OpenSubdiv/opensubdiv/osd/clKernel.cl
David G. Yu 236afb2c06 Osd Varying and FaceVarying Patch Evaluation
Implemented EvalPatchesVarying and EvalPatchesFaceVarying
methods for Osd::*Evaluator classes, i.e. cpu, omp, tbb,
GLXFB, GLSLCompute, OpenCL, and CUDA.

Also, the GPU Kernel implementations have been updated to use
the common patchBasis implementation instead of re-implementing
methods to compute patch basis weights locally.
2016-09-29 09:56:15 -07:00

279 lines
8.1 KiB
Common Lisp

//
// Copyright 2013 Pixar
//
// Licensed under the Apache License, Version 2.0 (the "Apache License")
// with the following modification; you may not use this file except in
// compliance with the Apache License and the following modification to it:
// Section 6. Trademarks. is deleted and replaced with:
//
// 6. Trademarks. This License does not grant permission to use the trade
// names, trademarks, service marks, or product names of the Licensor
// and its affiliates, except as required to comply with Section 4(c) of
// the License and to reproduce the content of the NOTICE file.
//
// You may obtain a copy of the Apache License at
//
// http://www.apache.org/licenses/LICENSE-2.0
//
// Unless required by applicable law or agreed to in writing, software
// distributed under the Apache License with the above modification is
// distributed on an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
// KIND, either express or implied. See the Apache License for the specific
// language governing permissions and limitations under the Apache License.
//
struct Vertex {
float v[LENGTH];
};
static void clear(struct Vertex *vertex) {
for (int i = 0; i < LENGTH; i++) {
vertex->v[i] = 0.0f;
}
}
static void addWithWeight(struct Vertex *dst,
__global float *srcOrigin,
int index, float weight) {
__global float *src = srcOrigin + index * SRC_STRIDE;
for (int i = 0; i < LENGTH; ++i) {
dst->v[i] += src[i] * weight;
}
}
static void writeVertex(__global float *dstOrigin,
int index,
struct Vertex *src) {
__global float *dst = dstOrigin + index * DST_STRIDE;
for (int i = 0; i < LENGTH; ++i) {
dst[i] = src->v[i];
}
}
static void writeVertexStride(__global float *dstOrigin,
int index,
struct Vertex *src,
int stride) {
__global float *dst = dstOrigin + index * stride;
for (int i = 0; i < LENGTH; ++i) {
dst[i] = src->v[i];
}
}
__kernel void computeStencils(
__global float * src, int srcOffset,
__global float * dst, int dstOffset,
__global int * sizes,
__global int * offsets,
__global int * indices,
__global float * weights,
int batchStart, int batchEnd) {
int current = get_global_id(0) + batchStart;
if (current>=batchEnd) {
return;
}
struct Vertex v;
clear(&v);
int size = sizes[current],
offset = offsets[current];
src += srcOffset;
dst += dstOffset;
for (int i=0; i<size; ++i) {
addWithWeight(&v, src, indices[offset+i], weights[offset+i]);
}
writeVertex(dst, current, &v);
}
__kernel void computeStencilsDerivatives(
__global float * src, int srcOffset,
__global float * dst, int dstOffset,
__global float * du, int duOffset, int duStride,
__global float * dv, int dvOffset, int dvStride,
__global int * sizes,
__global int * offsets,
__global int * indices,
__global float * weights,
__global float * duWeights,
__global float * dvWeights,
int batchStart, int batchEnd) {
int current = get_global_id(0) + batchStart;
if (current>=batchEnd) {
return;
}
struct Vertex v, vdu, vdv;
clear(&v);
clear(&vdu);
clear(&vdv);
int size = sizes[current],
offset = offsets[current];
if (src) src += srcOffset;
if (dst) dst += dstOffset;
if (du) du += duOffset;
if (dv) dv += dvOffset;
for (int i=0; i<size; ++i) {
int ofs = offset + i;
int vid = indices[ofs];
if (weights) addWithWeight( &v, src, vid, weights[ofs]);
if (duWeights) addWithWeight(&vdu, src, vid, duWeights[ofs]);
if (dvWeights) addWithWeight(&vdv, src, vid, dvWeights[ofs]);
}
if (dst) writeVertex (dst, current, &v);
if (du) writeVertexStride(du, current, &vdu, duStride);
if (dv) writeVertexStride(dv, current, &vdv, dvStride);
}
// ---------------------------------------------------------------------------
struct PatchArray {
int patchType;
int numPatches;
int indexBase; // an offset within the index buffer
int primitiveIdBase; // an offset within the patch param buffer
};
struct PatchCoord {
int arrayIndex;
int patchIndex;
int vertIndex;
float s;
float t;
};
struct PatchParam {
uint field0;
uint field1;
float sharpness;
};
static int getDepth(uint patchBits) {
return (patchBits & 0xf);
}
static float getParamFraction(uint patchBits) {
bool nonQuadRoot = (patchBits >> 4) & 0x1;
int depth = getDepth(patchBits);
if (nonQuadRoot) {
return 1.0f / (float)( 1 << (depth-1) );
} else {
return 1.0f / (float)( 1 << depth );
}
}
static void normalizePatchCoord(uint patchBits, float *uv) {
float frac = getParamFraction(patchBits);
int iu = (patchBits >> 22) & 0x3ff;
int iv = (patchBits >> 12) & 0x3ff;
// top left corner
float pu = (float)iu*frac;
float pv = (float)iv*frac;
// normalize u,v coordinates
uv[0] = (uv[0] - pu) / frac;
uv[1] = (uv[1] - pv) / frac;
}
static bool isRegular(uint patchBits) {
return ((patchBits >> 5) & 0x1) != 0;
}
static int getNumControlVertices(int patchType) {
return (patchType == 3) ? 4 :
(patchType == 6) ? 16 :
(patchType == 9) ? 20 : 0;
}
__kernel void computePatches(__global float *src, int srcOffset,
__global float *dst, int dstOffset,
__global float *du, int duOffset, int duStride,
__global float *dv, int dvOffset, int dvStride,
__global struct PatchCoord *patchCoords,
__global struct PatchArray *patchArrayBuffer,
__global int *patchIndexBuffer,
__global struct PatchParam *patchParamBuffer) {
int current = get_global_id(0);
if (src) src += srcOffset;
if (dst) dst += dstOffset;
if (du) du += duOffset;
if (dv) dv += dvOffset;
struct PatchCoord coord = patchCoords[current];
struct PatchArray array = patchArrayBuffer[coord.arrayIndex];
uint patchBits = patchParamBuffer[coord.patchIndex].field1;
int patchType = isRegular(patchBits) ? 6 : array.patchType;
float uv[2] = {coord.s, coord.t};
normalizePatchCoord(patchBits, uv);
float dScale = (float)(1 << getDepth(patchBits));
int boundary = (patchBits >> 8) & 0xf;
float wP[20], wDs[20], wDt[20], wDss[20], wDst[20], wDtt[20];
int numControlVertices = 0;
if (patchType == 3) {
OsdGetBilinearPatchWeights(uv[0], uv[1], dScale,
wP, wDs, wDt, wDss, wDst, wDtt);
numControlVertices = 4;
} else if (patchType == 6) {
OsdGetBSplinePatchWeights(uv[0], uv[1], dScale, boundary,
wP, wDs, wDt, wDss, wDst, wDtt);
numControlVertices = 16;
} else if (patchType == 9) {
OsdGetGregoryPatchWeights(uv[0], uv[1], dScale,
wP, wDs, wDt, wDss, wDst, wDtt);
numControlVertices = 20;
}
int indexStride = getNumControlVertices(array.patchType);
int indexBase = array.indexBase + indexStride *
(coord.patchIndex - array.primitiveIdBase);
struct Vertex v;
clear(&v);
for (int i = 0; i < numControlVertices; ++i) {
int index = patchIndexBuffer[indexBase + i];
addWithWeight(&v, src, index, wP[i]);
}
writeVertex(dst, current, &v);
if (du) {
struct Vertex vdu;
clear(&vdu);
for (int i = 0; i < numControlVertices; ++i) {
int index = patchIndexBuffer[indexBase + i];
addWithWeight(&vdu, src, index, wDs[i]);
}
writeVertexStride(du, current, &vdu, duStride);
}
if (dv) {
struct Vertex vdv;
clear(&vdv);
for (int i = 0; i < numControlVertices; ++i) {
int index = patchIndexBuffer[indexBase + i];
addWithWeight(&vdv, src, index, wDt[i]);
}
writeVertexStride(dv, current, &vdv, dvStride);
}
}