Fall03-quiz2

# Fall03-quiz2 - University of California Berkeley College of...

• Homework Help
• davidvictor
• 15

This preview shows pages 1–5. Sign up to view the full content.

1 University of California, Berkeley College of Engineering Computer Science Division EECS Spring 2003 John Kubiatowicz Midterm II SOLUTIONS December 1, 2003 CS252 Graduate Computer Architecture Your Name: SID Number: Problem Possible Score 1 25 2 25 3 20 4 30 Total 100

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

3 Problem #1: Checking the bits The error correction coding process can be viewed as a transformation from one space of bits (the unencoded data) to another (the coded data). 1a) For linear block codes, there are two matrices, the Generator ( G ) and Parity Check ( H ). What are they, how big are they (in terms parameters n and k ), and how are they used? The Generator, G , is a k u n matrix used to produce code words from data words: Let d be a k -bit data vector and C be the resulting n- bit code word. Then C=G d. H is an n u (n-k) matrix used to check the parity of a code world and produce a syndrome indicating where errors are located. Let S be an ( n-k) -bit vector, then S=H C. 1b) Define the minimum Hamming distance, d min , of an error correction code (make sure that this definition makes sense for codes like Reed-Solomon of RAID-5 that have more than one bit per symbol): d min is the minimum number of positions (symbols) of difference between two valid code word. If the symbols are bits, this is the number of different bit positions, but can be number of different disk blocks for something like RAID-5. 1c) Name a constraint on the columns of the parity check matrix ( H ) which would indicate that a code had minimum distance d min . If a code with parity check matrix H has minimum distance d min , then every combination of d min -1 columns of H must be independent. This means that no combination of d min -1 columns can be combined to yield zero. Similarly, no combination of d min -2 columns, d min -3 columns, etc. This also means that every column is unique and non-zero. 1d) For a code with d min minimum distance, what is the formula for the maximum number of errors, E detect , that can be reliably detected? Why might you be able to detect more than the E detect errors? 1 ± PLQ GHWHFW G ( . You might be able to detect more if the code words are not evenly distributed. However, more simply, you can almost always detect a number of errors that is not evenly divisible by d min.

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

4 1e) For a code with d min minimum distance, what is the formula for the maximum number of random errors, E correct , that can be corrected reliably? » ¼ » « ¬ « ± 2 1 PLQ GHWHFW G ( 1f) How does your answer to (1e) change if you have knowledge of which symbols are bad (i.e. which symbols have been erased ). How might erased symbols be detected? If you know which symbols are bad, then you can often correct up to 1 ± PLQ G errors. Erased symbols might be detected by some other means such as checksums over individual symbols (especially makes sense for large symbols such as disk blocks) or even another intra-symbol-level error code.
This is the end of the preview. Sign up to access the rest of the document.
• Spring '07
• Kubiatowicz
• Computer Architecture, Central processing unit, CPU cache, Cache algorithms, Hamming Code, write-back cache

{[ snackBarMessage ]}

### What students are saying

• As a current student on this bumpy collegiate pathway, I stumbled upon Course Hero, where I can find study resources for nearly all my courses, get online help from tutors 24/7, and even share my old projects, papers, and lecture notes with other students.

Kiran Temple University Fox School of Business ‘17, Course Hero Intern

• I cannot even describe how much Course Hero helped me this summer. It’s truly become something I can always rely on and help me. In the end, I was not only able to survive summer classes, but I was able to thrive thanks to Course Hero.

Dana University of Pennsylvania ‘17, Course Hero Intern

• The ability to access any university’s resources through Course Hero proved invaluable in my case. I was behind on Tulane coursework and actually used UCLA’s materials to help me move forward and get everything together on time.

Jill Tulane University ‘16, Course Hero Intern