hw6 - CS 6143 COMPUTER ARCHITECTURE II HOMEWORK VI FALL...

Info iconThis preview shows pages 1–3. Sign up to view the full content.

View Full Document Right Arrow Icon
CS 6143 COMPUTER ARCHITECTURE II FALL 2010 HOMEWORK VI Polytechnic Institute of NYU Page 1 of 14 Handout No : 14 December 8, 2010 DUE : By 6PM, Monday, December 20, 2010 READ : i) Related portions of Chapters 2, 4 and Appendix H of the Hennessy book ii) Related portions of Chapters 2, 4, 5, 7, 8, 9 and 10 of the Jordan book ASSIGNMENT: There are four problems. Solve all homework and exam problems as shown in class and past exam solutions. 1) a) Describe the Illinois cache coherence protocol, including its state diagram, as done in class. b) Assume that a multiprocessor system consists of two processors on a single bus. Cache mem- ories use the Illinois cache coherence protocol. The processors have id numbers 0 and 1. Assume that a program is run on these two processors and they access main memory blocks “ k and “ m .” The two blocks map to the same area in the two cache memories. Processor 0 has already started the execution but not yet used blocks “k” and “m.” Processor 1 starts after proces- sor 0. Continue the following table that shows operations with respect to time : PO p B l o c k Hit/ Miss Cache 0 state changes Cache 1 state changes Any cache or memory action 0R k 1R k 0W k 1W k m m m m k k k k
Background image of page 1

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

View Full DocumentRight Arrow Icon
Polytechnic Institute of NYU Page 2 of 14 CS6143 Handout No : 14 December 8, 2010 2) Develop a dot-product algorithm on two vectors, A and B . Store the result in “ k ” for a UMA MIMD system with “ p ” processors. Í Indicate the time complexity of your algorithm. Í Make observations relevant to the execution of your UMA MIMD algorithm, including the data decomposition, load balancing, synchronization, etc. 3) Develop a matrix multiply algorithm on two square matrices, A and B . Store the result in “ C for a UMA MIMD system with “ p ” processors. Í Indicate the time complexity of your algorithm. Í Make observations relevant to the execution of your UMA MIMD algorithm, including the data decomposition, load balancing, synchronization, etc. 4) Develop a dot-product algorithm on two vectors, A and B. Store the result in “ k ” for a 2-d tor- oidal-mesh NORMA MIMD system with “ p ” processors. Í Indicate the time complexity of your algorithm. Í Make observations relevant to the execution of your NORMA MIMD algorithm, including the data decomposition, load balancing, the communication graph, etc. RELEVANT QUESTIONS AND ANSWERS Q1) A multiprocessor system consists of two processors on a single bus. The cache memories use the snoopy cache coherence protocol discussed in class (the protocol with three states). The processors have id numbers 0 and 1. Assume that a program will be run on the two processors and they will access main memory blocks “k” and “m.” These two blocks map to the same area in the two cache memories. Processor P0 has already started the execution but not yet used blocks “k” and “m.” The other processor, P1, starts after processor P0. Continue on the following table that shows operations with respect to time : PO p B l o c k H / M Cache 0 state changes Cache 1 state changes
Background image of page 2
Image of page 3
This is the end of the preview. Sign up to access the rest of the document.

This note was uploaded on 02/02/2011 for the course CS 6143 taught by Professor Hadimioglu during the Fall '10 term at NYU Poly.

Page1 / 14

hw6 - CS 6143 COMPUTER ARCHITECTURE II HOMEWORK VI FALL...

This preview shows document pages 1 - 3. Sign up to view the full document.

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