[llvm-commits] [llvm] r77397 - in /llvm/trunk: include/llvm/Support/FormattedStream.h lib/Support/FormattedStream.cpp

Evan Cheng evan.cheng at apple.com
Wed Jul 29 13:02:16 PDT 2009


On Jul 29, 2009, at 1:00 PM, David Greene wrote:

> On Wednesday 29 July 2009 12:58, Daniel Dunbar wrote:
>
>> Ok. When reenabling the formatted ostream for the asm printer, please
>> do some timing of llc on some reasonable file to verify that it
>> doesn't hurt the (non-verbose?) performance and provide that in the
>> commit.
>
> I have nightly compile times for pre- and post-commit.  They're pretty
> much identical AFAICT.  Good enough?
>
>> I don't think it is onerous to expect changes not to cause  
>> performance
>> regressions. From a testing and quality control stand point, this is
>> just like wanting the test suite to be free of failures. The problem
>> is that once regressions are allowed to stand, they compound and
>> become much harder to analyze.
>
> I have no problem with this in principle, but running test-suite  
> before
> every commit is not practical.  What we need is something in the LLVM
> tests that handles compile-time and performance regression.

Why not? If the change has a potential to cause noticeable changes to  
performance / compile time / code size. Then running the test suite is  
a requirement.

Evan

>
>                                        -Dave
> _______________________________________________
> llvm-commits mailing list
> llvm-commits at cs.uiuc.edu
> http://lists.cs.uiuc.edu/mailman/listinfo/llvm-commits




More information about the llvm-commits mailing list