phynexis-deps/igl
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity
Files
3c1e35d568b815038eae2d94de1ba128932873fd
igl/external/embree/kernels/common/globals.cpp
T

298 lines
9.1 KiB
C++
Executable File
Raw Blame History

// ======================================================================== //
// Copyright 2009-2014 Intel Corporation //
// //
// Licensed under the Apache License, Version 2.0 (the "License"); //
// you may not use this file except in compliance with the License. //
// You may obtain a copy of the License at //
// //
// http://www.apache.org/licenses/LICENSE-2.0 //
// //
// Unless required by applicable law or agreed to in writing, software //
// distributed under the License is distributed on an "AS IS" BASIS, //
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. //
// See the License for the specific language governing permissions and //
// limitations under the License. //
// ======================================================================== //
#include "sys/platform.h"
#if defined(__MIC__)
#include "simd/mic.h"
#else
#include "simd/sse.h"
#endif
namespace embree
{
#if !defined(__MIC__)
struct __aligned(32) avxf
{
__forceinline avxf() { }
__forceinline avxf(float a) { for (size_t i=0; i<8; i++) v[i] = a; }
__forceinline avxf(StepTy ) { for (size_t i=0; i<8; i++) v[i] = i; }
friend __forceinline const avxf operator-(const avxf& a) { avxf r; for (size_t i=0; i<8; i++) r.v[i] = -a.v[i]; return r; }
friend __forceinline const avxf operator+(const avxf& a, const avxf& b) { avxf r; for (size_t i=0; i<8; i++) r.v[i] = a.v[i] + b.v[i]; return r; }
friend __forceinline const avxf operator-(const avxf& a, const avxf& b) { avxf r; for (size_t i=0; i<8; i++) r.v[i] = a.v[i] - b.v[i]; return r; }
friend __forceinline const avxf operator*(const avxf& a, const avxf& b) { avxf r; for (size_t i=0; i<8; i++) r.v[i] = a.v[i] * b.v[i]; return r; }
friend __forceinline const avxf operator*(const float a, const avxf& b) { avxf r; for (size_t i=0; i<8; i++) r.v[i] = a * b.v[i]; return r; }
friend __forceinline const avxf operator*(const avxf& a, const float b) { avxf r; for (size_t i=0; i<8; i++) r.v[i] = a.v[i] * b ; return r; }
friend __forceinline const avxf operator/(const avxf& a, const avxf& b) { avxf r; for (size_t i=0; i<8; i++) r.v[i] = a.v[i] / b.v[i]; return r; }
float v[8];
};
avxf coeff0[4];
avxf coeff1[4];
avxf coeff_P0[4];
avxf coeff_P1[4];
avxf coeff_P2[4];
avxf coeff_P3[4];
ssef sse_coeff0[4];
ssef sse_coeff1[4];
ssef sse_coeff_P0[4];
ssef sse_coeff_P1[4];
ssef sse_coeff_P2[4];
ssef sse_coeff_P3[4];
#else
mic_f coeff0[4];
mic_f coeff1[4];
mic_f coeff01[4];
mic_f coeff_P0[4];
mic_f coeff_P1[4];
mic_f coeff_P2[4];
mic_f coeff_P3[4];
#endif
void init_globals()
{
#if defined(__MIC__)
{
__aligned(64) float STEP[16] = {0/16.0f,1/16.0f,2/16.0f,3/16.0f,4/16.0f,5/16.0f,6/16.0f,7/16.0f,8/16.0f,9/16.0f,10/16.0f,11/16.0f,12/16.0f,13/16.0f,14/16.0f,15/16.0f};
mic_f step16 = load16f(STEP);
const mic_f t1 = step16;
const mic_f t0 = 1.0f-t1;
coeff01[0] = t0 * t0 * t0;
coeff01[1] = 3.0f * t1 * t0 * t0;
coeff01[2] = 3.0f * t1 * t1 * t0;
coeff01[3] = t1 * t1 * t1;
}
mic_f step16 = mic_f(step);
mic_f dt = 1.0f/16.0f;
{
const mic_f t1 = step16*dt;
const mic_f t0 = 1.0f-t1;
coeff0[0] = t0 * t0 * t0;
coeff0[1] = 3.0f * t1 * t0 * t0;
coeff0[2] = 3.0f * t1 * t1 * t0;
coeff0[3] = t1 * t1 * t1;
}
{
const mic_f t1 = step16*dt+mic_f(dt);
const mic_f t0 = 1.0f-t1;
coeff1[0] = t0 * t0 * t0;
coeff1[1] = 3.0f * t1 * t0 * t0;
coeff1[2] = 3.0f * t1 * t1 * t0;
coeff1[3] = t1 * t1 * t1;
}
{
const mic_f t1 = step16*dt;
const mic_f t0 = 1.0f-t1;
coeff_P0[0] = t0 * t0 * t0;
coeff_P0[1] = 3.0f * t1 * t0 * t0;
coeff_P0[2] = 3.0f * t1 * t1 * t0;
coeff_P0[3] = t1 * t1 * t1;
}
{
const mic_f t1 = step16*dt;
const mic_f t0 = 1.0f-t1;
const mic_f d0 = -t0*t0;
const mic_f d1 = t0*t0 - 2.0f*t0*t1;
const mic_f d2 = 2.0f*t1*t0 - t1*t1;
const mic_f d3 = t1*t1;
coeff_P1[0] = coeff_P0[0] + dt*d0;
coeff_P1[1] = coeff_P0[1] + dt*d1;
coeff_P1[2] = coeff_P0[2] + dt*d2;
coeff_P1[3] = coeff_P0[3] + dt*d3;
}
{
const mic_f t1 = step16*dt+mic_f(dt);
const mic_f t0 = 1.0f-t1;
coeff_P3[0] = t0 * t0 * t0;
coeff_P3[1] = 3.0f * t1 * t0 * t0;
coeff_P3[2] = 3.0f * t1 * t1 * t0;
coeff_P3[3] = t1 * t1 * t1;
}
{
const mic_f t1 = step16*dt+mic_f(dt);
const mic_f t0 = 1.0f-t1;
const mic_f d0 = -t0*t0;
const mic_f d1 = t0*t0 - 2.0f*t0*t1;
const mic_f d2 = 2.0f*t1*t0 - t1*t1;
const mic_f d3 = t1*t1;
coeff_P2[0] = coeff_P3[0] - dt*d0;
coeff_P2[1] = coeff_P3[1] - dt*d1;
coeff_P2[2] = coeff_P3[2] - dt*d2;
coeff_P2[3] = coeff_P3[3] - dt*d3;
}
#else
{
const float dt = 1.0f/8.0f;
const avxf t1 = avxf(step)*dt;
const avxf t0 = 1.0f-t1;
coeff0[0] = t0 * t0 * t0;
coeff0[1] = 3.0f * t1 * t0 * t0;
coeff0[2] = 3.0f * t1 * t1 * t0;
coeff0[3] = t1 * t1 * t1;
}
{
const float dt = 1.0f/8.0f;
const avxf t1 = avxf(step)*dt+avxf(dt);
const avxf t0 = 1.0f-t1;
coeff1[0] = t0 * t0 * t0;
coeff1[1] = 3.0f * t1 * t0 * t0;
coeff1[2] = 3.0f * t1 * t1 * t0;
coeff1[3] = t1 * t1 * t1;
}
{
const float dt = 1.0f/8.0f;
const avxf t1 = avxf(step)*dt;
const avxf t0 = 1.0f-t1;
coeff_P0[0] = t0 * t0 * t0;
coeff_P0[1] = 3.0f * t1 * t0 * t0;
coeff_P0[2] = 3.0f * t1 * t1 * t0;
coeff_P0[3] = t1 * t1 * t1;
}
{
const float dt = 1.0f/8.0f;
const avxf t1 = avxf(step)*dt;
const avxf t0 = 1.0f-t1;
const avxf d0 = -t0*t0;
const avxf d1 = t0*t0 - 2.0f*t0*t1;
const avxf d2 = 2.0f*t1*t0 - t1*t1;
const avxf d3 = t1*t1;
coeff_P1[0] = coeff_P0[0] + dt*d0;
coeff_P1[1] = coeff_P0[1] + dt*d1;
coeff_P1[2] = coeff_P0[2] + dt*d2;
coeff_P1[3] = coeff_P0[3] + dt*d3;
}
{
const float dt = 1.0f/8.0f;
const avxf t1 = avxf(step)*dt+avxf(dt);
const avxf t0 = 1.0f-t1;
coeff_P3[0] = t0 * t0 * t0;
coeff_P3[1] = 3.0f * t1 * t0 * t0;
coeff_P3[2] = 3.0f * t1 * t1 * t0;
coeff_P3[3] = t1 * t1 * t1;
}
{
const float dt = 1.0f/8.0f;
const avxf t1 = avxf(step)*dt+avxf(dt);
const avxf t0 = 1.0f-t1;
const avxf d0 = -t0*t0;
const avxf d1 = t0*t0 - 2.0f*t0*t1;
const avxf d2 = 2.0f*t1*t0 - t1*t1;
const avxf d3 = t1*t1;
coeff_P2[0] = coeff_P3[0] - dt*d0;
coeff_P2[1] = coeff_P3[1] - dt*d1;
coeff_P2[2] = coeff_P3[2] - dt*d2;
coeff_P2[3] = coeff_P3[3] - dt*d3;
}
////////////////////////// SSE ////////////////////////
{
const float dt = 1.0f/4.0f;
const ssef t1 = ssef(step)*dt;
const ssef t0 = 1.0f-t1;
sse_coeff0[0] = t0 * t0 * t0;
sse_coeff0[1] = 3.0f * t1 * t0 * t0;
sse_coeff0[2] = 3.0f * t1 * t1 * t0;
sse_coeff0[3] = t1 * t1 * t1;
}
{
const float dt = 1.0f/4.0f;
const ssef t1 = ssef(step)*dt+ssef(dt);
const ssef t0 = 1.0f-t1;
sse_coeff1[0] = t0 * t0 * t0;
sse_coeff1[1] = 3.0f * t1 * t0 * t0;
sse_coeff1[2] = 3.0f * t1 * t1 * t0;
sse_coeff1[3] = t1 * t1 * t1;
}
{
const float dt = 1.0f/4.0f;
const ssef t1 = ssef(step)*dt;
const ssef t0 = 1.0f-t1;
sse_coeff_P0[0] = t0 * t0 * t0;
sse_coeff_P0[1] = 3.0f * t1 * t0 * t0;
sse_coeff_P0[2] = 3.0f * t1 * t1 * t0;
sse_coeff_P0[3] = t1 * t1 * t1;
}
{
const float dt = 1.0f/4.0f;
const ssef t1 = ssef(step)*dt;
const ssef t0 = 1.0f-t1;
const ssef d0 = -t0*t0;
const ssef d1 = t0*t0 - 2.0f*t0*t1;
const ssef d2 = 2.0f*t1*t0 - t1*t1;
const ssef d3 = t1*t1;
sse_coeff_P1[0] = sse_coeff_P0[0] + dt*d0;
sse_coeff_P1[1] = sse_coeff_P0[1] + dt*d1;
sse_coeff_P1[2] = sse_coeff_P0[2] + dt*d2;
sse_coeff_P1[3] = sse_coeff_P0[3] + dt*d3;
}
{
const float dt = 1.0f/4.0f;
const ssef t1 = ssef(step)*dt+ssef(dt);
const ssef t0 = 1.0f-t1;
sse_coeff_P3[0] = t0 * t0 * t0;
sse_coeff_P3[1] = 3.0f * t1 * t0 * t0;
sse_coeff_P3[2] = 3.0f * t1 * t1 * t0;
sse_coeff_P3[3] = t1 * t1 * t1;
}
{
const float dt = 1.0f/4.0f;
const ssef t1 = ssef(step)*dt+ssef(dt);
const ssef t0 = 1.0f-t1;
const ssef d0 = -t0*t0;
const ssef d1 = t0*t0 - 2.0f*t0*t1;
const ssef d2 = 2.0f*t1*t0 - t1*t1;
const ssef d3 = t1*t1;
sse_coeff_P2[0] = sse_coeff_P3[0] - dt*d0;
sse_coeff_P2[1] = sse_coeff_P3[1] - dt*d1;
sse_coeff_P2[2] = sse_coeff_P3[2] - dt*d2;
sse_coeff_P2[3] = sse_coeff_P3[3] - dt*d3;
}
#endif
}
}
Reference in New Issue View Git Blame Copy Permalink