AuroraMiddleware/skia2
1
0
Fork 0
Code Issues Pull Requests Projects Releases Wiki Activity

Have Windows use _mm_rsqrt_ss too.

Browse Source 
Tidy up a little while I'm in here:
  1) SIMD headers are now included by SkTypes.h as appropriate.
  2) _mm_cvtss_f32() is pithier and generates the same code.

Looks like this is the only code checking for SSE wrong.  After this CL:
    ~/skia (sse) $ git grep __SSE
    include/core/SkPreConfig.h:    #if defined(__SSE4_2__)
    include/core/SkPreConfig.h:    #elif defined(__SSE4_1__)
    include/core/SkPreConfig.h:    #elif defined(__SSE3__)
    include/core/SkPreConfig.h:    #elif defined(__SSE2__)
every other check is in SkPreConfig.h where it belongs.

This is going to affect some GMs subtly on Windows.

BUG=chromium:511458

No public API changes.
TBR=reed@google.com

Review URL: https://codereview.chromium.org/1248503004
This commit is contained in:
mtklein 2015-07-23 08:37:02 -07:00 committed by Commit bot
parent 11dfc8e57a
commit b9c47f9968
1 changed files with 2 additions and 10 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

12
include/core/SkFloatingPoint.h
View File

@ -143,12 +143,6 @@ extern const uint32_t gIEEENegativeInfinity;
#define SK_FloatInfinity (*SkTCast<const float*>(&gIEEEInfinity))
#define SK_FloatNegativeInfinity (*SkTCast<const float*>(&gIEEENegativeInfinity))
#if defined(__SSE__)
#include <xmmintrin.h>
#elif defined(SK_ARM_HAS_NEON)
#include <arm_neon.h>
#endif
// Fast, approximate inverse square root.
// Compare to name-brand "1.0f / sk_float_sqrt(x)". Should be around 10x faster on SSE, 2x on NEON.
static inline float sk_float_rsqrt(const float x) {
@ -157,10 +151,8 @@ static inline float sk_float_rsqrt(const float x) {
//
// We do one step of Newton's method to refine the estimates in the NEON and null paths. No
// refinement is faster, but very innacurate. Two steps is more accurate, but slower than 1/sqrt.
#if defined(__SSE__)
float result;
_mm_store_ss(&result, _mm_rsqrt_ss(_mm_set_ss(x)));
return result;
#if SK_CPU_SSE_LEVEL >= SK_CPU_SSE_LEVEL_SSE1
return _mm_cvtss_f32(_mm_rsqrt_ss(_mm_set_ss(x)));
#elif defined(SK_ARM_HAS_NEON)
// Get initial estimate.
const float32x2_t xx = vdup_n_f32(x); // Clever readers will note we're doing everything 2x.