Python version required: python3.6 If you are about to autotune parameters for a new GPU (i.e. a GPU for which there are no autotuned parameters yet), please first follow these instructions.

$ cd dbcsr/src/acc/libsmm_acc/libcusmm

The tune_setup.py script generates job files. You have to adapt the script to the environment of your supercomputer and your personal settings.

...
def gen_jobfile(outdir, m, n, k):
    t = "/tune_%dx%dx%d"%(m,n,k)
    all_exe_src = [basename(fn) for fn in glob(outdir+t+"_*_main.cu")]
    all_exe = sorted([fn.replace("_main.cu", "") for fn in all_exe_src])
 
    output = "#!/bin/bash -l\n"
    output += "#SBATCH --nodes=%d\n"%len(all_exe)
    output += "#SBATCH --time=0:30:00\n"
    output += "#SBATCH --account=s441\n"
    output += "\n"
    output += "source ${MODULESHOME}/init/sh;\n"
    output += "module unload PrgEnv-cray\n"
    output += "module load cudatoolkit PrgEnv-gnu\n"
    output += "module list\n"
    output += "cd $SLURM_SUBMIT_DIR \n"
    output += "\n"
    output += "date\n"
    for exe in all_exe:
        output += "aprun -b -n 1 -N 1 -d 8 make -j 16 %s &\n"%exe
   ...

The script takes as arguments the blocksizes you want to add to libcusmm. For example, if your system contains blocks of size 5 and 8 type:

$ ./tune_setup.py 5 8
Found 23 parameter sets for 5x5x5
Found 31 parameter sets for 5x5x8
Found 107 parameter sets for 5x8x5
Found 171 parameter sets for 5x8x8
Found 75 parameter sets for 8x5x5
Found 107 parameter sets for 8x5x8
Found 248 parameter sets for 8x8x5
Found 424 parameter sets for 8x8x8

The script will create a directory for each combination of the blocksizes:

$ ls -d tune_*
tune_5x5x5  tune_5x5x8  tune_5x8x5  tune_5x8x8  tune_8x5x5  tune_8x5x8  tune_8x8x5  tune_8x8x8

Each directory contains a number of files:

$ ls -1 tune_8x8x8/
Makefile
tune_8x8x8_exe0_main.cu
tune_8x8x8_exe0_part0.cu
tune_8x8x8_exe0_part1.cu
tune_8x8x8_exe0_part2.cu
tune_8x8x8_exe0_part3.cu
tune_8x8x8_exe0_part4.cu
tune_8x8x8.job

For each possible parameter-set a launcher is generated. A launcher is a small snippet of C code, which launches the kernel by using the cuda specific <<< >>>-notation. It also instantiates the C++ template which contains the actual kernel code.

In order to parallelize the benchmarking, the launchers are distributed over multiple executables. Currently, up to 10000 launchers are benchmarked by one executable. Each executable is linked together from several tune_*_part???.o and a tune_*_main.o. Each part-files contains up to 100 launchers. This allows to parallelize the compilation over multiple CPU cores.

The script tune_submit.py was written for the slurm batch system as used e.g. by CRAY supercomputers. If your computer runs a different batch system you have to adapt tune_submit.py accordingly.

Each tune-directory contains a job file. Since there might be many tune-directories, the convenience script tune_submit.py can be used to submit jobs. It will go through all the tune_*-directories and check if its job has already been submitted or run. For this, the script calls squeue in the background and it searches for slurm-*.out files.

When tune_submit.py is called without arguments, it will just list the jobs that could be submitted:

$ ./tune_submit.py 
          tune_5x5x5: Would submit, run with "doit!"
          tune_5x5x8: Would submit, run with "doit!"
          tune_5x8x5: Would submit, run with "doit!"
          tune_5x8x8: Would submit, run with "doit!"
          tune_8x5x5: Would submit, run with "doit!"
          tune_8x5x8: Would submit, run with "doit!"
          tune_8x8x5: Would submit, run with "doit!"
          tune_8x8x8: Would submit, run with "doit!"
Number of jobs submitted: 8

Only when tune_submit.py is called with doit! as its first argument, will it actually submit jobs:

$ ./tune_submit.py doit!
          tune_5x5x5: Submitting
Submitted batch job 277987
          tune_5x5x8: Submitting
Submitted batch job 277988
          tune_5x8x5: Submitting
Submitted batch job 277989
          tune_5x8x8: Submitting
Submitted batch job 277990
          tune_8x5x5: Submitting
Submitted batch job 277991
          tune_8x5x8: Submitting
Submitted batch job 277992
          tune_8x8x5: Submitting
Submitted batch job 277993
          tune_8x8x8: Submitting
Submitted batch job 277994
Number of jobs submitted: 8

Run tune_collect.py to parse all log files and to determine the best kernel for each blocksize:

$ ./tune_collect.py
Reading: tune_5x5x5/tune_5x5x5_exe0.log
Reading: tune_5x5x8/tune_5x5x8_exe0.log
Reading: tune_5x8x5/tune_5x8x5_exe0.log
Reading: tune_5x8x8/tune_5x8x8_exe0.log
Reading: tune_8x5x5/tune_8x5x5_exe0.log
Reading: tune_8x5x8/tune_8x5x8_exe0.log
Reading: tune_8x8x5/tune_8x8x5_exe0.log
Reading: tune_8x8x8/tune_8x8x8_exe0.log
Kernel_dnt_tiny(m=5, n=5, k=5, split_thread=32, threads=64, grouping=16, minblocks=1) , # 27.9623 GFlops 
Kernel_dnt_tiny(m=5, n=5, k=8, split_thread=32, threads=96, grouping=16, minblocks=1) , # 37.8978 GFlops
Kernel_dnt_medium(m=5, n=8, k=5, tile_m=1, tile_n=1, threads=96, grouping=16, minblocks=8) , # 32.9231 GFlops 
Kernel_dnt_tiny(m=5, n=8, k=8, split_thread=32, threads=96, grouping=16, minblocks=1) , # 47.0366 GFlops
Kernel_dnt_medium(m=8, n=5, k=5, tile_m=1, tile_n=1, threads=96, grouping=16, minblocks=12) , # 33.1999 GFlops 
Kernel_dnt_medium(m=8, n=5, k=8, tile_m=1, tile_n=1, threads=96, grouping=16, minblocks=12) , # 49.3499 GFlops
Kernel_dnt_tiny(m=8, n=8, k=5, split_thread=32, threads=96, grouping=16, minblocks=1) , # 62.8469 GFlops 
Kernel_dnt_tiny(m=8, n=8, k=8, split_thread=32, threads=128, grouping=16, minblocks=1) , # 90.7763 GFlops