Dr. Zornitza Genova Prodanoff

EVALUATION TOOLS:

CRC-related tools

gencrc.c generates CRC32s for a list of strings

clipcrc.c clips CRC32s to M bits where the input is list of CRC32s

Hashing

A program to implement simple chained hashing is hash1.c. Input is a list of 32-bit values with associated 32-bit values, output is storage statistics for the hash table of 2^M bits (where M is a settable parameter).

hash1a.c hashes 32-bit key values into a chained hash table of 2^M bits size and adds performance measurement for look-ups.

hash1b.c hashes 32-bit key values into a chained hash table of 2^M bits size and puts access list into memory before timing accesses.

hash2.c hashes variable length strings into a hash table of 2^M bits size. Input is a list of URL strings and 32-bit signatures with associated 32-bit values, output is storage statistics for the hash table of 2^M bits.

hash2a.c hashes variable length strings into a hash table of 2^M bits size and adds performance measurement for look-ups.

hash2b.c hashes variable length strings into a hash table of 2^M bits size and puts access list into memory before timing accesses.

2level_h.c - constant time look-up implementation as described in "Storing a Sparse Table with O(1) worst case access time." by Michael Freedman and Janos Komlos, Journal of the ACM, Vol 31, No 3, July 1984, pp. 538-544.

Memory access related programs

alloc1.c times malloc and index accesses alloc2.c times sequential index accesses alloc3.c times sequential and random writes alloc4.c times sequential and random reads

Web programming

rawstuff.zip demonstrates raw sends and receives web_router.zip a simple web router that redirects a single http request

Miscellaneous

dupstr.c finds duplicate strings in a sorted input list birthday.c a solution to the birthday problem

TRACE FILES:

csee_urlsu.txt - set of URLs hosted by CSEE, and extracted from a two hour server access log

pb.sanitized-access.20010110.gz [ 22.8MB ]- Applied Networks Research cache log, format is described at ftp link

pb20010110url [ 38.2MB ASCII ] extracted with per00.pl URL access list

pb20010110su [ 24.6MB ASCII ] cleaned list of unique URLs

University of Virginia - server log

NASA - server log

Calgary - cache log

OTHER RESOURCES:

WinXs32 - UNIX-like tools for Windows implementing 32-bit versions of UNIX tools including cut, diff, head, ls, sort, uniq, and many others (shareware $30 per end user license).