| 1 |
#!/bin/sh |
| 2 |
|
| 3 |
BENCHMARK_OUT=benchmark_out |
| 4 |
HS=../heatshrink |
| 5 |
|
| 6 |
cd ${BENCHMARK_OUT} |
| 7 |
|
| 8 |
# Files in the Canterbury Corpus |
| 9 |
# http://corpus.canterbury.ac.nz/resources/cantrbry.tar.gz |
| 10 |
FILES='alice29.txt |
| 11 |
asyoulik.txt |
| 12 |
cp.html |
| 13 |
fields.c |
| 14 |
grammar.lsp |
| 15 |
kennedy.xls |
| 16 |
lcet10.txt |
| 17 |
plrabn12.txt |
| 18 |
ptt5 |
| 19 |
sum |
| 20 |
xargs.1' |
| 21 |
|
| 22 |
rm -f benchmark.output |
| 23 |
|
| 24 |
# Run several combinations of -w W -l L, |
| 25 |
# note compression ratios and check uncompressed output matches input |
| 26 |
for W in 6 7 8 9 10 11 12; do |
| 27 |
for L in 5 6 7 8; do |
| 28 |
if [ $L -lt $W ]; then |
| 29 |
for f in ${FILES} |
| 30 |
do |
| 31 |
IN_FILE="${f}" |
| 32 |
COMPRESSED_FILE="${f}.hsz.${W}_${L}" |
| 33 |
UNCOMPRESSED_FILE="${f}.orig.${W}_${L}" |
| 34 |
${HS} -e -v -w ${W} -l ${L} ${IN_FILE} ${COMPRESSED_FILE} >> benchmark.output |
| 35 |
${HS} -d -v -w ${W} -l ${L} ${COMPRESSED_FILE} ${UNCOMPRESSED_FILE} > /dev/null |
| 36 |
|
| 37 |
# compare file sizes |
| 38 |
if [ $(ls -l ${IN_FILE} | awk '{print($5)}') != $(ls -l ${UNCOMPRESSED_FILE} | awk '{print($5)}') ]; |
| 39 |
then |
| 40 |
printf "\n\n\nWARNING: size of %s does not match size of %s\n\n\n" \ |
| 41 |
${IN_FILE} ${UNCOMPRESSED_FILE} |
| 42 |
else |
| 43 |
printf "pass: -w %2d -l %2d %s\n" ${W} ${L} "${f}" |
| 44 |
fi |
| 45 |
|
| 46 |
rm ${COMPRESSED_FILE} ${UNCOMPRESSED_FILE} |
| 47 |
done |
| 48 |
fi |
| 49 |
done |
| 50 |
done |
| 51 |
|
| 52 |
# Print totals and averages |
| 53 |
awk '{ t += $2; c++ }; END { printf("====\nTotal compression: %.02f%% for %d documents (avg. %0.2f%%)\n", t, c, t / c) }' benchmark.output |