evas: add SSE3 op_blend_color blend functions

_op_blend_c_dp_sse3
_op_blend_rel_c_dp_sse3

Patch by: Jim Kukunas <james.t.kukunas@linux.intel.com>

git-svn-id: http://svn.enlightenment.org/svn/e/trunk/evas@63699 7cbeb6ba-43b4-40fd-8cce-4c39aea84d33

diff --git a/src/lib/engines/common/evas_op_blend/op_blend_color_sse3.c b/src/lib/engines/common/evas_op_blend/op_blend_color_sse3.c
new file mode 100644 (file)
index 0000000..215c170
--- /dev/null
+++ b/src/lib/engines/common/evas_op_blend/op_blend_color_sse3.c
@@ -0,0 +1,165 @@
+/* blend color -> dst */
+
+#ifdef BUILD_SSE3
+
+static void
+_op_blend_c_dp_sse3(DATA32 *s __UNUSED__, DATA8 *m __UNUSED__, DATA32 c, DATA32 *d, int l) {
+
+   DATA32 a = 256 - (c >> 24);
+
+   const __m128i c_packed = _mm_set_epi32(c, c, c, c);
+   const __m128i a_packed = _mm_set_epi32(a, a, a, a);
+
+   LOOP_ALIGNED_U1_A48_SSE3(d, l,
+      { /* UOP */
+
+         *d = c + MUL_256(a, *d);
+         d++; l--;
+      },
+      { /* A4OP */
+
+         __m128i d0 = _mm_load_si128((__m128i *)d);
+
+         d0 = mul_256_sse3(a_packed, d0);
+         d0 = _mm_add_epi32(d0, c_packed);
+
+         _mm_store_si128((__m128i *)d, d0);
+
+         d += 4; l -= 4;
+      },
+      { /* A8OP */
+
+         __m128i d0 = _mm_load_si128((__m128i *)d);
+         __m128i d1 = _mm_load_si128((__m128i *)(d+4));
+
+         d0 = mul_256_sse3(a_packed, d0);
+         d1 = mul_256_sse3(a_packed, d1);
+
+         d0 = _mm_add_epi32(d0, c_packed);
+         d1 = _mm_add_epi32(d1, c_packed);
+
+         _mm_store_si128((__m128i *)d, d0);
+         _mm_store_si128((__m128i *)(d+4), d1);
+
+         d += 8; l -= 8;
+      })
+}
+
+#define _op_blend_caa_dp_sse3 _op_blend_c_dp_sse3
+
+#define _op_blend_c_dpan_sse3 _op_blend_c_dp_sse3
+#define _op_blend_caa_dpan_sse3 _op_blend_c_dpan_sse3
+
+static void
+init_blend_color_span_funcs_sse3(void)
+{
+   op_blend_span_funcs[SP_N][SM_N][SC][DP][CPU_SSE3] = _op_blend_c_dp_sse3;
+   op_blend_span_funcs[SP_N][SM_N][SC_AA][DP][CPU_SSE3] = _op_blend_caa_dp_sse3;
+
+   op_blend_span_funcs[SP_N][SM_N][SC][DP_AN][CPU_SSE3] = _op_blend_c_dpan_sse3;
+   op_blend_span_funcs[SP_N][SM_N][SC_AA][DP_AN][CPU_SSE3] = _op_blend_caa_dpan_sse3;
+}
+
+#define _op_blend_pt_c_dp_sse3 NULL
+#define _op_blend_pt_caa_dp_sse3 _op_blend_pt_c_dp_sse3
+
+#define _op_blend_pt_c_dpan_sse3 _op_blend_pt_c_dp_sse3
+#define _op_blend_pt_caa_dpan_sse3 _op_blend_pt_c_dpan_sse3
+
+#define _op_blend_pt_c_dpas_sse3 _op_blend_pt_c_dp_sse3
+#define _op_blend_pt_caa_dpas_sse3 _op_blend_pt_c_dp_sse3
+
+static void
+init_blend_color_pt_funcs_sse3(void)
+{
+   op_blend_pt_funcs[SP_N][SM_N][SC][DP][CPU_SSE3] = _op_blend_pt_c_dp_sse3;
+   op_blend_pt_funcs[SP_N][SM_N][SC_AA][DP][CPU_SSE3] = _op_blend_pt_caa_dp_sse3;
+
+   op_blend_pt_funcs[SP_N][SM_N][SC][DP_AN][CPU_SSE3] = _op_blend_pt_c_dpan_sse3;
+   op_blend_pt_funcs[SP_N][SM_N][SC_AA][DP_AN][CPU_SSE3] = _op_blend_pt_caa_dpan_sse3;
+}
+
+
+/*-----*/
+
+/* blend_rel color -> dst */
+
+static void
+_op_blend_rel_c_dp_sse3(DATA32 *s __UNUSED__, DATA8 *m __UNUSED__, DATA32 c, DATA32 *d, int l) {
+
+   int alpha = 256 - (c >> 24);
+
+   const __m128i c_packed = _mm_set_epi32(c, c, c, c);
+   const __m128i alpha_packed = _mm_set_epi32(alpha, alpha, alpha, alpha);
+
+   LOOP_ALIGNED_U1_A48_SSE3(d, l,
+      {  /* UOP */
+
+         *d = MUL_SYM(*d >> 24, c) + MUL_256(alpha, *d);
+         d++; l--;
+      },
+      { /* A4OP */
+
+         __m128i d0 = _mm_load_si128((__m128i *)d);
+
+         __m128i mul0 = mul_256_sse3(alpha_packed, d0);
+         __m128i sym0 = mul_sym_sse3(_mm_srli_epi32(d0, 24), c_packed);
+
+         d0 = _mm_add_epi32(mul0, sym0);
+
+         _mm_store_si128((__m128i *)d, d0);
+
+         d += 4; l -= 4;
+      },
+      { /* A8OP */
+
+         __m128i d0 = _mm_load_si128((__m128i *)d);
+         __m128i d1 = _mm_load_si128((__m128i *)(d+4));
+
+         __m128i mul0 = mul_256_sse3(alpha_packed, d0);
+         __m128i mul1 = mul_256_sse3(alpha_packed, d1);
+
+         __m128i sym0 = mul_sym_sse3(_mm_srli_epi32(d0, 24), c_packed);
+         __m128i sym1 = mul_sym_sse3(_mm_srli_epi32(d1, 24), c_packed);
+
+         d0 = _mm_add_epi32(mul0, sym0);
+         d1 = _mm_add_epi32(mul1, sym1);
+
+         _mm_store_si128((__m128i *)d, d0);
+         _mm_store_si128((__m128i *)(d+4), d1);
+
+         d += 8; l -= 8;
+      })
+}
+
+#define _op_blend_rel_caa_dp_sse3 _op_blend_rel_c_dp_sse3
+#define _op_blend_rel_c_dpan_sse3 _op_blend_c_dpan_sse3
+#define _op_blend_rel_caa_dpan_sse3 _op_blend_caa_dpan_sse3
+
+static void
+init_blend_rel_color_span_funcs_sse3(void)
+{
+   op_blend_rel_span_funcs[SP_N][SM_N][SC][DP][CPU_SSE3] = _op_blend_rel_c_dp_sse3;
+   op_blend_rel_span_funcs[SP_N][SM_N][SC_AA][DP][CPU_SSE3] = _op_blend_rel_caa_dp_sse3;
+
+   op_blend_rel_span_funcs[SP_N][SM_N][SC][DP_AN][CPU_SSE3] = _op_blend_rel_c_dpan_sse3;
+   op_blend_rel_span_funcs[SP_N][SM_N][SC_AA][DP_AN][CPU_SSE3] = _op_blend_rel_caa_dpan_sse3;
+}
+
+#define _op_blend_rel_pt_c_dp_sse3 NULL
+#define _op_blend_rel_pt_caa_dp_sse3 _op_blend_rel_pt_c_dp_sse3
+
+#define _op_blend_rel_pt_c_dpan_sse3 _op_blend_pt_c_dpan_sse3
+#define _op_blend_rel_pt_caa_dpan_sse3 _op_blend_pt_caa_dpan_sse3
+
+static void
+init_blend_rel_color_pt_funcs_sse3(void)
+{
+   op_blend_rel_pt_funcs[SP_N][SM_N][SC][DP][CPU_SSE3] = _op_blend_rel_pt_c_dp_sse3;
+   op_blend_rel_pt_funcs[SP_N][SM_N][SC_AA][DP][CPU_SSE3] = _op_blend_rel_pt_caa_dp_sse3;
+
+   op_blend_rel_pt_funcs[SP_N][SM_N][SC][DP_AN][CPU_SSE3] = _op_blend_rel_pt_c_dpan_sse3;
+   op_blend_rel_pt_funcs[SP_N][SM_N][SC_AA][DP_AN][CPU_SSE3] = _op_blend_rel_pt_caa_dpan_sse3;
+}
+
+#endif