Problem in writing a parallel version of my Bash code

I am trying to parallelise my sample Bash script and have tried commands like & and wait. Please let me know what is an effective way to make it parallel My current code is working fine for limited entries in the reg2 variable. But I have millions of entries in the reg2 variable. So I want to make my outermost loop parallel. To get the same output i.e., 0,1,2,:,3,4,:,5,6, after parallelizing the code #!/bin/bash # array1=$1 # array2=($2) # reg2=($3) array1=('bam1' 'bam2' 'bam3' 'bam4' 'bam5' 'bam6' 'bam7') array2=('cell1' 'cell1' 'cell1' 'cell2' 'cell2' 'cell3' 'cell3') reg2=('chr1:10484-10572' 'chr1:10589-10632' 'chr1:10636-10661' 'chr1:10665-10690' 'chr1:10694-10719') start=date +%s.%N l=${#reg2[@]} # number of regions is 30 million on real data reg_cov=() j=0 for r in ${reg2[@]}; do (cov_array=() old_array2_element=${array2} for i in ${!array1[*]}; do new_array2_element=${array2[$i]} if [[ "$new_array2_element" != "$old_array2_element" ]]; then cov_array+=(":") old_array2_element=$new_array2_element fi cov_array+=($i) # in actual code this step takes 4-5 seconds to process sleep 2 done reg_cov+=($(IFS=, ; echo "${cov_array[*]}")) ) wait ((j++)) echo "$j/$l" done #echo ${reg_cov[@]} cov=() cov+=(${reg_cov[@]}) echo $cov end=date +%s.%N; runtime=$( echo "$end - $start" | bc -l ); runtime=${runtime%.*}; hours=$((runtime / 3600)); minutes=$(( (runtime % 3600) / 60 )); seconds=$(( (runtime % 3600) % 60 )) echo "==> completed Runtime: $hours:$minutes:$seconds (hh:mm:ss)"