On Tue, 15 Apr 2014, Malek Musleh wrote:



On April 10, 2014, 4:58 p.m., Nilay Vaish wrote:
configs/ruby/Ruby.py, line 102
<http://reviews.gem5.org/r/2224/diff/2/?file=39527#file39527line102>

    I am not in favour of adding such ad hoc checks to the config file.

I am not quite sure what you mean. There are other instances in the config scripts where string comparisons are done, so there is a long precedent for that (including a patch that was recently committed). Do you prefer doing something like if type(topology) == Cluster? If so, I tried that initially, but it required adding additional imports to the top of Ruby.py, and created some other complications that prevents it from running.


It is like having multiple different cpu configurations and then having checks in configuration files to support these different configs. A case in point is that of the arm detailed cpu configuration, for which changes were made to the file configs/common/Simulation.py when the configuration was initially added.

I am not against having the configuration, but it should be added in a fashion that the config files like Simulation.py or fs/se.py or Ruby.py do not need to have checks specific to individual configuration. If we make the change that you have suggested, it would mean that we are open to having checks related to different topologies in Ruby.py. And I am not in favour of having such checks.

--
Nilay
_______________________________________________
gem5-dev mailing list
gem5-dev@gem5.org
http://m5sim.org/mailman/listinfo/gem5-dev

Reply via email to