Solution Manual for Programming Massively Parallel Processors

treatments. This principle is enshrined in the legal concept of informed
consent. Informed consent requires healthcare providers to disclose all
relevant information about the risks, benefits, and alternatives to a
medical treatment or procedure, allowing patients to make informed
decisions.However, challenges arise when patients are not fully capable o
making informed decisions (e.g., due to age, mental illness, or language
barriers). In such cases, ethical dilemmas can arise regarding whether a
third party (e.g., a parent or guardian) should make the decision on the
patient’s behalf, and whether the legal framework supports such
decisions.#### 2.2 **End-of-Life Decisions and Euthanasia**End-of-life
care, particularly decisions regarding euthanasia, brings about
significant ethical and legal debates. While some argue that
Chapter Exercises Solution Manual
Programming Massively Parallel Processors: A Hands-on Approach
By Wen-mei W. Hwu, David B. Kirk, and Izzat El Hajj


Chapter 1: Introduction
No exercises.

Chapter 2: Heterogeneous data parallel computing
1. (C) i=blockIdx.x*blockDim.x + threadIdx.x;
2. (C) i=(blockIdx.x*blockDim.x + threadIdx.x)*2;
3. (D) i=blockIdx.x*blockDim.x*2 + threadIdx.x;
4. (C) 8192
5. (D) v * sizeof(int)
6. (D) (void **) &A_d
7. (C) cudaMemcpy(A_d, A_h, 3000, cudaMemcpyHostToDevice);
8. (C) cudaError_t err;
9.
a. 128
b. 200,064
c. 1,563
d. 200,064
e. 200,000
10. The intern should use both the “ host ” keyword and the “ device ” keyword in declaring the
functions that need to be executed on both the host and the device. The CUDA compiler will generate
both versions of the functions.

Chapter 3: Multidimensional grids and data
1. Not included
2.
global void matrixVectorMulKernel(float *A, float *B, float *C, int vectorLen) {
int i = threadIdx.x + blockIdx.x*blockDim.x;
float sum = 0.0f;
if (i < vectorLen) {
for (int j = 0; j < vectorLen; j++)
sum += B[i*vectorLen + j] * C[j];
A[i] = sum;
}
}

void matrixVectorMul(float *h_A, float *h_B, float *h_C, int vectorLen) {
float *d_A, *d_B, *d_C;
int matrixSize = vectorLen*vectorLen*sizeof(float);
int vectorSize = vectorLen*sizeof(float);
cudaMalloc((void**)&d_A, vectorSize);
cudaMalloc((void**)&d_B, matrixSize);
cudaMalloc((void**)&d_C, vectorSize);