Deep learning neural network training requires a lot of matrix calculations. Generally, we need to use hardware to enable the computer to have parallel computing capabilities. The following hardware…

The correct answer is C. GPU. GPUs (Graphics Processing Units) are purpose-built for massive parallelism - they contain thousands of smaller cores designed to perform many arithmetic operations simultaneously, which maps perfectly onto the matrix multiplications that dominate neural network training. A CPU…

Deep Learning Basics

Question

Deep learning neural network training requires a lot of matrix calculations. Generally, we need to use hardware to enable the computer to have parallel computing capabilities. The following hardware devices can provide parallel computing capabilities.:

Options

AMotherboard
BRAM
CGPU
DCPU

How the community answered

(55 responses)

A
16% (9)
B
4% (2)
C
73% (40)
D
7% (4)

Explanation

GPUs (Graphics Processing Units) are purpose-built for massive parallelism - they contain thousands of smaller cores designed to perform many arithmetic operations simultaneously, which maps perfectly onto the matrix multiplications that dominate neural network training. A CPU, while powerful, has only a handful of cores optimized for sequential, low-latency tasks, making it poorly suited for the thousands of concurrent operations required during forward/backward passes. RAM is simply memory storage - it holds data but performs no computation. The motherboard is the physical backbone connecting components but contributes no processing capability of its own.

Memory tip: Think of the "G" in GPU as standing for "Grid of cores" - a GPU's grid of thousands of cores lets it crunch entire matrices at once, while a CPU's few cores must work through them row by row.

Topics

#GPU#Parallel Computing#Deep Learning Hardware#Matrix Operations

Community Discussion

No community discussion yet for this question.

Full H13-311_V3.5 Practice