-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
http://reviews.gem5.org/r/2749/#review6086
-----------------------------------------------------------


Why make such a large and ugly change to fix a seemingly simple problem in 
configuration?  The Ruby Tester has always been a single tester, with multiple 
ports, rather than a set of testers.  This patch makes a set of testers, but 
then uses an ugly combination of static variables and conditions to fall back 
on the single tester logic.

We have a patch in our patch queue that we will soon post that I believe fixes 
this problem in a much better way.


src/cpu/testers/rubytest/RubyTester.hh (line 140)
<http://reviews.gem5.org/r/2749/#comment5292>

    We should avoid static variables.  At some point, we could be interested in 
running multiple instances of Ruby with the random tester.



src/cpu/testers/rubytest/RubyTester.cc (line 96)
<http://reviews.gem5.org/r/2749/#comment5293>

    This is pretty ugly hack.  The tester is not meant to be created per cpu.  
There is a single tester and check table that is used to coordinate the races.


- Brad Beckmann


On April 27, 2015, 7:23 p.m., Nilay Vaish wrote:
> 
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> http://reviews.gem5.org/r/2749/
> -----------------------------------------------------------
> 
> (Updated April 27, 2015, 7:23 p.m.)
> 
> 
> Review request for Default.
> 
> 
> Repository: gem5
> 
> 
> Description
> -------
> 
> Changeset 10811:b826d3a54c0c
> ---------------------------
> cpu: testers: rubytest: fix the test
> 
> The test was broken.  When run with more than a cpu, it resulted in error 
> while
> configuring the system.  Coherence protocols assume that multiple cpus are
> present when options.num_cpus > 1.  But the test only creates one tester even
> when we were testing a multicpu system.  Now, we will create individual 
> testers
> for each cpu.
> 
> 
> Diffs
> -----
> 
>   src/cpu/testers/rubytest/RubyTester.hh df2aa91dba5b 
>   src/cpu/testers/rubytest/RubyTester.cc df2aa91dba5b 
>   src/cpu/testers/rubytest/RubyTester.py df2aa91dba5b 
>   tests/configs/rubytest-ruby.py df2aa91dba5b 
>   configs/example/ruby_random_test.py df2aa91dba5b 
>   src/cpu/testers/rubytest/Check.hh df2aa91dba5b 
>   src/cpu/testers/rubytest/Check.cc df2aa91dba5b 
>   src/cpu/testers/rubytest/CheckTable.hh df2aa91dba5b 
>   src/cpu/testers/rubytest/CheckTable.cc df2aa91dba5b 
> 
> Diff: http://reviews.gem5.org/r/2749/diff/
> 
> 
> Testing
> -------
> 
> 
> Thanks,
> 
> Nilay Vaish
> 
>

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

Reply via email to