diff --git a/RigidBodyController.h b/RigidBodyController.h
index 90a76b54f2e2ec395404aa6794d64b1f361f0243..41cbcbf524e06bc6e6234dce35a01b6956aaf88d 100644
--- a/RigidBodyController.h
+++ b/RigidBodyController.h
@@ -5,8 +5,8 @@
#include <cuda.h>
#include <cuda_runtime.h>
-// #define NUMTHREADS 256 /* try with 64, every 32+ */
-#define NUMTHREADS 64
+#define NUMTHREADS 128 /* try with 64, every 32+ */
+// #define NUMTHREADS 64
#define NUMSTREAMS 8
class Configuration;
diff --git a/RigidBodyGrid.cu b/RigidBodyGrid.cu
index 3721144099a629fcef0210ed3d2ac6b22f1cea17..2a793feb790b8e2876510e5dec292bc3f1801060 100644
--- a/RigidBodyGrid.cu
+++ b/RigidBodyGrid.cu
@@ -370,18 +370,20 @@ DEVICE Vector3 RigidBodyGrid::interpolateForceD(const Vector3 l) const {
const int homeX = int(floor(l.x));
const int homeY = int(floor(l.y));
const int homeZ = int(floor(l.z));
-
+ const float wx = l.x - homeX;
+ const float wy = l.y - homeY;
+ const float wz = l.z - homeZ;
{ /* f.x */
float g3[4];
// #pragma unroll {1,4} doesn't seem to matter
- for (int iz = 0; iz < 4; iz++) {
+ for (int iz = -1; iz < 3; iz++) {
float g2[4];
- for (int iy = 0; iy < 4; iy++) {
+ for (int iy = -1; iy < 3; iy++) {
float g1[4];
- for (volatile int ix = 0; ix < 4; ix++) { /* RBTODO: unsigned/short? Best practices guide says no */
- volatile int jx = (ix-1 + homeX);
- volatile int jy = (iy-1 + homeY);
- volatile int jz = (iz-1 + homeZ);
+ for (volatile int ix = -1; ix < 3; ix++) { /* RBTODO: unsigned/short? Best practices guide says no */
+ volatile int jx = (ix + homeX);
+ volatile int jy = (iy + homeY);
+ volatile int jz = (iz + homeZ);
const int ind = jz + jy*nz + jx*nz*ny;
g1[ix] = jz < 0 || jz >= nz || jy < 0 || jy >= ny || jx < 0 || jx >= nx ?
0 : val[ind];
@@ -391,7 +393,6 @@ DEVICE Vector3 RigidBodyGrid::interpolateForceD(const Vector3 l) const {
const float a2 = 0.5f*(2.0f*g1[0] - 5.0f*g1[1] + 4.0f*g1[2] - g1[3]);
const float a1 = 0.5f*(-g1[0] + g1[2]);
/* const float a0 = g1[1]; */
- const float wx = l.x - homeX;
// g2[iy] = a3*wx*wx*wx + a2*wx*wx + a1*wx + a0;
g2[iy] = 3.0f*a3*wx*wx + 2.0f*a2*wx + a1; /* derivative */
}
@@ -401,7 +402,6 @@ DEVICE Vector3 RigidBodyGrid::interpolateForceD(const Vector3 l) const {
const float a2 = 0.5f*(2.0f*g2[0] - 5.0f*g2[1] + 4.0f*g2[2] - g2[3]);
const float a1 = 0.5f*(-g2[0] + g2[2]);
const float a0 = g2[1];
- const float wy = l.y - homeY;
g3[iz] = a3*wy*wy*wy + a2*wy*wy + a1*wy + a0;
}
@@ -410,21 +410,20 @@ DEVICE Vector3 RigidBodyGrid::interpolateForceD(const Vector3 l) const {
const float a2 = 0.5f*(2.0f*g3[0] - 5.0f*g3[1] + 4.0f*g3[2] - g3[3]);
const float a1 = 0.5f*(-g3[0] + g3[2]);
const float a0 = g3[1];
- const float wz = l.z - homeZ;
f.x = -(a3*wz*wz*wz + a2*wz*wz + a1*wz + a0);
- }
+ }
{ /* f.y */
float g3[4];
- for (short iz = 0; iz < 4; iz++) {
+ for (int iz = -1; iz < 3; iz++) {
float g2[4];
- for (short iy = 0; iy < 4; iy++) {
+ for (int iy = -1; iy < 3; iy++) {
// Fetch values from nearby
float g1[4];
- for (volatile unsigned short ix = 0; ix < 4; ix++) {
- volatile short jx = (ix-1 + homeX);
- volatile short jy = (iy-1 + homeY);
- volatile short jz = (iz-1 + homeZ);
+ for (volatile int ix = -1; ix < 3; ix++) {
+ volatile int jx = (ix + homeX);
+ volatile int jy = (iy + homeY);
+ volatile int jz = (iz + homeZ);
const int ind = jz + jy*nz + jx*nz*ny;
g1[ix] = jz < 0 || jz >= nz || jy < 0 || jy >= ny || jx < 0 || jx >= nx ?
0 : val[ind];
@@ -434,7 +433,6 @@ DEVICE Vector3 RigidBodyGrid::interpolateForceD(const Vector3 l) const {
const float a2 = 0.5f*(2.0f*g1[0] - 5.0f*g1[1] + 4.0f*g1[2] - g1[3]);
const float a1 = 0.5f*(-g1[0] + g1[2]);
const float a0 = g1[1];
- const float wx = l.x - homeX;
g2[iy] = a3*wx*wx*wx + a2*wx*wx + a1*wx + a0;
}
@@ -443,7 +441,6 @@ DEVICE Vector3 RigidBodyGrid::interpolateForceD(const Vector3 l) const {
const float a2 = 0.5f*(2.0f*g2[0] - 5.0f*g2[1] + 4.0f*g2[2] - g2[3]);
const float a1 = 0.5f*(-g2[0] + g2[2]);
/* const float a0 = g2[1]; */
- const float wy = l.y - homeY;
// g3[iz] = a3*wy*wy*wy + a2*wy*wy + a1*wy + a0;
g3[iz] = 3.0f*a3*wy*wy + 2.0f*a2*wy + a1; /* derivative */
@@ -453,21 +450,19 @@ DEVICE Vector3 RigidBodyGrid::interpolateForceD(const Vector3 l) const {
const float a2 = 0.5f*(2.0f*g3[0] - 5.0f*g3[1] + 4.0f*g3[2] - g3[3]);
const float a1 = 0.5f*(-g3[0] + g3[2]);
const float a0 = g3[1];
- const float wz = l.z - homeZ;
-
f.y = -(a3*wz*wz*wz + a2*wz*wz + a1*wz + a0);
}
{ /* f.z */
float g3[4];
- for (short iz = 0; iz < 4; iz++) {
+ for (int iz = -1; iz < 3; iz++) {
float g2[4];
- for (short iy = 0; iy < 4; iy++) {
+ for (int iy = -1; iy < 3; iy++) {
// Fetch values from nearby
float g1[4];
- for (volatile unsigned short ix = 0; ix < 4; ix++) {
- volatile short jx = (ix-1 + homeX);
- volatile short jy = (iy-1 + homeY);
- volatile short jz = (iz-1 + homeZ);
+ for (volatile int ix = -1; ix < 3; ix++) {
+ volatile int jx = (ix + homeX);
+ volatile int jy = (iy + homeY);
+ volatile int jz = (iz + homeZ);
const int ind = jz + jy*nz + jx*nz*ny;
g1[ix] = jz < 0 || jz >= nz || jy < 0 || jy >= ny || jx < 0 || jx >= nx ?
0 : val[ind];
@@ -477,7 +472,6 @@ DEVICE Vector3 RigidBodyGrid::interpolateForceD(const Vector3 l) const {
const float a2 = 0.5f*(2.0f*g1[0] - 5.0f*g1[1] + 4.0f*g1[2] - g1[3]);
const float a1 = 0.5f*(-g1[0] + g1[2]);
const float a0 = g1[1];
- const float wx = l.x - homeX;
g2[iy] = a3*wx*wx*wx + a2*wx*wx + a1*wx + a0;
}
@@ -486,17 +480,13 @@ DEVICE Vector3 RigidBodyGrid::interpolateForceD(const Vector3 l) const {
const float a2 = 0.5f*(2.0f*g2[0] - 5.0f*g2[1] + 4.0f*g2[2] - g2[3]);
const float a1 = 0.5f*(-g2[0] + g2[2]);
const float a0 = g2[1];
- const float wy = l.y - homeY;
-
g3[iz] = a3*wy*wy*wy + a2*wy*wy + a1*wy + a0;
}
// Mix along z.
const float a3 = 0.5f*(-g3[0] + 3.0f*g3[1] - 3.0f*g3[2] + g3[3]);
const float a2 = 0.5f*(2.0f*g3[0] - 5.0f*g3[1] + 4.0f*g3[2] - g3[3]);
const float a1 = 0.5f*(-g3[0] + g3[2]);
- /* const float a0 = g3[1]; */
- const float wz = l.z - homeZ;
-
+ /* const float a0 = g3[1]; */
f.z = -(3.0f*a3*wz*wz + 2.0f*a2*wz + a1); /* derivative */
}
return f;
diff --git a/makefile b/makefile
index d7c89a1d3794ca395130588ca7265917b820d012..a7de250d1b0280edfcadf243c2f76c0bb5bbaf75 100644
--- a/makefile
+++ b/makefile
@@ -14,7 +14,8 @@ INCLUDE = $(CUDA_PATH)/include
# DEBUG = -g -O0
CC_FLAGS = -Wall -Wno-write-strings -I$(INCLUDE) $(DEBUG) -std=c++0x -pedantic
# NV_FLAGS = -g -G #debug
-NV_FLAGS = --maxrregcount 63 -Xptxas -v # -v,-abi=no
+# NV_FLAGS = --maxrregcount 63 -Xptxas -v # -v,-abi=no
+NV_FLAGS = -Xptxas -v # -v,-abi=no
EX_FLAGS = -O3 -m$(OS_SIZE)
# EX_FLAGS = -O0 -m$(OS_SIZE)