diff --git a/makeplots.R b/makeplots.R index b627d0b3f7d988bc8eba5a45896cb9f2d7e4578c..0427845d0cb65f65557882bb652b1b3ada730e17 100755 --- a/makeplots.R +++ b/makeplots.R @@ -27,6 +27,33 @@ se <- function(x) sqrt(var(x)/length(x)) +# survey +d <- all %>% filter(exp=="survey") %>% group_by(sys, s, conf) %>% + summarize(meant=mean(timesec), se=se(timesec)) %>% + mutate(tool=paste(sys,conf), + datasize=sapply(s, switch, alabama=47512, california=1060060, acs3yr=9093077)) %>% + filter(tool == "Renjin jitopt" | sys != "Renjin") + +print(d %>% select(sys,s,meant) %>% arrange(s)) + +limits <- aes(ymax = meant + se, ymin=meant - se, width=2) + +pdf("survey.pdf",width=10,height=7) +ggplot(d,aes(datasize,meant,group=tool)) + + geom_point(size=4) + geom_line(size=1.5, aes(group=tool, linetype=tool)) + + geom_pointrange(limits) + + #scale_y_log10(breaks=ybreaks, labels=ylabels) + + scale_x_log10(breaks=c(47512,1060060,9093077)) + + + xlab("Dataset Size (elements, log scale)") + ylab("Execution Time (s)") + theme + + annotate("text", x=10^6.6, y=45, label="GNU R", family="serif", size=10)+ + annotate("text", x=10^6.4, y=100, label="sqlsurvey", family="serif", size=10)+ + annotate("text", x=10^6.8, y=15, label="Renjin", family="serif", size=10) + + #scale_color_brewer(palette=cBrwPl) + + #guides(colour=guide_legend(keywidth=3.5)) + +dev.off() @@ -79,7 +106,6 @@ dev.off() # dev.off() -stop() # recycling d <- all %>% filter(exp=="recycling", s > 5) %>% group_by(exp, sys, s, conf) %>% summarize(meant=mean(timesec), se=se(timesec)) %>% @@ -105,33 +131,6 @@ p <- ggplot(d,aes(datasize,meant,group=tool)) + print(p) dev.off() -# survey -d <- all %>% filter(exp=="survey") %>% group_by(sys, s, conf) %>% - summarize(meant=mean(timesec), se=se(timesec)) %>% - mutate(tool=paste(sys,conf), - datasize=sapply(s, switch, alabama=47512, california=1060060, acs3yr=9093077)) %>% - filter(tool == "Renjin jitopt" | sys != "Renjin") - -print(d) - -limits <- aes(ymax = meant + se, ymin=meant - se, width=2) - -pdf("survey.pdf",width=10,height=7) -ggplot(d,aes(datasize,meant,group=tool)) + - geom_point(size=4) + geom_line(size=1.5, aes(group=tool, linetype=tool)) + - geom_pointrange(limits) + - #scale_y_log10(breaks=ybreaks, labels=ylabels) + - scale_x_log10(breaks=c(47512,1060060,9093077)) + - - xlab("Dataset Size (elements, log scale)") + ylab("Execution Time (s)") + theme + - annotate("text", x=10^6.6, y=45, label="GNU R", family="serif", size=10)+ - annotate("text", x=10^6.4, y=100, label="sqlsurvey", family="serif", size=10)+ - annotate("text", x=10^6.8, y=15, label="Renjin", family="serif", size=10) - - #scale_color_brewer(palette=cBrwPl) + - #guides(colour=guide_legend(keywidth=3.5)) - -dev.off() # # identity @@ -163,7 +162,7 @@ dev.off() # pushdown -d <- all %>% filter(exp=="pushdown", s > 4) %>% group_by(exp, sys, s) %>% +d <- all %>% filter(exp=="pushdown", s > 5) %>% group_by(exp, sys, s) %>% summarize(meant=mean(timesec), se=se(timesec)) %>% mutate(tool=sys, datasize=10^as.integer(s)) @@ -179,7 +178,7 @@ p <- ggplot(d,aes(datasize,meant,group=tool)) + scale_x_log10(breaks=xbreaks, labels=xlabels) + xlab("Dataset Size (elements, log scale)") + ylab("Execution Time (s)") + theme + - annotate("text", x=10^7.5, y=5, label="GNU R", family="serif", size=10)+ + annotate("text", x=10^7.6, y=5, label="GNU R", family="serif", size=10)+ annotate("text", x=10^7.6, y=.4, label="Renjin ", family="serif", size=10) print(p)