Cleanup some cuda compile warnings.

cuda/spgpu/kernels/caxpby.cu

@@ -63,7 +63,6 @@ void spgpuCaxpby(spgpuHandle_t handle,
 	cuFloatComplex alpha,
 	__device cuFloatComplex* x)
 {
-	int msize = (n+BLOCK_SIZE-1)/BLOCK_SIZE;
 	int num_mp, max_threads_mp, num_blocks_mp, num_blocks;
 	dim3 block(BLOCK_SIZE);
 	num_mp         = getGPUMultiProcessors();

cuda/spgpu/kernels/cupd_xyz.cu

@@ -65,7 +65,6 @@ void spgpuCupd_xyz(spgpuHandle_t handle,
 		   __device cuFloatComplex * y,
 		   __device cuFloatComplex  *z)
 {
-	int msize = (n+BLOCK_SIZE-1)/BLOCK_SIZE;
 	int num_mp, max_threads_mp, num_blocks_mp, num_blocks;
 	dim3 block(BLOCK_SIZE);
 	num_mp         = getGPUMultiProcessors();

cuda/spgpu/kernels/cxyzw.cu

@@ -63,7 +63,6 @@ void spgpuCxyzw(spgpuHandle_t handle,
 		__device cuFloatComplex * z,
 		__device cuFloatComplex *w)
 {
-	int msize = (n+BLOCK_SIZE-1)/BLOCK_SIZE;
 	int num_mp, max_threads_mp, num_blocks_mp, num_blocks;
 	dim3 block(BLOCK_SIZE);
 	num_mp         = getGPUMultiProcessors();

cuda/spgpu/kernels/daxpby.cu

@@ -60,7 +60,6 @@ void spgpuDaxpby(spgpuHandle_t handle,
 	double alpha,
 	__device double* x)
 {
-	int msize = (n+BLOCK_SIZE-1)/BLOCK_SIZE;
 	int num_mp, max_threads_mp, num_blocks_mp, num_blocks;
 	dim3 block(BLOCK_SIZE);
 	num_mp         = getGPUMultiProcessors();

cuda/spgpu/kernels/dupd_xyz.cu

@@ -64,7 +64,6 @@ void spgpuDupd_xyz(spgpuHandle_t handle,
 		   __device double* y,
 		   __device double *z)
 {
-	int msize = (n+BLOCK_SIZE-1)/BLOCK_SIZE;
 	int num_mp, max_threads_mp, num_blocks_mp, num_blocks;
 	dim3 block(BLOCK_SIZE);
 	num_mp         = getGPUMultiProcessors();

cuda/spgpu/kernels/dxyzw.cu

@@ -63,7 +63,6 @@ void spgpuDxyzw(spgpuHandle_t handle,
 		__device double * z,
 		__device double *w)
 {
-	int msize = (n+BLOCK_SIZE-1)/BLOCK_SIZE;
 	int num_mp, max_threads_mp, num_blocks_mp, num_blocks;
 	dim3 block(BLOCK_SIZE);
 	num_mp         = getGPUMultiProcessors();

cuda/spgpu/kernels/saxpby.cu

@@ -62,7 +62,6 @@ void spgpuSaxpby(spgpuHandle_t handle,
 	float alpha,
 	__device float* x)
 {
-	int msize = (n+BLOCK_SIZE-1)/BLOCK_SIZE;
 	int num_mp, max_threads_mp, num_blocks_mp, num_blocks;
 	dim3 block(BLOCK_SIZE);
 	num_mp         = getGPUMultiProcessors();

cuda/spgpu/kernels/supd_xyz.cu

@@ -64,7 +64,6 @@ void spgpuSupd_xyz(spgpuHandle_t handle,
 		   __device float* y,
 		   __device float *z)
 {
-	int msize = (n+BLOCK_SIZE-1)/BLOCK_SIZE;
 	int num_mp, max_threads_mp, num_blocks_mp, num_blocks;
 	dim3 block(BLOCK_SIZE);
 	num_mp         = getGPUMultiProcessors();

cuda/spgpu/kernels/sxyzw.cu

@@ -63,7 +63,6 @@ void spgpuSxyzw(spgpuHandle_t handle,
 		__device float * z,
 		__device float *w)
 {
-	int msize = (n+BLOCK_SIZE-1)/BLOCK_SIZE;
 	int num_mp, max_threads_mp, num_blocks_mp, num_blocks;
 	dim3 block(BLOCK_SIZE);
 	num_mp         = getGPUMultiProcessors();

cuda/spgpu/kernels/zaxpby.cu

@@ -64,7 +64,6 @@ void spgpuZaxpby(spgpuHandle_t handle,
 	cuDoubleComplex alpha,
 	__device cuDoubleComplex* x)
 {
-	int msize = (n+BLOCK_SIZE-1)/BLOCK_SIZE;
 	int num_mp, max_threads_mp, num_blocks_mp, num_blocks;
 	dim3 block(BLOCK_SIZE);
 	num_mp         = getGPUMultiProcessors();

cuda/spgpu/kernels/zupd_xyz.cu

@@ -65,7 +65,6 @@ void spgpuZupd_xyz(spgpuHandle_t handle,
 		   __device cuDoubleComplex * y,
 		   __device cuDoubleComplex  *z)
 {
-	int msize = (n+BLOCK_SIZE-1)/BLOCK_SIZE;
 	int num_mp, max_threads_mp, num_blocks_mp, num_blocks;
 	dim3 block(BLOCK_SIZE);
 	num_mp         = getGPUMultiProcessors();

cuda/spgpu/kernels/zxyzw.cu

@@ -63,7 +63,6 @@ void spgpuZxyzw(spgpuHandle_t handle,
 		__device cuDoubleComplex * z,
 		__device cuDoubleComplex *w)
 {
-	int msize = (n+BLOCK_SIZE-1)/BLOCK_SIZE;
 	int num_mp, max_threads_mp, num_blocks_mp, num_blocks;
 	dim3 block(BLOCK_SIZE);
 	num_mp         = getGPUMultiProcessors();