Simple vector addition n Cuda

Accelerated Computing CUDA CUDA Programming and Performance
#1

I searched but I didnt found anything…

I wonder how carry is handled in a simple vector addition (i.e. the one in sdk example)

__global__ void add_vectors_kernel(float *A, float *B, float *C, int N) {

	// Determine which element this thread is computing

	int block_id = blockIdx.x + gridDim.x * blockIdx.y;

	int thread_id = blockDim.x * block_id + threadIdx.x;

	

	// Compute a single element of the result vector (if the element is valid)

	if (thread_id < N) C[thread_id] = A[thread_id] + B[thread_id];

}

is it because they are float?

I need for a unsigned char byte array…

#2

I am sorry, i don’t understand what you are trying to say. Is it that you need the vector addition for unsigned chars?

#3

If you want to add unsigned char array,

__global__ void add_vectors_kernel(unsigned char *A, unsigned char *B, unsigned char *C, int N)

{

    // Determine which element this thread is computing

    int block_id = blockIdx.x + gridDim.x * blockIdx.y;

    int thread_id = blockDim.x * block_id + threadIdx.x;

    // Compute a single element of the result vector (if the element is valid)

    if (thread_id < N){

       C[thread_id] = A[thread_id] + B[thread_id];

    }

}

then compiler would load A[thread_id], B[thread_id] to 32-bit register and

do addition of 32-bit register, finally write LSB 8-bit of 32-bit register to C[thread_id].

you can try following code to compute carry bit

__global__ void add_vectors_kernel(unsigned char *A, unsigned char *B, unsigned char *C, int N)

{

    // Determine which element this thread is computing

    int block_id = blockIdx.x + gridDim.x * blockIdx.y;

    int thread_id = blockDim.x * block_id + threadIdx.x;

    // Compute a single element of the result vector (if the element is valid)

    if (thread_id < N){

       int Areg =  A[thread_id] ;

       int Breg =  B[thread_id] ;

       int Creg = Areg + Breg ;

       int carry = (Creg > 255) ;

       C[thread_id] = Creg & 255 ;

    }

}
#4

@tdhd: I apologize if I didnt write properly what I meant

@LSChien: thanks man, that’s what I was looking for =)

#5

…and what about if I want subtract?

i.e.: c − b = a

Conditions are the following:

  • I am working on big numbers (many limbs)

  • Always byte array (unsigned char)

  • The total minuend ‘c’ is always bigger than the total subtrahend ‘b’, but obviously is possible that the singular limb ‘c’ is smaller than the singular limb ‘b’.