Bart/bullet3
1
0 Fork 0
Code Issues Pull Requests Releases Wiki Activity

Make is easier to compile Bullet/Extras/sph:

Browse Source 
1) use $(CUDA_INC_PATH) for CUDA include path
2) avoid using cuda util, but use BT_GPU_SAFE_CALL and BT_GPU_CHECK_ERROR (defined in Bullet/Extras/CUDA/btCudaUtil.h)
3) remove pre-linked glee*.lib, but include glee.c directly
This commit is contained in:
erwin.coumans committed 2009-06-10 22:44:27 +00:00
1 parent 656d6b8336
commit 47a4435a9a
7 files changed
+17703 -40
Download Patch File Download Diff File Expand all files Collapse all files

No files matched your search

Extras/sph/common/GLee.h
+17647
View File
File diff suppressed because it is too large. Load diff
Extras/sph/common/gl_helper.h
+1 -1
View File
@@ -25,7 +25,7 @@
#include "common_defs.h" #include "common_defs.h"
#include <gl/glee.h> #include <glee.h>
#include <gl/glext.h> #include <gl/glext.h>
#ifdef _MSC_VER // Windows #ifdef _MSC_VER // Windows
Extras/sph/fluids.vcproj
+6 -2
View File
@@ -125,7 +125,7 @@
EnableIntrinsicFunctions="true" EnableIntrinsicFunctions="true"
FavorSizeOrSpeed="1" FavorSizeOrSpeed="1"
OmitFramePointers="true" OmitFramePointers="true"
AdditionalIncludeDirectories="fluids;common;marching;sphere_scan;marching_tris" AdditionalIncludeDirectories="fluids;common;marching;sphere_scan;marching_tris;&quot;$(CUDA_INC_PATH)&quot;;../../Glut"
PreprocessorDefinitions="WIN32;NDEBUG;_CONSOLE" PreprocessorDefinitions="WIN32;NDEBUG;_CONSOLE"
StringPooling="true" StringPooling="true"
RuntimeLibrary="0" RuntimeLibrary="0"
@@ -146,9 +146,9 @@
/> />
<Tool <Tool
Name="VCLinkerTool" Name="VCLinkerTool"
AdditionalDependencies="glee_2008.lib"
OutputFile="$(OutDir)/fluids.exe" OutputFile="$(OutDir)/fluids.exe"
LinkIncremental="1" LinkIncremental="1"
AdditionalLibraryDirectories="../../Glut"
IgnoreDefaultLibraryNames="libcmtd.lib" IgnoreDefaultLibraryNames="libcmtd.lib"
GenerateDebugInformation="false" GenerateDebugInformation="false"
SubSystem="1" SubSystem="1"
@@ -292,6 +292,10 @@
> >
</File> </File>
</Filter> </Filter>
<File
RelativePath=".\common\GLee.c"
>
</File>
<File <File
RelativePath=".\main.cpp" RelativePath=".\main.cpp"
> >
Extras/sph/fluids/fluid_system.cpp
+1 -1
View File
@@ -134,7 +134,7 @@ int FluidSystem::AddPointReuse ()
void FluidSystem::Run () void FluidSystem::Run ()
{ {
bool bTiming = true; bool bTiming = false;//true;
mint::Time start, stop; mint::Time start, stop;
Extras/sph/fluids/fluid_system_host.cu
+35 -32
View File
@@ -21,8 +21,10 @@
*/ */
#include "C:\CUDA\common\inc\cutil.h" // cutil32.lib //#include "C:\CUDA\common\inc\cutil.h" // cutil32.lib
#include <string.h> #include <string.h>
#include "../CUDA/btCudaDefines.h"
#if defined(__APPLE__) || defined(MACOSX) #if defined(__APPLE__) || defined(MACOSX)
@@ -43,12 +45,13 @@ __device__ uint* bufHash[2]; // point grid hash
__device__ int* bufGrid; __device__ int* bufGrid;
extern "C" extern "C"
{ {
// Initialize CUDA // Initialize CUDA
void cudaInit(int argc, char **argv) void cudaInit(int argc, char **argv)
{ {
CUT_DEVICE_INIT(argc, argv); //CUT_DEVICE_INIT(argc, argv);
cudaDeviceProp p; cudaDeviceProp p;
cudaGetDeviceProperties ( &p, 0); cudaGetDeviceProperties ( &p, 0);
@@ -65,10 +68,10 @@ void cudaInit(int argc, char **argv)
printf ( "Const Mem: %d\n", p.totalConstMem ); printf ( "Const Mem: %d\n", p.totalConstMem );
printf ( "Clock Rate: %d\n", p.clockRate ); printf ( "Clock Rate: %d\n", p.clockRate );
CUDA_SAFE_CALL ( cudaMalloc ( (void**) &bufPnts, 10 ) ); BT_GPU_SAFE_CALL ( cudaMalloc ( (void**) &bufPnts, 10 ) );
CUDA_SAFE_CALL ( cudaMalloc ( (void**) &bufPntSort, 10 ) ); BT_GPU_SAFE_CALL ( cudaMalloc ( (void**) &bufPntSort, 10 ) );
CUDA_SAFE_CALL ( cudaMalloc ( (void**) &bufHash, 10 ) ); BT_GPU_SAFE_CALL ( cudaMalloc ( (void**) &bufHash, 10 ) );
CUDA_SAFE_CALL ( cudaMalloc ( (void**) &bufGrid, 10 ) ); BT_GPU_SAFE_CALL ( cudaMalloc ( (void**) &bufGrid, 10 ) );
}; };
// Compute number of blocks to create // Compute number of blocks to create
@@ -83,11 +86,11 @@ void computeNumBlocks (int numPnts, int maxThreads, int &numBlocks, int &numThre
void FluidClearCUDA () void FluidClearCUDA ()
{ {
CUDA_SAFE_CALL ( cudaFree ( bufPnts ) ); BT_GPU_SAFE_CALL ( cudaFree ( bufPnts ) );
CUDA_SAFE_CALL ( cudaFree ( bufPntSort ) ); BT_GPU_SAFE_CALL ( cudaFree ( bufPntSort ) );
CUDA_SAFE_CALL ( cudaFree ( bufHash[0] ) ); BT_GPU_SAFE_CALL ( cudaFree ( bufHash[0] ) );
CUDA_SAFE_CALL ( cudaFree ( bufHash[1] ) ); BT_GPU_SAFE_CALL ( cudaFree ( bufHash[1] ) );
CUDA_SAFE_CALL ( cudaFree ( bufGrid ) ); BT_GPU_SAFE_CALL ( cudaFree ( bufGrid ) );
} }
@@ -114,11 +117,11 @@ void FluidSetupCUDA ( int num, int stride, float3 min, float3 max, float3 res, f
printf ( "pnts: %d, t:%dx%d=%d, bufPnts:%d, bufHash:%d\n", fcuda.pnts, fcuda.numBlocks, fcuda.numThreads, fcuda.numBlocks*fcuda.numThreads, fcuda.szPnts, fcuda.szHash ); printf ( "pnts: %d, t:%dx%d=%d, bufPnts:%d, bufHash:%d\n", fcuda.pnts, fcuda.numBlocks, fcuda.numThreads, fcuda.numBlocks*fcuda.numThreads, fcuda.szPnts, fcuda.szHash );
printf ( "grds: %d, t:%dx%d=%d, bufGrid:%d, Res: %dx%dx%d\n", fcuda.cells, fcuda.gridBlocks, fcuda.gridThreads, fcuda.gridBlocks*fcuda.gridThreads, fcuda.szGrid, (int) fcuda.res.x, (int) fcuda.res.y, (int) fcuda.res.z ); printf ( "grds: %d, t:%dx%d=%d, bufGrid:%d, Res: %dx%dx%d\n", fcuda.cells, fcuda.gridBlocks, fcuda.gridThreads, fcuda.gridBlocks*fcuda.gridThreads, fcuda.szGrid, (int) fcuda.res.x, (int) fcuda.res.y, (int) fcuda.res.z );
CUDA_SAFE_CALL ( cudaMalloc ( (void**) &bufPnts, fcuda.szPnts ) ); BT_GPU_SAFE_CALL ( cudaMalloc ( (void**) &bufPnts, fcuda.szPnts ) );
CUDA_SAFE_CALL ( cudaMalloc ( (void**) &bufPntSort, fcuda.szPnts ) ); BT_GPU_SAFE_CALL ( cudaMalloc ( (void**) &bufPntSort, fcuda.szPnts ) );
CUDA_SAFE_CALL ( cudaMalloc ( (void**) &bufHash[0], fcuda.szHash ) ); BT_GPU_SAFE_CALL ( cudaMalloc ( (void**) &bufHash[0], fcuda.szHash ) );
CUDA_SAFE_CALL ( cudaMalloc ( (void**) &bufHash[1], fcuda.szHash ) ); BT_GPU_SAFE_CALL ( cudaMalloc ( (void**) &bufHash[1], fcuda.szHash ) );
CUDA_SAFE_CALL ( cudaMalloc ( (void**) &bufGrid, fcuda.szGrid ) ); BT_GPU_SAFE_CALL ( cudaMalloc ( (void**) &bufGrid, fcuda.szGrid ) );
printf ( "POINTERS\n"); printf ( "POINTERS\n");
printf ( "bufPnts: %p\n", bufPnts ); printf ( "bufPnts: %p\n", bufPnts );
@@ -127,7 +130,7 @@ void FluidSetupCUDA ( int num, int stride, float3 min, float3 max, float3 res, f
printf ( "bufHash1: %p\n", bufHash[1] ); printf ( "bufHash1: %p\n", bufHash[1] );
printf ( "bufGrid: %p\n", bufGrid ); printf ( "bufGrid: %p\n", bufGrid );
CUDA_SAFE_CALL ( cudaMemcpyToSymbol ( simData, &fcuda, sizeof(FluidParams) ) ); BT_GPU_SAFE_CALL ( cudaMemcpyToSymbol ( simData, &fcuda, sizeof(FluidParams) ) );
cudaThreadSynchronize (); cudaThreadSynchronize ();
} }
@@ -146,46 +149,46 @@ void FluidParamCUDA ( float sim_scale, float smooth_rad, float mass, float rest,
fcuda.spikykern = -45.0f / (3.141592 * pow( smooth_rad, 6.0f) ); fcuda.spikykern = -45.0f / (3.141592 * pow( smooth_rad, 6.0f) );
fcuda.lapkern = 45.0f / (3.141592 * pow( smooth_rad, 6.0f) ); fcuda.lapkern = 45.0f / (3.141592 * pow( smooth_rad, 6.0f) );
CUDA_SAFE_CALL( cudaMemcpyToSymbol ( simData, &fcuda, sizeof(FluidParams) ) ); BT_GPU_SAFE_CALL( cudaMemcpyToSymbol ( simData, &fcuda, sizeof(FluidParams) ) );
cudaThreadSynchronize (); cudaThreadSynchronize ();
} }
void TransferToCUDA ( char* data, int* grid, int numPoints ) void TransferToCUDA ( char* data, int* grid, int numPoints )
{ {
CUDA_SAFE_CALL( cudaMemcpy ( bufPnts, data, numPoints * fcuda.stride, cudaMemcpyHostToDevice ) ); BT_GPU_SAFE_CALL( cudaMemcpy ( bufPnts, data, numPoints * fcuda.stride, cudaMemcpyHostToDevice ) );
cudaThreadSynchronize (); cudaThreadSynchronize ();
} }
void TransferFromCUDA ( char* data, int* grid, int numPoints ) void TransferFromCUDA ( char* data, int* grid, int numPoints )
{ {
CUDA_SAFE_CALL( cudaMemcpy ( data, bufPntSort, numPoints * fcuda.stride, cudaMemcpyDeviceToHost ) ); BT_GPU_SAFE_CALL( cudaMemcpy ( data, bufPntSort, numPoints * fcuda.stride, cudaMemcpyDeviceToHost ) );
cudaThreadSynchronize (); cudaThreadSynchronize ();
CUDA_SAFE_CALL( cudaMemcpy ( grid, bufGrid, fcuda.cells * sizeof(uint), cudaMemcpyDeviceToHost ) ); BT_GPU_SAFE_CALL( cudaMemcpy ( grid, bufGrid, fcuda.cells * sizeof(uint), cudaMemcpyDeviceToHost ) );
} }
void Grid_InsertParticlesCUDA () void Grid_InsertParticlesCUDA ()
{ {
CUDA_SAFE_CALL( cudaMemset ( bufHash[0], 0, fcuda.szHash ) ); BT_GPU_SAFE_CALL( cudaMemset ( bufHash[0], 0, fcuda.szHash ) );
hashParticles<<< fcuda.numBlocks, fcuda.numThreads>>> ( bufPnts, (uint2*) bufHash[0], fcuda.pnts ); hashParticles<<< fcuda.numBlocks, fcuda.numThreads>>> ( bufPnts, (uint2*) bufHash[0], fcuda.pnts );
CUT_CHECK_ERROR( "Kernel execution failed"); BT_GPU_CHECK_ERROR( "Kernel execution failed");
cudaThreadSynchronize (); cudaThreadSynchronize ();
//int buf[20000]; //int buf[20000];
/*printf ( "HASH: %d (%d)\n", fcuda.pnts, fcuda.numBlocks*fcuda.numThreads ); /*printf ( "HASH: %d (%d)\n", fcuda.pnts, fcuda.numBlocks*fcuda.numThreads );
CUDA_SAFE_CALL( cudaMemcpy ( buf, bufHash[0], fcuda.pnts * 2*sizeof(uint), cudaMemcpyDeviceToHost ) ); BT_GPU_SAFE_CALL( cudaMemcpy ( buf, bufHash[0], fcuda.pnts * 2*sizeof(uint), cudaMemcpyDeviceToHost ) );
//for (int n=0; n < fcuda.numBlocks*fcuda.numThreads; n++) { //for (int n=0; n < fcuda.numBlocks*fcuda.numThreads; n++) {
for (int n=0; n < 100; n++) { for (int n=0; n < 100; n++) {
printf ( "%d: <%d,%d>\n", n, buf[n*2], buf[n*2+1] ); printf ( "%d: <%d,%d>\n", n, buf[n*2], buf[n*2+1] );
}*/ }*/
RadixSort( (KeyValuePair *) bufHash[0], (KeyValuePair *) bufHash[1], fcuda.pnts, 32); RadixSort( (KeyValuePair *) bufHash[0], (KeyValuePair *) bufHash[1], fcuda.pnts, 32);
CUT_CHECK_ERROR( "Kernel execution failed"); BT_GPU_CHECK_ERROR( "Kernel execution failed");
cudaThreadSynchronize (); cudaThreadSynchronize ();
/*printf ( "HASH: %d (%d)\n", fcuda.pnts, fcuda.numBlocks*fcuda.numThreads ); /*printf ( "HASH: %d (%d)\n", fcuda.pnts, fcuda.numBlocks*fcuda.numThreads );
CUDA_SAFE_CALL( cudaMemcpy ( buf, bufHash[0], fcuda.pnts * 2*sizeof(uint), cudaMemcpyDeviceToHost ) ); BT_GPU_SAFE_CALL( cudaMemcpy ( buf, bufHash[0], fcuda.pnts * 2*sizeof(uint), cudaMemcpyDeviceToHost ) );
//for (int n=0; n < fcuda.numBlocks*fcuda.numThreads; n++) { //for (int n=0; n < fcuda.numBlocks*fcuda.numThreads; n++) {
for (int n=0; n < 100; n++) { for (int n=0; n < 100; n++) {
printf ( "%d: <%d,%d>\n", n, buf[n*2], buf[n*2+1] ); printf ( "%d: <%d,%d>\n", n, buf[n*2], buf[n*2+1] );
@@ -193,14 +196,14 @@ void Grid_InsertParticlesCUDA ()
// insertParticles<<< fcuda.gridBlocks, fcuda.gridThreads>>> ( bufPnts, (uint2*) bufHash[0], bufGrid, fcuda.pnts, fcuda.cells ); // insertParticles<<< fcuda.gridBlocks, fcuda.gridThreads>>> ( bufPnts, (uint2*) bufHash[0], bufGrid, fcuda.pnts, fcuda.cells );
CUDA_SAFE_CALL( cudaMemset ( bufGrid, NULL_HASH, fcuda.cells * sizeof(uint) ) ); BT_GPU_SAFE_CALL( cudaMemset ( bufGrid, NULL_HASH, fcuda.cells * sizeof(uint) ) );
insertParticlesRadix<<< fcuda.numBlocks, fcuda.numThreads>>> ( bufPnts, (uint2*) bufHash[0], bufGrid, bufPntSort, fcuda.pnts, fcuda.cells ); insertParticlesRadix<<< fcuda.numBlocks, fcuda.numThreads>>> ( bufPnts, (uint2*) bufHash[0], bufGrid, bufPntSort, fcuda.pnts, fcuda.cells );
CUT_CHECK_ERROR( "Kernel execution failed"); BT_GPU_CHECK_ERROR( "Kernel execution failed");
cudaThreadSynchronize (); cudaThreadSynchronize ();
/*printf ( "GRID: %d\n", fcuda.cells ); /*printf ( "GRID: %d\n", fcuda.cells );
CUDA_SAFE_CALL( cudaMemcpy ( buf, bufGrid, fcuda.cells * sizeof(uint), cudaMemcpyDeviceToHost ) ); BT_GPU_SAFE_CALL( cudaMemcpy ( buf, bufGrid, fcuda.cells * sizeof(uint), cudaMemcpyDeviceToHost ) );
*for (int n=0; n < 100; n++) { *for (int n=0; n < 100; n++) {
printf ( "%d: %d\n", n, buf[n]); printf ( "%d: %d\n", n, buf[n]);
}*/ }*/
@@ -209,7 +212,7 @@ void Grid_InsertParticlesCUDA ()
void SPH_ComputePressureCUDA () void SPH_ComputePressureCUDA ()
{ {
computePressure<<< fcuda.numBlocks, fcuda.numThreads>>> ( bufPntSort, bufGrid, (uint2*) bufHash[0], fcuda.pnts ); computePressure<<< fcuda.numBlocks, fcuda.numThreads>>> ( bufPntSort, bufGrid, (uint2*) bufHash[0], fcuda.pnts );
CUT_CHECK_ERROR( "Kernel execution failed"); BT_GPU_CHECK_ERROR( "Kernel execution failed");
cudaThreadSynchronize (); cudaThreadSynchronize ();
} }
@@ -220,14 +223,14 @@ void SPH_ComputeForceCUDA ()
// Force using neighbor table // Force using neighbor table
computeForceNbr<<< fcuda.numBlocks, fcuda.numThreads>>> ( bufPntSort, fcuda.pnts ); computeForceNbr<<< fcuda.numBlocks, fcuda.numThreads>>> ( bufPntSort, fcuda.pnts );
CUT_CHECK_ERROR( "Kernel execution failed"); BT_GPU_CHECK_ERROR( "Kernel execution failed");
cudaThreadSynchronize (); cudaThreadSynchronize ();
} }
void SPH_AdvanceCUDA ( float dt, float ss ) void SPH_AdvanceCUDA ( float dt, float ss )
{ {
advanceParticles<<< fcuda.numBlocks, fcuda.numThreads>>> ( bufPntSort, fcuda.pnts, dt, ss ); advanceParticles<<< fcuda.numBlocks, fcuda.numThreads>>> ( bufPntSort, fcuda.pnts, dt, ss );
CUT_CHECK_ERROR( "Kernel execution failed"); BT_GPU_CHECK_ERROR( "Kernel execution failed");
cudaThreadSynchronize (); cudaThreadSynchronize ();
} }
Extras/sph/fluids_2005.vcproj
+4 -4
View File
@@ -40,7 +40,7 @@
<Tool <Tool
Name="VCCLCompilerTool" Name="VCCLCompilerTool"
Optimization="0" Optimization="0"
AdditionalIncludeDirectories="fluids;common;marching;marching_tris;C:\CUDA\common\inc" AdditionalIncludeDirectories="fluids;common;marching;marching_tris;&quot;$(CUDA_INC_PATH)&quot;"
PreprocessorDefinitions="WIN32;_DEBUG;_CONSOLE" PreprocessorDefinitions="WIN32;_DEBUG;_CONSOLE"
MinimalRebuild="true" MinimalRebuild="true"
BasicRuntimeChecks="3" BasicRuntimeChecks="3"
@@ -61,7 +61,7 @@
/> />
<Tool <Tool
Name="VCLinkerTool" Name="VCLinkerTool"
AdditionalDependencies="glee_2005.lib cudart.lib cutil32D.lib" AdditionalDependencies="cudart.lib"
LinkIncremental="2" LinkIncremental="2"
IgnoreDefaultLibraryNames="libcmt.lib" IgnoreDefaultLibraryNames="libcmt.lib"
GenerateDebugInformation="true" GenerateDebugInformation="true"
@@ -137,7 +137,7 @@
/> />
<Tool <Tool
Name="VCLinkerTool" Name="VCLinkerTool"
AdditionalDependencies="cudart.lib cutil32.lib" AdditionalDependencies="cudart.lib "
LinkIncremental="1" LinkIncremental="1"
AdditionalLibraryDirectories="C:\CUDA\common\lib;." AdditionalLibraryDirectories="C:\CUDA\common\lib;."
IgnoreDefaultLibraryNames="libcmtd.lib" IgnoreDefaultLibraryNames="libcmtd.lib"
@@ -215,7 +215,7 @@
/> />
<Tool <Tool
Name="VCLinkerTool" Name="VCLinkerTool"
AdditionalDependencies="glee_VS2005.lib cg.lib cgGL.lib cudart.lib cutil32D.lib" AdditionalDependencies="cg.lib cgGL.lib cudart.lib "
LinkIncremental="1" LinkIncremental="1"
GenerateDebugInformation="true" GenerateDebugInformation="true"
ProgramDatabaseFile="$(OutDir)/fluids_2005.pdb" ProgramDatabaseFile="$(OutDir)/fluids_2005.pdb"
Extras/sph/main.cpp
+9
View File
@@ -583,3 +583,12 @@ int main ( int argc, char **argv )
return 0; return 0;
} }
extern "C" {
void btCuda_exit(int val)
{
fprintf(stderr, "Press ENTER key to terminate the program\n");
getchar();
exit(val);
}
}