Hi Drizzle, During a merge request for:
https://blueprints.launchpad.net/drizzle/+spec/replace-macro-functions Stewart questioned updates of SHA1, see: https://code.launchpad.net/~gthorslund/drizzle/replace-macro-functions/+merge/39894 I've done some benchmarking now. Attached test case. Unpack and do: shell> ./run_bench.sh Should give a result like: ---------------------------------------- -O0 sha1 -rwxr-xr-x 1 gustaf gustaf 23748 2010-11-05 17:33 benchmark_sha1 -rw-r--r-- 1 gustaf gustaf 10928 2010-11-05 17:33 sha1.o b43859b53bf90da1951b6ab59cc964f3a8671a 0.836730975 ---------------------------------------- -O0 sha1-inline-static -rwxr-xr-x 1 gustaf gustaf 24286 2010-11-05 17:33 benchmark_sha1 -rw-r--r-- 1 gustaf gustaf 8976 2010-11-05 17:33 sha1-inline-static.o b43859b53bf90da1951b6ab59cc964f3a8671a 3.391155227 ---------------------------------------- -O1 sha1 -rwxr-xr-x 1 gustaf gustaf 18821 2010-11-05 17:33 benchmark_sha1 -rw-r--r-- 1 gustaf gustaf 7936 2010-11-05 17:33 sha1.o b43859b53bf90da1951b6ab59cc964f3a8671a 0.248819991 ---------------------------------------- -O1 sha1-inline-static -rwxr-xr-x 1 gustaf gustaf 18835 2010-11-05 17:33 benchmark_sha1 -rw-r--r-- 1 gustaf gustaf 7952 2010-11-05 17:33 sha1-inline-static.o b43859b53bf90da1951b6ab59cc964f3a8671a 0.247948874 ---------------------------------------- -O2 sha1 -rwxr-xr-x 1 gustaf gustaf 22969 2010-11-05 17:33 benchmark_sha1 -rw-r--r-- 1 gustaf gustaf 8976 2010-11-05 17:33 sha1.o b43859b53bf90da1951b6ab59cc964f3a8671a 0.353121104 ---------------------------------------- -O2 sha1-inline-static -rwxr-xr-x 1 gustaf gustaf 22983 2010-11-05 17:33 benchmark_sha1 -rw-r--r-- 1 gustaf gustaf 8992 2010-11-05 17:33 sha1-inline-static.o b43859b53bf90da1951b6ab59cc964f3a8671a 0.350880366 ---------------------------------------- -O3 sha1 -rwxr-xr-x 1 gustaf gustaf 22969 2010-11-05 17:33 benchmark_sha1 -rw-r--r-- 1 gustaf gustaf 9288 2010-11-05 17:33 sha1.o b43859b53bf90da1951b6ab59cc964f3a8671a 0.351405901 ---------------------------------------- -O3 sha1-inline-static -rwxr-xr-x 1 gustaf gustaf 22983 2010-11-05 17:33 benchmark_sha1 -rw-r--r-- 1 gustaf gustaf 9304 2010-11-05 17:33 sha1-inline-static.o b43859b53bf90da1951b6ab59cc964f3a8671a 0.348646827 ---------------------------------------- -Os sha1 -rwxr-xr-x 1 gustaf gustaf 18583 2010-11-05 17:33 benchmark_sha1 -rw-r--r-- 1 gustaf gustaf 8776 2010-11-05 17:33 sha1.o b43859b53bf90da1951b6ab59cc964f3a8671a 0.337403311 ---------------------------------------- -Os sha1-inline-static -rwxr-xr-x 1 gustaf gustaf 18807 2010-11-05 17:33 benchmark_sha1 -rw-r--r-- 1 gustaf gustaf 7248 2010-11-05 17:33 sha1-inline-static.o b43859b53bf90da1951b6ab59cc964f3a8671a 0.602943397 So except for -O0 and -Os they give similar results. Interesting -O1 goes fastest. I removed the code from my branch before merging, but I can put it in a separate branch if interesting. Someone might also be interested in comparing with other implementations of SHA1. /Gustaf
sha1_test.tar.bz2
Description: application/bzip
_______________________________________________ Mailing list: https://launchpad.net/~drizzle-discuss Post to : [email protected] Unsubscribe : https://launchpad.net/~drizzle-discuss More help : https://help.launchpad.net/ListHelp
