search for: _mm_empty

...ne asm can still have MMX > operands/results/clobbers, and can still put the processor in MMX mode. If > code is mixing MMX intrinsics and inline asm, there could be a significant > penalty to moving values across register files. And it’s not clear what we > want to do with _mm_empty(): under full emulation, it should be a no-op, > but if there’s MMX asm, we need to actually clear the register file. > > Moving data between the register files in order to call an inline asm is not a correctness issue, however, just a potential performance issue. The compiler will inse...

I recently diagnosed a bug in someone else's software, which turned out to be due to incorrect MMX intrinsics usage: if you use any of the x86 intrinsics that accept or return __m64 values, then you, the *programmer* are required to call _mm_empty() before using any x87 floating point instructions or leaving the function. I was aware that this was required at the assembly-level, but not that the compiler forced users to deal with this when using intrinsics. This is a real nasty footgun -- if you get this wrong, your program doesn't cras...

...e current ARM code to ARM SIMD Alan Coopersmith (1): Define force_inline for compilers other than gcc & MS Visual C Benjamin Otte (3): force alignment of arg pointer in potential SSE users update .gitignore unswitch red and blue David M?ller (1): Move _mm_empty() to correct place David Woodhouse (1): Less fragile Linux altivec detection Jeff Muizelaar (5): Add support for ARMv6 SIMD fastpaths. Fix inner branch code. Cleanup inline definitions [arm-simd] Fix typo found by ??? Unify scanline buffer types Ke...

...p); sign = _m_paddw (sign, cmp); cmp = _m_pmaddwd (diff, sign); sad = _m_paddd (sad, cmp); return _m_to_int (sad) + _m_to_int (_m_psrlqi (sad, 32)); } but it isn't faster. Does anyone of you have got a hint to make it faster? I've got another question: why don't you call _mm_empty when you use intrinsic asm? Thank you and excuse me for the OT. -- Ottavio Campana Telecommunication Engineer Lab. Immagini Dept. of Information Engineering University of Padova Via Gradenigo 6/B 35131 Padova Italy

...Move PIXMAN_EXPORT into pixman-region.c Move all the PIXMAN_EXPORT into .c files Some formatting fixing Add pixman_region32_copy_from_region16 Use 32 bit regions internally Add pixman_image_set_clip_region32 Call _mm_empty() at the end of fbComposeSetupSSE Fix bug in pixman_compute_composite_region() Export pixman_compute_composite_region32() and use it in walk_region Add infinite loop test Fix uses of short variables in pixman-region.c...

...to Mpx_per_sec() lowlevel-blt-bench: add CSV output mode lowlevel-blt-bench: add option to skip memcpy measurement Siarhei Siamashka (3): configure.ac: Check if the compiler supports GCC vector extensions test: Fix OpenMP clauses for the tolerance-test mmx: Fix _mm_empty problems for over_8888_8888/over_8888_n_8888 Søren Sandmann (14): Post-release version bump to 0.33.1 demos/Makefile.am: Move EXTRA_DIST outside "if HAVE_GTK" test/utils.c: Make the stack unaligned only on 32 bit Windows general: Support component alpha for...

...utils.[ch]: add fence_get_page_size() implementation: add PIXMAN_DISABLE=wholeops armv6: enable over_n_8888 Siarhei Siamashka (5): configure.ac: Check if the compiler supports GCC vector extensions test: Fix OpenMP clauses for the tolerance-test mmx: Fix _mm_empty problems for over_8888_8888/over_8888_n_8888 pixman-general: Fix stack related pointer arithmetic overflow vmx: implement fast path vmx_composite_over_n_8888 Søren Sandmann (24): Post-release version bump to 0.33.1 demos/Makefile.am: Move EXTRA_DIST outside "if HA...