OK, that's obviously insane. Yet here we are. See this Reddit thread for discussion.
The original README describes the original usage and some genesis of this code. I've changed it quite a bit.
As before you need to start by generating the data to be
benchmarked over. Go to the gen/
subdirectory and say
cargo run --release ../acgt
This will take a while and generate 3GB of data.
Now go back to the top and say
sh try.sh
This will build the two binary versions of the counter
program (counts instances of G
and C
in the datafile)
as debug
and ndebug
, and run each of them in sequential
mode with timing information.
From there, have fun with it. My friend and I did.