Parallel Computing (Quiz-2)

What does CUDA stand for?

• Compute Unified Device Architecture
• Central Unified Device Architecture
• Compute Unit Device Application
• Complex Unified Device Array

Answer explanation

Explanation for the answer (optional)

What type of processing is CUDA primarily used for?

• Serial processing
• Parallel processing
• Quantum processing
• Cloud computing

What is a CUDA kernel?

• A CPU function executed on a single core
• A GPU function that executes in parallel across multiple threads
• A function for memory allocation
• A function for managing CPU-GPU communication

Which keyword is used to declare a CUDA kernel function?

• __kernel__
• __global__
• __device__
• __shared__

Which CUDA function is used to allocate memory on the GPU?

• cudaMemcpy
• cudaMalloc
• cudaInit
• cudaFree

In CUDA, what are threads grouped into?

• Blocks
• Kernels
• Grids
• Streams

What is a grid in CUDA?

• A collection of blocks
• A single thread
• A shared memory space
• A global memory array

Which built-in variable gives the thread's unique ID within a block?

• blockIdx
• threadIdx
• gridDim
• blockDim

How is the total number of threads calculated in a CUDA kernel?

• blockIdx.x * blockDim.x + threadIdx.x
• gridDim.x * blockDim.x + threadIdx.x
• (gridDim.x * blockDim.x) + threadIdx.x
• (blockIdx.x * blockDim.x) + threadIdx.x

What is the maximum number of threads per block in CUDA (as of recent architectures)?

• 256
• 512
• 1024
• 2048