Difference between revisions of "CSC352 Homework 5 Solution 2013"
(Created page with "--~~~~ ---- <onlysmith> The solution this week is provided by Gavi, with an application that generates a very good graph. More on the graph in class. =Script= <source lang="...") |
|||
Line 2: | Line 2: | ||
---- | ---- | ||
<onlysmith> | <onlysmith> | ||
− | The solution this week is provided by Gavi, with an application that generates a very good graph. More on the graph in class. | + | The solution this week is provided by Gavi, with an application that generates a very good graph. More on the graph in class. I am also sharing Emily's scripts for running the experiments and generating the graph. |
=Script= | =Script= | ||
Line 416: | Line 416: | ||
<center>[[Image:CSC352hw5graphGavi.png|700px]]</center> | <center>[[Image:CSC352hw5graphGavi.png|700px]]</center> | ||
+ | =More Scripts= | ||
+ | |||
+ | ===Script to run the experiment=== | ||
+ | <source lang="bash"> | ||
+ | #!/bin/bash | ||
+ | # | ||
+ | # Emily Flynn | ||
+ | # CSC 352 | ||
+ | # 11/14/2013 | ||
+ | # | ||
+ | # Gets timing data for hadoop (I only did one run on aws to save time). | ||
+ | |||
+ | |||
+ | N=$1 | ||
+ | header="M run1 run2 run3" | ||
+ | echo $header > timing${N}.txt | ||
+ | |||
+ | for m in "10" "50" "100" "250" "500" "1000" "2500" "5000" | ||
+ | do | ||
+ | &>out time mpirun -np 8 ./hw5 ${N} ${m} > /dev/null | ||
+ | result1=$(head -1 out | awk '{print $1;}') #grabs only the user time | ||
+ | &>out time mpirun -np 8 ./hw5 ${N} ${m} > /dev/null | ||
+ | result2=$(head -1 out | awk '{print $1;}') | ||
+ | &>out time mpirun -np 8 ./hw5 ${N} ${m} > /dev/null | ||
+ | result3=$(head -1 out | awk '{print $1;}') | ||
+ | |||
+ | # write it out to a file | ||
+ | printf "%s\t%s\t%s\t%s\n" "$m" "$result1" "$result2" "$result3" >> timing${N}.txt | ||
+ | |||
+ | done | ||
+ | |||
+ | rm out | ||
+ | |||
+ | exit 0 | ||
+ | |||
+ | </source> | ||
+ | |||
+ | |||
+ | |||
+ | ===Script to generate a Graph in R=== | ||
+ | <source lang="rsplus"> | ||
+ | # plotHw5TimingData.R | ||
+ | # Emily Flynn | ||
+ | # CSC 352 | ||
+ | # 11/14/2013 | ||
+ | # | ||
+ | # Plots the timing data from hadoop0, aws timing with N=5000. | ||
+ | |||
+ | # helpful package | ||
+ | require(mosaic) | ||
+ | hadoop0 <- read.table("timing5000.txt", header=TRUE) | ||
+ | aws <- read.table("timing_aws5000.txt", header=TRUE) | ||
+ | |||
+ | # process timing data to remove "user" and convert data to numeric | ||
+ | removeUser<- function(x){ | ||
+ | return(as.numeric(gsub("user", "", x))) | ||
+ | } | ||
+ | hadoop0 <- data.frame(apply(hadoop0, c(1,2), removeUser)) | ||
+ | aws <-data.frame(apply(aws, c(1,2), removeUser)) | ||
+ | colnames(aws) <- c("M", "time") # aws only contains one run, not complete | ||
+ | aws$system <- "aws" | ||
+ | |||
+ | hadoop0$avg <- rowMeans(hadoop0[,2:4]) # hadoop contains multiple runs | ||
+ | |||
+ | hadoop0_short <- data.frame(hadoop0$M, hadoop0$avg) | ||
+ | colnames(hadoop0_short) <- c("M", "time") | ||
+ | hadoop0_short$system <- "hadoop" | ||
+ | |||
+ | result <- rbind(aws, hadoop0_short) | ||
+ | my.col=c("blue","magenta") | ||
+ | |||
+ | plotPoints(time ~ M, groups=system, data=result, ylab="time (s)", xlab="M - number per packet", | ||
+ | main="Hadoop0 vs. AWS times for varying packet size, \nN=5000", | ||
+ | key = list(space = "right", adj = 1, | ||
+ | text = list(c("aws", "hadoop0")), | ||
+ | points = list(pch = 21, col=my.col)) | ||
+ | |||
+ | |||
+ | ) | ||
+ | |||
+ | </source> | ||
</onlysmith> | </onlysmith> | ||
Revision as of 12:34, 26 November 2013
--D. Thiebaut (talk) 09:45, 26 November 2013 (EST)