Re: Request for code review - (ZEE patch : Redundant Zero extension elimination)

[Sriraman Tallam <tmsriram at google dot com>]

Hi,

    I am reopening this. I have attached the new ZEE patch generated
for GCC 4.6.0. Again, this patch is x86_64 specific now. I can modify
it to be a generic one if necessary. I would appreciate any feed-back
on how this can be taken forward.

Thanks,
-Sriraman.


On Mon, Jan 18, 2010 at 7:39 PM, Jie Zhang <jie.zhang@analog.com> wrote:
> Hi Sriraman,
>
> On 01/14/2010 03:55 AM, Sriraman Tallam wrote:
>>
>> Hi Richard,
>>
>> With regards to your comments in :
>> http://gcc.gnu.org/ml/gcc/2009-12/msg00028.html
>>
>>>> " which also did something stupid like move the patch down into the i386
>>>> subdirectory. It'll never be applied in that form if I have anything to say
>>>> on the subject, because obviously this is a generic sort of optimization."
>>
>> I moved it to i386 based on comments from Paolo and Richard Guenther.
>> I really don't care where the patch goes and I was just doing what
>> Paolo and Richard felt strongly about at that time about ?this being
>> very specific to x86 and using x86 patterns.
>>
>> We could come to a consensus on this and I am more than willing to
>> move it back if necessary.
>>
>> Please let me know what you think,
>
> I would like to see your patch to be extended to be a generic optimization
> pass. Apparently it is a x86_64 specific one now. I have request from
> Blackfin kernel developers to do similar optimizations, not only ZEE but
> also SEE. So I'm looking into both your proposed ZEE patch and the already
> removed SEE pass. Your ZEE patch needs generalization. The SEE pass needs
> many nontrivial fixes. I have not decided which way to go.
>
>
> Jie
>

Attachment: gcc_zee_patch.txt
Description: Text document