generate-traces
No OneTemporary
Actions

Subscribers

None

File Metadata

	#!/bin/bash

	#nr_jobs #shuffle_size(MB) #job_duration(s) #reduce_time


	parse(){
	FILE=$1
	RATIO=$2

	echo "--- --- --$FILE-- --- --- ---"
	for line in `cat $FILE \| grep -v "#" \| sed 's/[ \t]/*/g'`
	do
	REDUCER_INPUT_SIZE=`echo "$line" \| tr "" " "\| awk -vratio=$RATIO '{redrt=$3/ratio;nrtasks=int($4/redrt); red_in=$21024*1024/nrtasks;print red_in;}'`
	OPTIMAL_MEM=`java MainGetOpt $REDUCER_INPUT_SIZE`
	echo "$line" \| tr "" " "\| awk -vratio=$RATIO -vopt=$OPTIMAL_MEM '{redrt=$3/ratio;nrtasks=int($4/redrt);proc=opt/32000000000;print $1,nrtasks,opt,redrt,proc,int(int(proc10000)/100)*100;}'
	done
	echo
	}

	for idsds in 1
	do
	echo "#jobs #tasks_per_job #memory_per_task(bytes) #task_duration(sec) #%_of_32GB 10000"
	echo "#1 1 1000000000 1 1"
	echo "--- --- --------- ---"
	#2nd parameter to parse function is ratio between job running time and reducer running time
	parse CCa 1
	parse CCb 1
	parse CCc 1
	parse CCd 1
	parse CCe 1
	parse YH 1.02
	parse FB2009 0.36
	parse FB2010 0.36
	done \| column -t