From 142b3c3caff164deb9bc7b2848e58e52387723ff Mon Sep 17 00:00:00 2001 From: Stan Seibert Date: Fri, 16 Sep 2011 15:02:02 -0400 Subject: Move CUDA source inside chroma package, rename tests directory to test --- test/linalg_test.cu | 128 ++++++++++++++++++++++++++++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 test/linalg_test.cu (limited to 'test/linalg_test.cu') diff --git a/test/linalg_test.cu b/test/linalg_test.cu new file mode 100644 index 0000000..4e9c983 --- /dev/null +++ b/test/linalg_test.cu @@ -0,0 +1,128 @@ +//-*-c-*- + +#include "linalg.h" + +extern "C" +{ + +__global__ void float3add(float3 *a, float3 *b, float3 *dest) +{ + int idx = blockIdx.x*blockDim.x + threadIdx.x; + dest[idx] = a[idx] + b[idx]; +} + +__global__ void float3addequal(float3 *a, float3 *b) +{ + int idx = blockIdx.x*blockDim.x + threadIdx.x; + a[idx] += b[idx]; +} + +__global__ void float3sub(float3 *a, float3 *b, float3 *dest) +{ + int idx = blockIdx.x*blockDim.x + threadIdx.x; + dest[idx] = a[idx] - b[idx]; +} + +__global__ void float3subequal(float3 *a, float3 *b) +{ + int idx = blockIdx.x*blockDim.x + threadIdx.x; + a[idx] -= b[idx]; +} + +__global__ void float3addfloat(float3 *a, float c, float3 *dest) +{ + int idx = blockIdx.x*blockDim.x + threadIdx.x; + dest[idx] = a[idx] + c; +} + +__global__ void float3addfloatequal(float3 *a, float c) +{ + int idx = blockIdx.x*blockDim.x + threadIdx.x; + a[idx] += c; +} + +__global__ void floataddfloat3(float3 *a, float c, float3 *dest) +{ + int idx = blockIdx.x*blockDim.x + threadIdx.x; + dest[idx] = c + a[idx]; +} + +__global__ void float3subfloat(float3 *a, float c, float3 *dest) +{ + int idx = blockIdx.x*blockDim.x + threadIdx.x; + dest[idx] = a[idx] - c; +} + +__global__ void float3subfloatequal(float3 *a, float c) +{ + int idx = blockIdx.x*blockDim.x + threadIdx.x; + a[idx] -= c; +} + +__global__ void floatsubfloat3(float3 *a, float c, float3 *dest) +{ + int idx = blockIdx.x*blockDim.x + threadIdx.x; + dest[idx] = c - a[idx]; +} + +__global__ void float3mulfloat(float3 *a, float c, float3 *dest) +{ + int idx = blockIdx.x*blockDim.x + threadIdx.x; + dest[idx] = a[idx]*c; +} + +__global__ void float3mulfloatequal(float3 *a, float c) +{ + int idx = blockIdx.x*blockDim.x + threadIdx.x; + a[idx] *= c; +} + +__global__ void floatmulfloat3(float3 *a, float c, float3 *dest) +{ + int idx = blockIdx.x*blockDim.x + threadIdx.x; + dest[idx] = c*a[idx]; +} + +__global__ void float3divfloat(float3 *a, float c, float3 *dest) +{ + int idx = blockIdx.x*blockDim.x + threadIdx.x; + dest[idx] = a[idx]/c; +} + +__global__ void float3divfloatequal(float3 *a, float c) +{ + int idx = blockIdx.x*blockDim.x + threadIdx.x; + a[idx] /= c; +} + +__global__ void floatdivfloat3(float3 *a, float c, float3 *dest) +{ + int idx = blockIdx.x*blockDim.x + threadIdx.x; + dest[idx] = c/a[idx]; +} + +__global__ void dot(float3 *a, float3 *b, float *dest) +{ + int idx = blockIdx.x*blockDim.x + threadIdx.x; + dest[idx] = dot(a[idx],b[idx]); +} + +__global__ void cross(float3 *a, float3 *b, float3 *dest) +{ + int idx = blockIdx.x*blockDim.x + threadIdx.x; + dest[idx] = cross(a[idx],b[idx]); +} + +__global__ void norm(float3 *a, float *dest) +{ + int idx = blockIdx.x*blockDim.x + threadIdx.x; + dest[idx] = norm(a[idx]); +} + +__global__ void minusfloat3(float3 *a, float3 *dest) +{ + int idx = blockIdx.x*blockDim.x + threadIdx.x; + dest[idx] = -a[idx]; +} + +} // extern "c" -- cgit