the
default for command line option --l1d_size as 64kB. CacheConfig.py sets
the dcache size based on the command line. The result is that size,
associativity, and block size for caches, when not set on the command
line, default to the default values for the options, *not* the values
set in the configuration script. This could be problematic given that
arm_detailed has apparently been tuned to something, but running with
--cpu-type arm_detailed does not actually run the configuration defined
in O3_ARM_v7a.py. I don't immediately see a good solution. This behavior
appears to be present in the latest dev version, but I don't have time
atm to fight through the compilation to check. Has anyone else run into
this and/or think it's a problem and/or have an idea for a solution?
-Erik -- The University of Edinburgh is a charitable body, registered in
Scotland, with registration number SC005336.
______________________________**_________________ gem5-dev mailing list
[email protected] http://m5sim.org/mailman/
[1]**listinfo/gem5-dev<http://m5sim.org/mailman/listinfo/gem5-dev [2]>
>

> _______________________________________________
> gem5-dev mailing
list
> [email protected]
> http://m5sim.org/mailman/listinfo/gem5-dev
[2]

 

Links:
------
[1] http://m5sim.org/mailman/
[2]
http://m5sim.org/mailman/listinfo/gem5-dev
_______________________________________________
gem5-dev mailing list
[email protected]
http://m5sim.org/mailman/listinfo/gem5-dev

Reply via email to