1. Download the simulation infrastructure: http://hpca23.cse.tamu.edu/cbp2016/cbp2016.final.tar.gz
2. Download the training traces : http://hpca23.cse.tamu.edu/cbp2016/trainingTraces.Final.tar
3. Download the evaluation traces : http://hpca23.cse.tamu.edu/cbp2016/evaluationTraces.Final.tar
4. Download the simulation results from the 2016 competition : http://hpca23.cse.tamu.edu/cbp2016/cbp2016_evaluation_results.tar.gz
5. MD5SUM for all of the above : http://hpca23.cse.tamu.edu/cbp2016/MD5SUM.txt
6. Unpack the kit:
tar -xzvf cbp2016.tar.gz
7. There should be five directories: sim, scripts, traces, bin, and results
8. You will need to install the BOOST library and add a link to the library to the Makefile in the sim directory. See comments in the Makefile.
9. The sim directory contains the simulator.
make clean; make
10. The scripts directory contains scripts that can help you run
your predictor for all 223 traces. Check out the doit.sh file
in the scripts directory.
11. We will use the AMEAN over all traces as the figure of merit. You can compute this using the getdata.pl script. Usage is as follows.
./getdata.pl -d ../results/GSHARE*
In your sim directory, there are two files you can modify: predictor.cc and predictor.h
Only these two files should be submitted with your contest
submission. Any changes you make to any other files will not
be reflected in our simulator.