[LLVMdev] [clang] SSE2 intrinsics (emmintrin.h): _mm_movpi64_pi64 should be _mm_movpi64_epi64?

[Dimitry Andric]

On 01 Nov 2013, at 17:15, Alexey Dokuchaev <danfe at nsu.ru> wrote:
> I've recently encountered a piece of code that uses some SSE2 intrinsics
> and builds with gcc46, but not clang: clang can't find _mm_movpi64_epi64(),
> while gcc46 defines it in its lib/gcc46/gcc/.../4.6.3/include/emmintrin.h:
> 
>  extern __inline __m128i __attribute__((__gnu_inline__, __always_inline__, __artificial__))
>  _mm_movpi64_epi64 (__m64 __A)
>  {
>    return _mm_set_epi64 ((__m64)0LL, __A);
>  }
> 
>  extern __inline __m128i __attribute__((__gnu_inline__, __always_inline__, __artificial__))
>  _mm_set_epi64x (long long __q1, long long __q0)
>  {
>    return __extension__ (__m128i)(__v2di){ __q0, __q1 };
>  }
> 
> Now, Clang in /usr/include/clang/3.3/emmintrin.h defines similar function,
> but without the `e', _mm_movpi64_pi64():
> 
>  static __inline__ __m128i __attribute__((__always_inline__, __nodebug__))
>  _mm_movpi64_pi64(__m64 __a)
>  {
>    return (__m128i){ (long long)__a, 0 };
>  }
> 
> Microsoft (http://msdn.microsoft.com/en-us/library/has3d153(v=vs.90).aspx)
> defines these two:
> 
>  _mm_movepi64_pi64		MOVDQ2Q			Move
>  _mm_movpi64_epi64		MOVQ2DQ			Move
> 
> That is:
> 
>  __m64 _mm_movepi64_pi64 (__m128i a);
>  MOVDQ2Q
>  r0 := a0 ;
> 
>  __m128i _mm_movpi64_epi64 (__m64 a);
>  MOVDQ2Q
>  r0 := a0 ; r1 := 0X0 ;
> 
> Cf. Intel's manual [1]:
> 
>  _mm_movepi64_pi64		Move			MOVDQ2Q
>  _mm_movpi64_epi64		Move			MOVDQ2Q
> 
>  __m64 _mm_movepi64_pi64(__m128i a)
>  Returns the lower 64 bits of a as an __m64 type:	R0 := a0
> 
>  __m128i _mm_movpi64_pi64(__m64 a)			<<< TYPO?!
>  Moves the 64 bits of a to the lower 64 bits
>  of the result, zeroing the upper bits:		R0 := a0, R1 = 0X0
> 
> Assuming that both documents correctly assign instructions to function
> names (bonus clue: it also makes them symmetrical), then _mm_movpi64_pi64
> is indeed a typo and Clang's header is wrong, while GCC's is correct: it
> should read _mm_movpi64_epi64(), not _mm_movpi64_pi64().

Hi Alexey,

This belongs on the cfe-commits list, and I have already posted the same fix there:

http://lists.cs.uiuc.edu/pipermail/cfe-commits/Week-of-Mon-20131104/092514.html

-Dimitry

-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 203 bytes
Desc: Message signed with OpenPGP using GPGMail
URL: <http://lists.llvm.org/pipermail/cfe-commits/attachments/20131123/92b04d37/attachment.sig>