Why have avx, fma4, and sse4a been added to CK_BDVER1? xop implies all of
those. Can we use comments instead?

On Wed, Nov 5, 2014 at 9:21 AM, Andrea Di Biagio <
[email protected]> wrote:

> Hi rnk, craig.topper, qcolombet,
>
> Hi Reid, Quentin (and all),
>
> This patch slightly refactors how clang sets default features for AMD
> bdver cpus.
> Method 'getDefaultFeatures' (in Basic/Targets.cpp), already uses a
> fallthrough mechanism to set default features for 'bdver4', bdver3 and
> bdver2. This patch simply adds another fallthrough from case 'CK_BDVER2' to
> case 'CK_BDVER1'. That is because 'bdver2' has the same features available
> in 'bdver1' plus BMI,FMA,F16C and TBM etc.
>
> While at it, I added missing checks for features in test
> predefined-arch-macros.c.
> In the case of BTVER2, this patch adds explicit checks for __F16C__,
> __BMI__ and __PCLMUL__.
> In the case of BDVER3 and BDVER4, this patch checks for the presence of
> __FSGSBASE__ (added at r221130).
>
> Please let me know if ok to submit.
>
> Thanks,
> Andrea
>
> http://reviews.llvm.org/D6134
>
> Files:
>   lib/Basic/Targets.cpp
>   test/Preprocessor/predefined-arch-macros.c
>



-- 
~Craig
_______________________________________________
cfe-commits mailing list
[email protected]
http://lists.cs.uiuc.edu/mailman/listinfo/cfe-commits

Reply via email to