change lcpp Lua preprocessor, to keep #defines and comments, remove empty lines
remove duplicate data in b3Contact4 (now in btContact4Data shared between CPU/C++ and OpenCL) OpenCL kernels use #include "Bullet3Collision/NarrowPhaseCollision/shared/b3Contact4Data.h" Increase number of batches back to 250 (from 50), need to fix this hard coded number (see https://github.com/erwincoumans/bullet3/issues/12) Work towards GJK/EPA, in addition to SAT/clipping (early on)
This commit is contained in:
erwincoumans
@@ -2,37 +2,71 @@
|
||||
static const char* solverSetup2CL= \
|
||||
"/*\n"
|
||||
"Copyright (c) 2012 Advanced Micro Devices, Inc. \n"
|
||||
"\n"
|
||||
"This software is provided 'as-is', without any express or implied warranty.\n"
|
||||
"In no event will the authors be held liable for any damages arising from the use of this software.\n"
|
||||
"Permission is granted to anyone to use this software for any purpose, \n"
|
||||
"including commercial applications, and to alter it and redistribute it freely, \n"
|
||||
"subject to the following restrictions:\n"
|
||||
"\n"
|
||||
"1. The origin of this software must not be misrepresented; you must not claim that you wrote the original software. If you use this software in a product, an acknowledgment in the product documentation would be appreciated but is not required.\n"
|
||||
"2. Altered source versions must be plainly marked as such, and must not be misrepresented as being the original software.\n"
|
||||
"3. This notice may not be removed or altered from any source distribution.\n"
|
||||
"*/\n"
|
||||
"//Originally written by Takahiro Harada\n"
|
||||
"\n"
|
||||
"\n"
|
||||
"#ifndef B3_CONTACT4DATA_H\n"
|
||||
"#define B3_CONTACT4DATA_H\n"
|
||||
"#ifndef B3_FLOAT4_H\n"
|
||||
"#define B3_FLOAT4_H\n"
|
||||
"#ifndef B3_PLATFORM_DEFINITIONS_H\n"
|
||||
"#define B3_PLATFORM_DEFINITIONS_H\n"
|
||||
"struct MyTest\n"
|
||||
"{\n"
|
||||
" int bla;\n"
|
||||
"};\n"
|
||||
"#endif\n"
|
||||
"#ifdef __cplusplus\n"
|
||||
"#else//bla\n"
|
||||
" typedef float4 b3Float4;\n"
|
||||
"#endif \n"
|
||||
"#endif //B3_FLOAT4_H\n"
|
||||
"typedef struct b3Contact4Data b3Contact4Data_t;\n"
|
||||
"struct b3Contact4Data\n"
|
||||
"{\n"
|
||||
" b3Float4 m_worldPos[4];\n"
|
||||
"// b3Float4 m_localPosB[4];\n"
|
||||
" b3Float4 m_worldNormal; // w: m_nPoints\n"
|
||||
" unsigned short m_restituitionCoeffCmp;\n"
|
||||
" unsigned short m_frictionCoeffCmp;\n"
|
||||
" int m_batchIdx;\n"
|
||||
" int m_bodyAPtrAndSignBit;//x:m_bodyAPtr, y:m_bodyBPtr\n"
|
||||
" int m_bodyBPtrAndSignBit;\n"
|
||||
" int m_childIndexA;\n"
|
||||
" int m_childIndexB;\n"
|
||||
" int m_unused1;\n"
|
||||
" int m_unused2;\n"
|
||||
" b3Float4 m_localPosA;\n"
|
||||
"};\n"
|
||||
"inline int b3Contact4Data_getNumPoints(const struct b3Contact4Data* contact)\n"
|
||||
"{\n"
|
||||
" return (int)contact->m_worldNormal.w;\n"
|
||||
"};\n"
|
||||
"inline void b3Contact4Data_setNumPoints(struct b3Contact4Data* contact, int numPoints)\n"
|
||||
"{\n"
|
||||
" contact->m_worldNormal.w = (float)numPoints;\n"
|
||||
"};\n"
|
||||
"#endif //B3_CONTACT4DATA_H\n"
|
||||
"#pragma OPENCL EXTENSION cl_amd_printf : enable\n"
|
||||
"#pragma OPENCL EXTENSION cl_khr_local_int32_base_atomics : enable\n"
|
||||
"#pragma OPENCL EXTENSION cl_khr_global_int32_base_atomics : enable\n"
|
||||
"#pragma OPENCL EXTENSION cl_khr_local_int32_extended_atomics : enable\n"
|
||||
"#pragma OPENCL EXTENSION cl_khr_global_int32_extended_atomics : enable\n"
|
||||
"\n"
|
||||
"\n"
|
||||
"#ifdef cl_ext_atomic_counters_32\n"
|
||||
"#pragma OPENCL EXTENSION cl_ext_atomic_counters_32 : enable\n"
|
||||
"#else\n"
|
||||
"#define counter32_t volatile global int*\n"
|
||||
"#endif\n"
|
||||
"\n"
|
||||
"typedef unsigned int u32;\n"
|
||||
"typedef unsigned short u16;\n"
|
||||
"typedef unsigned char u8;\n"
|
||||
"\n"
|
||||
"#define GET_GROUP_IDX get_group_id(0)\n"
|
||||
"#define GET_LOCAL_IDX get_local_id(0)\n"
|
||||
"#define GET_GLOBAL_IDX get_global_id(0)\n"
|
||||
@@ -46,22 +80,15 @@ static const char* solverSetup2CL= \
|
||||
"#define AtomAdd(x, value) atom_add(&(x), value)\n"
|
||||
"#define AtomCmpxhg(x, cmp, value) atom_cmpxchg( &(x), cmp, value )\n"
|
||||
"#define AtomXhg(x, value) atom_xchg ( &(x), value )\n"
|
||||
"\n"
|
||||
"\n"
|
||||
"#define SELECT_UINT4( b, a, condition ) select( b,a,condition )\n"
|
||||
"\n"
|
||||
"#define make_float4 (float4)\n"
|
||||
"#define make_float2 (float2)\n"
|
||||
"#define make_uint4 (uint4)\n"
|
||||
"#define make_int4 (int4)\n"
|
||||
"#define make_uint2 (uint2)\n"
|
||||
"#define make_int2 (int2)\n"
|
||||
"\n"
|
||||
"\n"
|
||||
"#define max2 max\n"
|
||||
"#define min2 min\n"
|
||||
"\n"
|
||||
"\n"
|
||||
"///////////////////////////////////////\n"
|
||||
"// Vector\n"
|
||||
"///////////////////////////////////////\n"
|
||||
@@ -71,52 +98,43 @@ static const char* solverSetup2CL= \
|
||||
" return native_divide(numerator, denominator); \n"
|
||||
"// return numerator/denominator; \n"
|
||||
"}\n"
|
||||
"\n"
|
||||
"__inline\n"
|
||||
"float4 fastDiv4(float4 numerator, float4 denominator)\n"
|
||||
"{\n"
|
||||
" return native_divide(numerator, denominator); \n"
|
||||
"}\n"
|
||||
"\n"
|
||||
"__inline\n"
|
||||
"float fastSqrtf(float f2)\n"
|
||||
"{\n"
|
||||
" return native_sqrt(f2);\n"
|
||||
"// return sqrt(f2);\n"
|
||||
"}\n"
|
||||
"\n"
|
||||
"__inline\n"
|
||||
"float fastRSqrt(float f2)\n"
|
||||
"{\n"
|
||||
" return native_rsqrt(f2);\n"
|
||||
"}\n"
|
||||
"\n"
|
||||
"__inline\n"
|
||||
"float fastLength4(float4 v)\n"
|
||||
"{\n"
|
||||
" return fast_length(v);\n"
|
||||
"}\n"
|
||||
"\n"
|
||||
"__inline\n"
|
||||
"float4 fastNormalize4(float4 v)\n"
|
||||
"{\n"
|
||||
" return fast_normalize(v);\n"
|
||||
"}\n"
|
||||
"\n"
|
||||
"\n"
|
||||
"__inline\n"
|
||||
"float sqrtf(float a)\n"
|
||||
"{\n"
|
||||
"// return sqrt(a);\n"
|
||||
" return native_sqrt(a);\n"
|
||||
"}\n"
|
||||
"\n"
|
||||
"__inline\n"
|
||||
"float4 cross3(float4 a, float4 b)\n"
|
||||
"{\n"
|
||||
" return cross(a,b);\n"
|
||||
"}\n"
|
||||
"\n"
|
||||
"__inline\n"
|
||||
"float dot3F4(float4 a, float4 b)\n"
|
||||
"{\n"
|
||||
@@ -124,26 +142,22 @@ static const char* solverSetup2CL= \
|
||||
" float4 b1 = make_float4(b.xyz,0.f);\n"
|
||||
" return dot(a1, b1);\n"
|
||||
"}\n"
|
||||
"\n"
|
||||
"__inline\n"
|
||||
"float length3(const float4 a)\n"
|
||||
"{\n"
|
||||
" return sqrtf(dot3F4(a,a));\n"
|
||||
"}\n"
|
||||
"\n"
|
||||
"__inline\n"
|
||||
"float dot4(const float4 a, const float4 b)\n"
|
||||
"{\n"
|
||||
" return dot( a, b );\n"
|
||||
"}\n"
|
||||
"\n"
|
||||
"// for height\n"
|
||||
"__inline\n"
|
||||
"float dot3w1(const float4 point, const float4 eqn)\n"
|
||||
"{\n"
|
||||
" return dot3F4(point,eqn) + eqn.w;\n"
|
||||
"}\n"
|
||||
"\n"
|
||||
"__inline\n"
|
||||
"float4 normalize3(const float4 a)\n"
|
||||
"{\n"
|
||||
@@ -152,14 +166,12 @@ static const char* solverSetup2CL= \
|
||||
"// float length = sqrtf(dot3F4(a, a));\n"
|
||||
"// return 1.f/length * a;\n"
|
||||
"}\n"
|
||||
"\n"
|
||||
"__inline\n"
|
||||
"float4 normalize4(const float4 a)\n"
|
||||
"{\n"
|
||||
" float length = sqrtf(dot4(a, a));\n"
|
||||
" return 1.f/length * a;\n"
|
||||
"}\n"
|
||||
"\n"
|
||||
"__inline\n"
|
||||
"float4 createEquation(const float4 a, const float4 b, const float4 c)\n"
|
||||
"{\n"
|
||||
@@ -170,34 +182,25 @@ static const char* solverSetup2CL= \
|
||||
" eqn.w = -dot3F4(eqn,a);\n"
|
||||
" return eqn;\n"
|
||||
"}\n"
|
||||
"\n"
|
||||
"///////////////////////////////////////\n"
|
||||
"// Matrix3x3\n"
|
||||
"///////////////////////////////////////\n"
|
||||
"\n"
|
||||
"typedef struct\n"
|
||||
"{\n"
|
||||
" float4 m_row[3];\n"
|
||||
"}Matrix3x3;\n"
|
||||
"\n"
|
||||
"__inline\n"
|
||||
"Matrix3x3 mtZero();\n"
|
||||
"\n"
|
||||
"__inline\n"
|
||||
"Matrix3x3 mtIdentity();\n"
|
||||
"\n"
|
||||
"__inline\n"
|
||||
"Matrix3x3 mtTranspose(Matrix3x3 m);\n"
|
||||
"\n"
|
||||
"__inline\n"
|
||||
"Matrix3x3 mtMul(Matrix3x3 a, Matrix3x3 b);\n"
|
||||
"\n"
|
||||
"__inline\n"
|
||||
"float4 mtMul1(Matrix3x3 a, float4 b);\n"
|
||||
"\n"
|
||||
"__inline\n"
|
||||
"float4 mtMul3(float4 a, Matrix3x3 b);\n"
|
||||
"\n"
|
||||
"__inline\n"
|
||||
"Matrix3x3 mtZero()\n"
|
||||
"{\n"
|
||||
@@ -207,7 +210,6 @@ static const char* solverSetup2CL= \
|
||||
" m.m_row[2] = (float4)(0.f);\n"
|
||||
" return m;\n"
|
||||
"}\n"
|
||||
"\n"
|
||||
"__inline\n"
|
||||
"Matrix3x3 mtIdentity()\n"
|
||||
"{\n"
|
||||
@@ -217,7 +219,6 @@ static const char* solverSetup2CL= \
|
||||
" m.m_row[2] = (float4)(0,0,1,0);\n"
|
||||
" return m;\n"
|
||||
"}\n"
|
||||
"\n"
|
||||
"__inline\n"
|
||||
"Matrix3x3 mtTranspose(Matrix3x3 m)\n"
|
||||
"{\n"
|
||||
@@ -227,7 +228,6 @@ static const char* solverSetup2CL= \
|
||||
" out.m_row[2] = (float4)(m.m_row[0].z, m.m_row[1].z, m.m_row[2].z, 0.f);\n"
|
||||
" return out;\n"
|
||||
"}\n"
|
||||
"\n"
|
||||
"__inline\n"
|
||||
"Matrix3x3 mtMul(Matrix3x3 a, Matrix3x3 b)\n"
|
||||
"{\n"
|
||||
@@ -248,7 +248,6 @@ static const char* solverSetup2CL= \
|
||||
" }\n"
|
||||
" return ans;\n"
|
||||
"}\n"
|
||||
"\n"
|
||||
"__inline\n"
|
||||
"float4 mtMul1(Matrix3x3 a, float4 b)\n"
|
||||
"{\n"
|
||||
@@ -259,43 +258,30 @@ static const char* solverSetup2CL= \
|
||||
" ans.w = 0.f;\n"
|
||||
" return ans;\n"
|
||||
"}\n"
|
||||
"\n"
|
||||
"__inline\n"
|
||||
"float4 mtMul3(float4 a, Matrix3x3 b)\n"
|
||||
"{\n"
|
||||
" float4 colx = make_float4(b.m_row[0].x, b.m_row[1].x, b.m_row[2].x, 0);\n"
|
||||
" float4 coly = make_float4(b.m_row[0].y, b.m_row[1].y, b.m_row[2].y, 0);\n"
|
||||
" float4 colz = make_float4(b.m_row[0].z, b.m_row[1].z, b.m_row[2].z, 0);\n"
|
||||
"\n"
|
||||
" float4 ans;\n"
|
||||
" ans.x = dot3F4( a, colx );\n"
|
||||
" ans.y = dot3F4( a, coly );\n"
|
||||
" ans.z = dot3F4( a, colz );\n"
|
||||
" return ans;\n"
|
||||
"}\n"
|
||||
"\n"
|
||||
"///////////////////////////////////////\n"
|
||||
"// Quaternion\n"
|
||||
"///////////////////////////////////////\n"
|
||||
"\n"
|
||||
"typedef float4 Quaternion;\n"
|
||||
"\n"
|
||||
"__inline\n"
|
||||
"Quaternion qtMul(Quaternion a, Quaternion b);\n"
|
||||
"\n"
|
||||
"__inline\n"
|
||||
"Quaternion qtNormalize(Quaternion in);\n"
|
||||
"\n"
|
||||
"__inline\n"
|
||||
"float4 qtRotate(Quaternion q, float4 vec);\n"
|
||||
"\n"
|
||||
"__inline\n"
|
||||
"Quaternion qtInvert(Quaternion q);\n"
|
||||
"\n"
|
||||
"\n"
|
||||
"\n"
|
||||
"\n"
|
||||
"\n"
|
||||
"__inline\n"
|
||||
"Quaternion qtMul(Quaternion a, Quaternion b)\n"
|
||||
"{\n"
|
||||
@@ -306,7 +292,6 @@ static const char* solverSetup2CL= \
|
||||
" ans.w = a.w*b.w - dot3F4(a, b);\n"
|
||||
" return ans;\n"
|
||||
"}\n"
|
||||
"\n"
|
||||
"__inline\n"
|
||||
"Quaternion qtNormalize(Quaternion in)\n"
|
||||
"{\n"
|
||||
@@ -323,43 +308,33 @@ static const char* solverSetup2CL= \
|
||||
" float4 out = qtMul(qtMul(q,vcpy),qInv);\n"
|
||||
" return out;\n"
|
||||
"}\n"
|
||||
"\n"
|
||||
"__inline\n"
|
||||
"Quaternion qtInvert(Quaternion q)\n"
|
||||
"{\n"
|
||||
" return (Quaternion)(-q.xyz, q.w);\n"
|
||||
"}\n"
|
||||
"\n"
|
||||
"__inline\n"
|
||||
"float4 qtInvRotate(const Quaternion q, float4 vec)\n"
|
||||
"{\n"
|
||||
" return qtRotate( qtInvert( q ), vec );\n"
|
||||
"}\n"
|
||||
"\n"
|
||||
"\n"
|
||||
"\n"
|
||||
"\n"
|
||||
"#define WG_SIZE 64\n"
|
||||
"\n"
|
||||
"typedef struct\n"
|
||||
"{\n"
|
||||
" float4 m_pos;\n"
|
||||
" Quaternion m_quat;\n"
|
||||
" float4 m_linVel;\n"
|
||||
" float4 m_angVel;\n"
|
||||
"\n"
|
||||
" u32 m_shapeIdx;\n"
|
||||
" float m_invMass;\n"
|
||||
" float m_restituitionCoeff;\n"
|
||||
" float m_frictionCoeff;\n"
|
||||
"} Body;\n"
|
||||
"\n"
|
||||
"typedef struct\n"
|
||||
"{\n"
|
||||
" Matrix3x3 m_invInertia;\n"
|
||||
" Matrix3x3 m_initInvInertia;\n"
|
||||
"} Shape;\n"
|
||||
"\n"
|
||||
"typedef struct\n"
|
||||
"{\n"
|
||||
" float4 m_linear;\n"
|
||||
@@ -368,34 +343,13 @@ static const char* solverSetup2CL= \
|
||||
" float m_jacCoeffInv[4];\n"
|
||||
" float m_b[4];\n"
|
||||
" float m_appliedRambdaDt[4];\n"
|
||||
"\n"
|
||||
" float m_fJacCoeffInv[2]; \n"
|
||||
" float m_fAppliedRambdaDt[2]; \n"
|
||||
"\n"
|
||||
" u32 m_bodyA;\n"
|
||||
" u32 m_bodyB;\n"
|
||||
"\n"
|
||||
" int m_batchIdx;\n"
|
||||
" u32 m_paddings[1];\n"
|
||||
"} Constraint4;\n"
|
||||
"\n"
|
||||
"typedef struct\n"
|
||||
"{\n"
|
||||
" float4 m_worldPos[4];\n"
|
||||
" float4 m_worldNormal;\n"
|
||||
" u32 m_coeffs;\n"
|
||||
" int m_batchIdx;\n"
|
||||
"\n"
|
||||
" int m_bodyAPtrAndSignBit;\n"
|
||||
" int m_bodyBPtrAndSignBit;\n"
|
||||
"\n"
|
||||
" int m_childIndexA;\n"
|
||||
" int m_childIndexB;\n"
|
||||
" int m_unused1;\n"
|
||||
" int m_unused2;\n"
|
||||
"\n"
|
||||
"} Contact4;\n"
|
||||
"\n"
|
||||
"typedef struct\n"
|
||||
"{\n"
|
||||
" int m_nConstraints;\n"
|
||||
@@ -404,7 +358,6 @@ static const char* solverSetup2CL= \
|
||||
" int m_nSplit;\n"
|
||||
"// int m_paddings[1];\n"
|
||||
"} ConstBuffer;\n"
|
||||
"\n"
|
||||
"typedef struct\n"
|
||||
"{\n"
|
||||
" int m_solveFriction;\n"
|
||||
@@ -413,47 +366,35 @@ static const char* solverSetup2CL= \
|
||||
" int m_nSplit;\n"
|
||||
"// int m_paddings[1];\n"
|
||||
"} ConstBufferBatchSolve;\n"
|
||||
"\n"
|
||||
"\n"
|
||||
" \n"
|
||||
"\n"
|
||||
"\n"
|
||||
"typedef struct \n"
|
||||
"{\n"
|
||||
" int m_valInt0;\n"
|
||||
" int m_valInt1;\n"
|
||||
" int m_valInt2;\n"
|
||||
" int m_valInt3;\n"
|
||||
"\n"
|
||||
" float m_val0;\n"
|
||||
" float m_val1;\n"
|
||||
" float m_val2;\n"
|
||||
" float m_val3;\n"
|
||||
"} SolverDebugInfo;\n"
|
||||
"\n"
|
||||
"\n"
|
||||
"\n"
|
||||
"\n"
|
||||
"// others\n"
|
||||
"__kernel\n"
|
||||
"__attribute__((reqd_work_group_size(WG_SIZE,1,1)))\n"
|
||||
"void ReorderContactKernel(__global Contact4* in, __global Contact4* out, __global int2* sortData, int4 cb )\n"
|
||||
"void ReorderContactKernel(__global struct b3Contact4Data* in, __global struct b3Contact4Data* out, __global int2* sortData, int4 cb )\n"
|
||||
"{\n"
|
||||
" int nContacts = cb.x;\n"
|
||||
" int gIdx = GET_GLOBAL_IDX;\n"
|
||||
"\n"
|
||||
" if( gIdx < nContacts )\n"
|
||||
" {\n"
|
||||
" int srcIdx = sortData[gIdx].y;\n"
|
||||
" out[gIdx] = in[srcIdx];\n"
|
||||
" }\n"
|
||||
"}\n"
|
||||
"\n"
|
||||
"__kernel __attribute__((reqd_work_group_size(WG_SIZE,1,1)))\n"
|
||||
"void SetDeterminismSortDataChildShapeB(__global Contact4* contactsIn, __global int2* sortDataOut, int nContacts)\n"
|
||||
"void SetDeterminismSortDataChildShapeB(__global struct b3Contact4Data* contactsIn, __global int2* sortDataOut, int nContacts)\n"
|
||||
"{\n"
|
||||
" int gIdx = GET_GLOBAL_IDX;\n"
|
||||
"\n"
|
||||
" if( gIdx < nContacts )\n"
|
||||
" {\n"
|
||||
" int2 sd;\n"
|
||||
@@ -462,12 +403,10 @@ static const char* solverSetup2CL= \
|
||||
" sortDataOut[gIdx] = sd;\n"
|
||||
" }\n"
|
||||
"}\n"
|
||||
"\n"
|
||||
"__kernel __attribute__((reqd_work_group_size(WG_SIZE,1,1)))\n"
|
||||
"void SetDeterminismSortDataChildShapeA(__global Contact4* contactsIn, __global int2* sortDataInOut, int nContacts)\n"
|
||||
"void SetDeterminismSortDataChildShapeA(__global struct b3Contact4Data* contactsIn, __global int2* sortDataInOut, int nContacts)\n"
|
||||
"{\n"
|
||||
" int gIdx = GET_GLOBAL_IDX;\n"
|
||||
"\n"
|
||||
" if( gIdx < nContacts )\n"
|
||||
" {\n"
|
||||
" int2 sdIn;\n"
|
||||
@@ -478,12 +417,10 @@ static const char* solverSetup2CL= \
|
||||
" sortDataInOut[gIdx] = sdOut;\n"
|
||||
" }\n"
|
||||
"}\n"
|
||||
"\n"
|
||||
"__kernel __attribute__((reqd_work_group_size(WG_SIZE,1,1)))\n"
|
||||
"void SetDeterminismSortDataBodyA(__global Contact4* contactsIn, __global int2* sortDataInOut, int nContacts)\n"
|
||||
"void SetDeterminismSortDataBodyA(__global struct b3Contact4Data* contactsIn, __global int2* sortDataInOut, int nContacts)\n"
|
||||
"{\n"
|
||||
" int gIdx = GET_GLOBAL_IDX;\n"
|
||||
"\n"
|
||||
" if( gIdx < nContacts )\n"
|
||||
" {\n"
|
||||
" int2 sdIn;\n"
|
||||
@@ -494,14 +431,11 @@ static const char* solverSetup2CL= \
|
||||
" sortDataInOut[gIdx] = sdOut;\n"
|
||||
" }\n"
|
||||
"}\n"
|
||||
"\n"
|
||||
"\n"
|
||||
"__kernel\n"
|
||||
"__attribute__((reqd_work_group_size(WG_SIZE,1,1)))\n"
|
||||
"void SetDeterminismSortDataBodyB(__global Contact4* contactsIn, __global int2* sortDataInOut, int nContacts)\n"
|
||||
"void SetDeterminismSortDataBodyB(__global struct b3Contact4Data* contactsIn, __global int2* sortDataInOut, int nContacts)\n"
|
||||
"{\n"
|
||||
" int gIdx = GET_GLOBAL_IDX;\n"
|
||||
"\n"
|
||||
" if( gIdx < nContacts )\n"
|
||||
" {\n"
|
||||
" int2 sdIn;\n"
|
||||
@@ -512,10 +446,6 @@ static const char* solverSetup2CL= \
|
||||
" sortDataInOut[gIdx] = sdOut;\n"
|
||||
" }\n"
|
||||
"}\n"
|
||||
"\n"
|
||||
"\n"
|
||||
"\n"
|
||||
"\n"
|
||||
"typedef struct\n"
|
||||
"{\n"
|
||||
" int m_nContacts;\n"
|
||||
@@ -523,8 +453,6 @@ static const char* solverSetup2CL= \
|
||||
" float m_scale;\n"
|
||||
" int m_nSplit;\n"
|
||||
"} ConstBufferSSD;\n"
|
||||
"\n"
|
||||
"\n"
|
||||
"static __constant const int gridTable4x4[] = \n"
|
||||
"{\n"
|
||||
" 0,1,17,16,\n"
|
||||
@@ -532,7 +460,6 @@ static const char* solverSetup2CL= \
|
||||
" 17,18,32,3,\n"
|
||||
" 16,19,3,34\n"
|
||||
"};\n"
|
||||
"\n"
|
||||
"static __constant const int gridTable8x8[] = \n"
|
||||
"{\n"
|
||||
" 0, 2, 3, 16, 17, 18, 19, 1,\n"
|
||||
@@ -545,18 +472,12 @@ static const char* solverSetup2CL= \
|
||||
" 197,27,214,213,212,199,198,196\n"
|
||||
" \n"
|
||||
"};\n"
|
||||
"\n"
|
||||
"\n"
|
||||
"\n"
|
||||
"\n"
|
||||
"#define USE_SPATIAL_BATCHING 1\n"
|
||||
"#define USE_4x4_GRID 1\n"
|
||||
"\n"
|
||||
"__kernel\n"
|
||||
"__attribute__((reqd_work_group_size(WG_SIZE,1,1)))\n"
|
||||
"void SetSortDataKernel(__global Contact4* gContact, __global Body* gBodies, __global int2* gSortDataOut, \n"
|
||||
"void SetSortDataKernel(__global struct b3Contact4Data* gContact, __global Body* gBodies, __global int2* gSortDataOut, \n"
|
||||
"int nContacts,float scale,int4 nSplit,int staticIdx)\n"
|
||||
"\n"
|
||||
"{\n"
|
||||
" int gIdx = GET_GLOBAL_IDX;\n"
|
||||
" \n"
|
||||
@@ -564,13 +485,10 @@ static const char* solverSetup2CL= \
|
||||
" {\n"
|
||||
" int aPtrAndSignBit = gContact[gIdx].m_bodyAPtrAndSignBit;\n"
|
||||
" int bPtrAndSignBit = gContact[gIdx].m_bodyBPtrAndSignBit;\n"
|
||||
"\n"
|
||||
" int aIdx = abs(aPtrAndSignBit );\n"
|
||||
" int bIdx = abs(bPtrAndSignBit);\n"
|
||||
"\n"
|
||||
" bool aStatic = (aPtrAndSignBit<0) ||(aPtrAndSignBit==staticIdx);\n"
|
||||
" bool bStatic = (bPtrAndSignBit<0) ||(bPtrAndSignBit==staticIdx);\n"
|
||||
"\n"
|
||||
"#if USE_SPATIAL_BATCHING \n"
|
||||
" int idx = (aStatic)? bIdx: aIdx;\n"
|
||||
" float4 p = gBodies[idx].m_pos;\n"
|
||||
@@ -587,7 +505,6 @@ static const char* solverSetup2CL= \
|
||||
" aa = bb;\n"
|
||||
" if (bStatic)\n"
|
||||
" bb = aa;\n"
|
||||
"\n"
|
||||
" int gridIndex = aa + bb*4;\n"
|
||||
" int newIndex = gridTable4x4[gridIndex];\n"
|
||||
" #else//USE_4x4_GRID\n"
|
||||
@@ -597,13 +514,10 @@ static const char* solverSetup2CL= \
|
||||
" aa = bb;\n"
|
||||
" if (bStatic)\n"
|
||||
" bb = aa;\n"
|
||||
"\n"
|
||||
" int gridIndex = aa + bb*8;\n"
|
||||
" int newIndex = gridTable8x8[gridIndex];\n"
|
||||
" #endif//USE_4x4_GRID\n"
|
||||
"#endif//USE_SPATIAL_BATCHING\n"
|
||||
"\n"
|
||||
"\n"
|
||||
" gSortDataOut[gIdx].x = newIndex;\n"
|
||||
" gSortDataOut[gIdx].y = gIdx;\n"
|
||||
" }\n"
|
||||
@@ -612,10 +526,9 @@ static const char* solverSetup2CL= \
|
||||
" gSortDataOut[gIdx].x = 0xffffffff;\n"
|
||||
" }\n"
|
||||
"}\n"
|
||||
"\n"
|
||||
"__kernel\n"
|
||||
"__attribute__((reqd_work_group_size(WG_SIZE,1,1)))\n"
|
||||
"void CopyConstraintKernel(__global Contact4* gIn, __global Contact4* gOut, int4 cb )\n"
|
||||
"void CopyConstraintKernel(__global struct b3Contact4Data* gIn, __global struct b3Contact4Data* gOut, int4 cb )\n"
|
||||
"{\n"
|
||||
" int gIdx = GET_GLOBAL_IDX;\n"
|
||||
" if( gIdx < cb.x )\n"
|
||||
@@ -623,7 +536,4 @@ static const char* solverSetup2CL= \
|
||||
" gOut[gIdx] = gIn[gIdx];\n"
|
||||
" }\n"
|
||||
"}\n"
|
||||
"\n"
|
||||
"\n"
|
||||
"\n"
|
||||
;
|
||||
|
||||
Reference in New Issue
Block a user