Comment shell scripts.

fast/openmp.sh

@@ -1,30 +1,52 @@
 #!/bin/bash
 
+# Simple script to run multithreaded benchmarks locally,
+# for simple sanity checks.
+
+# Use OpenMP
 export DEVITO_LANGUAGE=openmp
+# Use the cray compiler, if available.
 export DEVITO_ARCH=cray
+# Enable debug logging.
 export DEVITO_LOGGING=DEBUG
+# Disable (tile size) autotuning for iteration speed.
+# I enable it sometimes; NB enabling it requires that no explicit tile size
+# is specified in the Operator constructor args.
 unset DEVITO_AUTOTUNING
 
+# Bind threads to physical cores
 export OMP_PLACES=cores
 export OMP_PROC_BIND=true
 
+# Just extract the reported runtime from the whole output of the passed command
 get_runtime() {
     $@ |& grep 'Operator.*ran' | rev | cut -d ' ' -f2 | rev
 }
 
+# Iterate over benchmarks and cases, print simple CSV data to stdout
+# Copy-pastes nicely in Google Sheets
 echo bench_name,so,threads,Devito,xDSL
 for bench in "wave2d_b.py -d 2048 2048 --nt 512" "wave3d_b.py -d 512 512 512 --nt 512" "diffusion_3D_wBCs.py -d 512 512 512 --nt 512" "diffusion_2D_wBCs.py -d 2048 2048 --nt 512"
 do
+  # Get the benchmark file for printing
   bench_name=$(echo $bench | cut -d ' ' -f1)
+  # Iterate over measured space orders
   for so in 2 4 8
   do
+    # Iterate over measured thread numbers
     for threads in 1 2 4 8 16 32
     do
+      # Set the thread number to use
       export OMP_NUM_THREADS=$threads
+
+      # To uncomment to check what's going on without capturing the output.
       # echo OMP_NUM_THREADS=$threads
       # python $bench -so $so --devito 1
+
+      # Get the runtimes
       devito_time=$(get_runtime python $bench -so $so --devito 1)
       xdsl_time=$(get_runtime python $bench -so $so --xdsl 1)
+      # print CSV line
       echo $bench_name,$so,$threads,$devito_time,$xdsl_time
     done
   done

fast/single_thread.sh

@@ -1,22 +1,38 @@
 #!/bin/bash
 
+# Simple script to run single threaded benchmarks locally,
+# for simple sanity checks.
+
+# Don't use OpenMP
 unset DEVITO_LANGUAGE
+# Use the cray compiler, if available.
 export DEVITO_ARCH=cray
+# Enable debug logging.
 export DEVITO_LOGGING=DEBUG
+# Enable (tile size) autotuning
+# NB enabling it requires that no explicit tile size
+# is specified in the Operator constructor args.
 export DEVITO_AUTOTUNING=aggressive
 
+# Just extract the reported runtime from the whole output of the passed command
 get_runtime() {
     $@ |& grep 'Operator.*ran' | rev | cut -d ' ' -f2 | rev
 }
 
+# Iterate over benchmarks and cases, print simple CSV data to stdout
+# Copy-pastes nicely in Google Sheets
 echo bench_name,so,Devito,xDSL
 for bench in "wave2d_b.py -d 2048 2048 --nt 512" "wave3d_b.py -d 512 512 512 --nt 512" "diffusion_3D_wBCs.py -d 512 512 512 --nt 512" "diffusion_2D_wBCs.py -d 2048 2048 --nt 512"
 do
+  # Get the benchmark file for printing
   bench_name=$(echo $bench | cut -d ' ' -f1)
+  # Iterate over measured space orders
   for so in 2 4 8
     do
+      # Get the runtimes
       devito_time=$(get_runtime python $bench -so $so --devito 1)
       xdsl_time=$(get_runtime python $bench -so $so --xdsl 1)
+      # print CSV line
       echo $bench_name,$so,$devito_time,$xdsl_time
   done
 done