Welcome to mirror list, hosted at ThFree Co, Russian Federation.

git.blender.org/blender.git - Unnamed repository; edit this file 'description' to name the repository.
summaryrefslogtreecommitdiff
diff options
context:
space:
mode:
authorThomas Dinges <blender@dingto.org>2014-06-13 23:13:18 +0400
committerThomas Dinges <blender@dingto.org>2014-06-13 23:59:12 +0400
commitcd5e1ff74e4f6443f3e4b836dd23fe46b56cb7ed (patch)
tree578ee132eab87d348147e49c91e1929660558c20 /intern/cycles/kernel/svm/svm_texture.h
parentd0573ce9054e325c0ad2fbb943087e0f8b9e159a (diff)
Cycles Refactor: Add SSE Utility code from Embree for cleaner SSE code.
This makes the code a bit easier to understand, and might come in handy if we want to reuse more Embree code. Differential Revision: https://developer.blender.org/D482 Code by Brecht, with fixes by Lockal, Sergey and myself.
Diffstat (limited to 'intern/cycles/kernel/svm/svm_texture.h')
-rw-r--r--intern/cycles/kernel/svm/svm_texture.h22
1 files changed, 11 insertions, 11 deletions
diff --git a/intern/cycles/kernel/svm/svm_texture.h b/intern/cycles/kernel/svm/svm_texture.h
index 5fd9204cbf6..d97c85db36a 100644
--- a/intern/cycles/kernel/svm/svm_texture.h
+++ b/intern/cycles/kernel/svm/svm_texture.h
@@ -140,15 +140,15 @@ ccl_device float voronoi_F1_distance(float3 p)
}
}
#else
- __m128 vec_p = load_m128(p);
- __m128i xyzi = quick_floor_sse(vec_p);
+ ssef vec_p = load4f(p);
+ ssei xyzi = quick_floor_sse(vec_p);
for (int xx = -1; xx <= 1; xx++) {
for (int yy = -1; yy <= 1; yy++) {
for (int zz = -1; zz <= 1; zz++) {
- __m128 ip = _mm_cvtepi32_ps(_mm_add_epi32(xyzi, _mm_setr_epi32(xx, yy, zz, 0)));
- __m128 vp = _mm_add_ps(ip, cellnoise_color(ip));
- float d = len_squared<1, 1, 1, 0>(_mm_sub_ps(vec_p, vp));
+ ssef ip = ssef(xyzi + ssei(xx, yy, zz, 0));
+ ssef vp = ip + cellnoise_color(ip);
+ float d = len_squared<1, 1, 1, 0>(vec_p - vp);
da = min(d, da);
}
}
@@ -184,15 +184,15 @@ ccl_device float3 voronoi_F1_color(float3 p)
return cellnoise_color(pa);
#else
- __m128 pa, vec_p = load_m128(p);
- __m128i xyzi = quick_floor_sse(vec_p);
+ ssef pa, vec_p = load4f(p);
+ ssei xyzi = quick_floor_sse(vec_p);
for (int xx = -1; xx <= 1; xx++) {
for (int yy = -1; yy <= 1; yy++) {
for (int zz = -1; zz <= 1; zz++) {
- __m128 ip = _mm_cvtepi32_ps(_mm_add_epi32(xyzi, _mm_setr_epi32(xx, yy, zz, 0)));
- __m128 vp = _mm_add_ps(ip, cellnoise_color(ip));
- float d = len_squared<1, 1, 1, 0>(_mm_sub_ps(vec_p, vp));
+ ssef ip = ssef(xyzi + ssei(xx, yy, zz, 0));
+ ssef vp = ip + cellnoise_color(ip);
+ float d = len_squared<1, 1, 1, 0>(vec_p - vp);
if(d < da) {
da = d;
@@ -202,7 +202,7 @@ ccl_device float3 voronoi_F1_color(float3 p)
}
}
- __m128 color = cellnoise_color(pa);
+ ssef color = cellnoise_color(pa);
return (float3 &)color;
#endif
}