From 37dd82c0a3363318958cd62cfdad2c73cc56338a Mon Sep 17 00:00:00 2001
From: Leandro Lucarella <llucax@gmail.com>
Date: Wed, 26 Aug 2009 21:06:17 -0300
Subject: [PATCH 1/1] micro: Improve generation of data to plot the histogram

If a program allocates a lot of cell with different  sizes the histogram
would have a lot of  bars to plot, making it very hard to read.

The script to generate the histogram data is improved to take a maximum
number of cell sizes. If that limit is passed, the cell sizes are grouped
together in ranges of cell sizes, keeping the histogram readable.
---
 micro/Makefile |  5 ++--
 micro/hist.awk | 78 ++++++++++++++++++++++++++++++++++++++++++++++++++
 2 files changed, 80 insertions(+), 3 deletions(-)
 create mode 100644 micro/hist.awk

diff --git a/micro/Makefile b/micro/Makefile
index e48a2e2..ddbbac7 100644
--- a/micro/Makefile
+++ b/micro/Makefile
@@ -72,9 +72,8 @@ $(STAT_DIR)/%.c.csv $(STAT_DIR)/%.a.csv: $(BIN_DIR)/%
 $(STAT_DIR)/split.c.csv $(STAT_DIR)/split.a.csv: override args := bible.txt
 
 .PRECIOUS: $(STAT_DIR)/%.h.csv
-$(STAT_DIR)/%.h.csv: $(STAT_DIR)/%.a.csv
-	$(P_AWK) awk -F, 'BEGIN  { print "Size,Count" } NR > 1 { a[$$3]++ } \
-			END { for (i in a) print i "," a[i] }' $< | sort > $@
+$(STAT_DIR)/%.h.csv: $(STAT_DIR)/%.a.csv hist.awk
+	$(P_AWK) awk -F, -f $(lastword $^) $< > $@
 
 .PHONY: plot
 plot: $(graphs)
diff --git a/micro/hist.awk b/micro/hist.awk
new file mode 100644
index 0000000..adeeb7b
--- /dev/null
+++ b/micro/hist.awk
@@ -0,0 +1,78 @@
+#!/usr/bin/env awk -F, -f
+
+BEGIN {
+	MAX_SAMPLES = 50
+	# output CSV header
+	print "Size,Scan,No Scan";
+}
+
+NR == 2 {
+	min = max = int($3)
+}
+
+NR > 1 { # skip the input CVS header
+	n = int($3)
+	if (int($6) > 0) { # cell has the NO_SCAN bit
+		no_scan[n]++
+		if (!(n in scan))
+			scan[n] = 0
+	} else { # cell doesn't have the NO_SCAN bit
+		scan[n]++
+		if (!(n in no_scan))
+			no_scan[n] = 0
+	}
+	if (n < min)
+		min = n
+	else if (n > max)
+		max = n
+}
+
+function h(val) {
+	if (val >= 1048576) # 1 M
+		r = sprintf("%uM", val / 1048576)
+	else if (val >= 1024) # 1 K
+		r = sprintf("%uK", val / 1024)
+	else
+		r = sprintf("%u", val)
+	return r
+}
+
+function p(s, ns, o, n) {
+	for (i = 1; i <= n; i++)
+		print o[i] "," s[o[i]] "," ns[o[i]]
+}
+
+END {
+	# reduce the number of elements in the histogram if there are too many
+	if (length(scan) > MAX_SAMPLES) {
+		step = int((max - min) / MAX_SAMPLES) + 1
+		for (i in scan) {
+			i = int(i)
+			for (from = min; from < max; from += step) {
+				to = from + step
+				if ((from <= i) && (i < to)) {
+					j = sprintf("%s-%s", h(from), h(to))
+					scan2[j] += scan[i]
+					no_scan2[j] += no_scan[i]
+					break
+				}
+			}
+		}
+		n = 1
+		for (from = min; from < max; from += step) {
+			v = sprintf("%s-%s", h(from), h(from + step))
+			if (v in scan2)
+				order[n++] = v
+		}
+	}
+	# print output data
+	if (length(scan2)) {
+		p(scan2, no_scan2, order, n)
+	} else {
+		for (i in scan)
+			order[i++] = int(i)
+		n = asort(order)
+		p(scan, no_scan, order, n)
+	}
+}
+
-- 
2.43.0