MT1_2010_sol

MT1_2010_sol - CSci 5451, S10 Midterm # 1 March 8th, 2010...

Info iconThis preview shows pages 1–2. Sign up to view the full content.

View Full Document Right Arrow Icon

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full DocumentRight Arrow Icon
This is the end of the preview. Sign up to access the rest of the document.

Unformatted text preview: CSci 5451, S10 Midterm # 1 March 8th, 2010 No documents are allowed - except for a 2-page formula sheet. Weights are indicated in bold at the end of each question. Base is: 100 1. General questions . Give a very brief answer for each of the following questions: (a) What is the main disadvantage of bus-based parallel architectures ? [4 pts] Answer: They do not scale well with number of processors. (b) Name two non-blocking switch-based networks. [4 pts] Answer: : Cross-bar and Benes switches. (c) What is the maximum block-size in CUDA? [4 pts] Answer: 512 (d) What does SIMD stand for? Give an example of SIMD-style programming. [4 pts] Answer: Single Instruction Multiple Stream. An example is CUDA (e) What is the effect of the openMP directive #pragma omp critical ? [4 pts] Answer: The section will be executed by one thread [to avoid racing] 2. Consider a 4-stage Omega network [16 inputs to 16 outputs]. Show how to route a mes- sage from input 0101 to ouput 1011. [List switches that are traversed at each stage and their state, e.g.: Stage 1: switch 0111, state == etc, Stage 2 switch 0110, state X, etc. . *No figures*] [20 pts] Answer: The input and output differ addresses in first, second, and third bits. Therefore the switches traversed in the 1st, 2nd and 3rd stages will be set to X and the last one to...
View Full Document

Page1 / 3

MT1_2010_sol - CSci 5451, S10 Midterm # 1 March 8th, 2010...

This preview shows document pages 1 - 2. Sign up to view the full document.

View Full Document Right Arrow Icon
Ask a homework question - tutors are online