diff options
| -rw-r--r-- | .hgignore | 3 | ||||
| -rw-r--r-- | linalg.h | 101 | ||||
| -rw-r--r-- | linalg_test.cu | 99 |
3 files changed, 203 insertions, 0 deletions
diff --git a/.hgignore b/.hgignore new file mode 100644 index 0000000..ea5e656 --- /dev/null +++ b/.hgignore @@ -0,0 +1,3 @@ +syntax:glob +*.pyc +*~
\ No newline at end of file diff --git a/linalg.h b/linalg.h new file mode 100644 index 0000000..4d0344e --- /dev/null +++ b/linalg.h @@ -0,0 +1,101 @@ +#ifndef __LINALG_H__ +#define __LINALG_H__ + +__device__ __host__ float3 operator+ (const float3 &a, const float3 &b) +{ + return make_float3(a.x+b.x, a.y+b.y, a.z+b.y); +} + +__device__ __host__ void operator+= (float3 &a, const float3 &b) +{ + a.x += b.x; + a.y += b.y; + a.z += b.z; +} + +__device__ __host__ float3 operator- (const float3 &a, const float3 &b) +{ + return make_float3(a.x-b.x, a.y-b.y, a.z-b.z); +} + +__device__ __host__ void operator-= (float3 &a, const float3 &b) +{ + a.x -= b.x; + a.y -= b.y; + a.z -= b.z; +} + +__device__ __host__ float3 operator+ (const float3 &a, const float &c) +{ + return make_float3(a.x+c, a.y+c, a.z+c); +} + +__device__ __host__ float3 operator+= (const float3 &a, const float &c) +{ + a.x += c; + a.y += c; + a.z += c; +} + +__device__ __host__ float3 operator+ (const float &c, const float3 &a) +{ + return make_float3(c+a.x, c+a.y, c+a.z); +} + +__device__ __host__ float3 operator- (const float3 &a, const float &c) +{ + return make_float3(a.x-c, a.y-c, a.z-c); +} + +__device__ __host__ float3 operator-= (const float3 &a, const float &c) +{ + a.x -= c; + a.y -= c; + a.z -= c; +} + +__device__ __host__ float3 operator- (const float &c, const float3& a) +{ + return make_float3(c-a.x, c-a.y, c-a.z); +} + +__device__ __host__ float3 operator* (const float3 &a, const float &c) +{ + return make_float3(a.x*c, a.y*c, a.z*c); +} + +__device__ __host__ float3 operator*= (const float3 &a, const float &c) +{ + a.x *= c; + a.y *= c; + a.z *= c; +} + +__device__ __host__ float3 operator* (const float &c, const float3& a) +{ + return make_float3(c*a.x, c*a.y, c*a.z); +} + +__device__ __host__ float3 operator/ (const float3 &a, const float &c) +{ + return make_float3(a.x/c, a.y/c, a.z/c); +} + +__device__ __host__ float3 operator/= (const float3 &a, const float &c) +{ + a.x /= c; + a.y /= c; + a.z /= c; +} + +__device__ __host__ float3 operator/ (const float &c, const float3 &a) +{ + return make_float3(c/a.x, c/a.y, c/a.z); +} + +__device__ __host__ float dot(const float3 &a, const float3 &b) +{ + return a.x*b.x + a.y*b.y + a.z*b.z; +} + +#endif diff --git a/linalg_test.cu b/linalg_test.cu new file mode 100644 index 0000000..c4647a7 --- /dev/null +++ b/linalg_test.cu @@ -0,0 +1,99 @@ +//-*-c-*- + +#include "linalg.h" + +extern "C" +{ + +__global__ void add(int n, float3 *a, float3 *b, float3 *out) +{ + int i; + for (i=0; i < n; i++) + out[i] = a[i] + b[i]; +} + +__global__ void addequal(int n, float3 *a, float3 *b) +{ + int i; + for (i=0; i < n; i++) + a[i] += b[i]; +} + +__global__ void sub(int n, float3 *a, float3 *b, float3 *out) +{ + int i; + for (i=0; i < n; i++) + out[i] = a[i] - b[i]; +} + +__global__ void subequal(int n, float3 *a, float3 *b) +{ + int i; + for (i=0; i < n; i++) + a[i] -= b[i]; +} + +__gloabl__ void addfloat(int n, float3 *a, float c, float3 *out) +{ + int i; + for (i=0; i < n; i++) + out[i] = a[i] + c; +} + +__global__ void addfloatequal(int n, float3 *a, float c) +{ + int i; + for (i=0; i < n; i++) + a += c; +} + +__global__ void subfloat(int n, float3 *a, float c, float3 *out) +{ + int i; + for (i=0; i < n; i++) + out[i] = a[i] - c; +} + +__global__ void subfloatequal(int n, float3 *a, float c) +{ + int i; + for (i=0; i < n; i++) + a[i] -= c; +} + +__global__ void mulfloat(int n, float3 *a, float c, float3 *out) +{ + int i; + for (i=0; i < n; i++) + out[i] = a[i]*c; +} + +__global__ void mulfloatequal(int n, float3 *a, float c) +{ + int i; + for (i=0; i < n; i++) + a *= c; +} + +__global__ void divfloat(int n, float3 *a, float c, float3 *out) +{ + int i; + for (i=0; i < n; i++) + out[i] = a[i]/c; +} + +__global__ void divfloatequal(int n, float3 *a, float c) +{ + int i; + for (i=0; i < n; i++) + a /= c; +} + +__global__ void dot(int n, float3 *a, float3 *b, float* out) +{ + int i; + for (i=0; i < n; i++) + out[i] = dot(a,b); +} + +} // extern "c" |