AES algorithm in cuda and CPU/GPU openmp
Developed based on https://github.com/franneck94/CUDA-AES
Book from http://www.crypto-textbook.com/ Chapter4
All the tests are under NYU Greene cluster. All makefiles are well-written in the subdirectories.
Note that there are some sample text files to test, so text_file_number
below can be 1,2,3,4,41,5,51,6,61
. Default set to 1
.
- AES_CUDA compile:
ssh burst
# distribute a GPU node
srun --account=csci_ga_3033_025-2022fa --partition=n1s8-v100-1 --gres=gpu:v100:1 --time=00:40:00 --pty /bin/bash
# Use a singularity image
/share/apps/gcc/run-cuda-11.6.bash
# cd to the AES_CUDA directory
make
Usage:
./AES_CUDA text_file_number
#use nvprof to profile GPU utilization:
nvprof ./AES_CUDA text_file_number
- AES_SEQ, AES_OMPC, AES_OMPG compile:
ssh burst
# distribute a GPU node
srun --account=csci_ga_3033_025-2022fa --partition=n1s8-v100-1 --gres=gpu:v100:1 --time=00:40:00 --pty /bin/bash
# Use a singularity image (different than above)
/share/apps/gcc/run-gcc-offload-nvptx-12.2.0.bash
# cd to the AES_() directory
make
Usage:
./AES_SEQ text_file_number
./AES_OMPC text_file_number num_threads
./AES_OMPG text_file_number num_threads
- The unit tests are in
AES/exec/run.sh
, the output is inAES/exec/out.txt
, and the executables in NYU Greene cluster GPU are inAES/exec/
.
sh run.sh > out.txt 2>&1 &