> On April 24, 2013, 9:37 p.m., Nathan Binkert wrote:
> > src/python/m5/stats/display.py, line 177
> > <http://reviews.gem5.org/r/1643/diff/3/?file=35285#file35285line177>
> >
> >     try/except may be faster here.  I haven't benchmarked that though.

I had a look, seems like try/except is very slightly slower in this case. Very 
marginal though.


- Sascha


-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
http://reviews.gem5.org/r/1643/#review4285
-----------------------------------------------------------


On April 25, 2013, 5 a.m., Andreas Hansson wrote:
> 
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> http://reviews.gem5.org/r/1643/
> -----------------------------------------------------------
> 
> (Updated April 25, 2013, 5 a.m.)
> 
> 
> Review request for Default.
> 
> 
> Description
> -------
> 
> Changeset 9714:68ea5c8f8d6c
> ---------------------------
> stats: Optimise the text-based stats output to improve speed
> 
> The text-based stats output is slow than the original C++ output. This patch
> addresses some of those issues by removing:
> 
>     Asserts that are not required
>     stats.ValueProxy, as it is no longer required
>     Unnecessary calls to __getattr__ and __hasattr__
> 
> It also changes the formula calculation to only calculate the values and 
> totals
> of formulas if they have not been passed in from the C++. This allows other
> formulas to be added, whilst ensuring that the stats output is sufficiently
> fast.
> 
> Testing has shown a 6x - 8x increase in stats output speed.
> 
> 
> Diffs
> -----
> 
>   src/python/m5/stats/__init__.py 773aae0990ae 
>   src/python/m5/stats/display.py PRE-CREATION 
>   src/python/m5/stats/info.py PRE-CREATION 
> 
> Diff: http://reviews.gem5.org/r/1643/diff/
> 
> 
> Testing
> -------
> 
> Ensure that the output files are identical
> 
> 
> Thanks,
> 
> Andreas Hansson
> 
>

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

Reply via email to