[PATCH] Check isTypeDesirableForOp before replacing mulhu/mulhs

[Fiona Glaser]

It looks like what’s going is there’s a urem i8, which we should be able to promote, but this doesn’t work because i16 isn’t a Good Type for x86. So the iffy thing here is that i8 is considered okay at this point, but i16 isn’t, but i32 is. So because it can’t get promoted to i16, it ends up being compiled differently, I think?

Fiona

> On Feb 17, 2015, at 11:28 PM, Owen Anderson <resistor at mac.com> wrote:
> 
> I’m fine with the change in principle, but it would be good to get someone with more X86 knowledge to weigh in on the regression there.
> 
> —Owen
> 
>> On Feb 17, 2015, at 3:34 PM, Matt Arsenault <arsenm2 at gmail.com <mailto:arsenm2 at gmail.com>> wrote:
>> 
>>> 
>>> On Jan 22, 2015, at 1:28 PM, Matt Arsenault <arsenm2 at gmail.com <mailto:arsenm2 at gmail.com>> wrote:
>>> 
>>> 
>>>> On Jan 8, 2015, at 4:54 PM, Matt Arsenault <arsenm2 at gmail.com <mailto:arsenm2 at gmail.com>> wrote:
>>>> 
>>>> 
>>>>> On Jan 5, 2015, at 7:51 AM, Matt Arsenault <arsenm2 at gmail.com <mailto:arsenm2 at gmail.com>> wrote:
>>>>> 
>>>>> 
>>>>>> On Dec 11, 2014, at 3:30 PM, Matt Arsenault <arsenm2 at gmail.com <mailto:arsenm2 at gmail.com>> wrote:
>>>>>> 
>>>>>> 
>>>>>>> On Dec 2, 2014, at 4:43 PM, Matt Arsenault <arsenm2 at gmail.com <mailto:arsenm2 at gmail.com>> wrote:
>>>>>>> 
>>>>>>> 
>>>>>> 
>>>>>> ping
>>>>> 
>>>>> ping
>>>> 
>>>> 
>>>> pre-branch ping
>>> 
>>> 
>>> ping
>> 
>> ping
>> _______________________________________________
>> llvm-commits mailing list
>> llvm-commits at cs.uiuc.edu <mailto:llvm-commits at cs.uiuc.edu>
>> http://lists.cs.uiuc.edu/mailman/listinfo/llvm-commits <http://lists.cs.uiuc.edu/mailman/listinfo/llvm-commits>
> _______________________________________________
> llvm-commits mailing list
> llvm-commits at cs.uiuc.edu
> http://lists.cs.uiuc.edu/mailman/listinfo/llvm-commits

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.llvm.org/pipermail/llvm-commits/attachments/20150218/53afa2ae/attachment.html>