[llvm-commits] [PATCH][APFloat] add APFloat::isDenormal()
Michael Ilseman
milseman at apple.com
Wed Jan 2 12:51:33 PST 2013
On Jan 2, 2013, at 12:45 PM, Shuxin Yang <shuxin.llvm at gmail.com> wrote:
> My following changes will include quite a few testing cases which disable fast-math optimization if constant-folding yields a denormal.
> These testing cases covers the "test" hard-coded in this patch.
>
I think Duncan was referring to the /unittests/ folder, which has places where you can write C++ code and assert on values. Your hard-coded tests could be moved directly into there, swapping out assert(…) for EXPECT_TRUE(…) where applicable. See /unittests/ADT/APFloatTest.cpp for examples.
> On 1/2/13 12:19 PM, Duncan Sands wrote:
>> Hi Shuxin,
>>
>> > The "test" is hard-coded in the InstCombineAddSub.cpp (this part is not
>>> going to be committed).
>>> and the test is triggered by any *.ll having a fadd instruction.
>>>
>>> The "test" covers the tricky cases about single/double/quadruple-precision
>>> as well as Intel 80-bit
>>> double-ext precision floating-point. I don't know how to test PPC FP. But I
>>> don't think it is necessary
>>> to test PPC FP as APFloat is a arch-neutral representations of all these formats.
>>
>> maybe turn it into a unit test (presumably in unittests/ADT/APFloatTest.cpp)?
>>
>> Ciao, Duncan.
>> _______________________________________________
>> llvm-commits mailing list
>> llvm-commits at cs.uiuc.edu
>> 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
More information about the llvm-commits
mailing list