made the labels of plots work better with how they are presented in the results chapter

--HG--
rename : plots/scalability/speedup.plt => plots/scalability/efficiency.plt

plots/resolution/resolution.2D.regular.plt

@@ -9,6 +9,7 @@ set key right bottom
 
 set xrange [0.008:0.18]
 
+set xlabel 'Mesh Spacing'
 set ylabel 'RMS of error'
 
 plot './resolution.2D.64.out' \

plots/resolution/resolution.3D.gmsh.plt

@@ -10,7 +10,7 @@ set key right bottom
 set xrange [0.004:0.055]
 
 set xlabel 'Mesh Spacing'
-set ylabel 'RMS of error'
+# set ylabel 'RMS of error'
 
 plot './resolution.3D.gmsh.out' \
    u 1:2 t '{/Symbol n} = 2' w l,\

plots/scalability/efficiency.plt

@@ -0,0 +1,12 @@
+load '../main.plt'
+
+set output 'efficiency.eps'
+
+set nokey
+
+set ylabel "Efficiency (E_p)"
+set xlabel "Number of Participating Minions"
+
+set yrange [0:1]
+
+plot 'scale.out' u 1:3 w lp

plots/scalability/hist.plt

@@ -0,0 +1,20 @@
+set terminal postscript enhanced eps color
+set output "hist.eps"
+
+set key off
+
+# Make some suitable labels.
+set title "Work Performed per Participant"
+set xlabel "Number of Interpolations Performed"
+set ylabel "Number of Minions"
+
+
+set style histogram clustered gap 1
+set style fill solid border -1
+
+binwidth=5
+set boxwidth binwidth
+bin(x,width)=width*floor(x/width) + binwidth/2.0
+
+plot 'hist1.out' using (bin($1,binwidth)):(1.0) smooth freq with boxes
+# plot 'hist2.out' using (bin($1,binwidth)):(1.0) smooth freq with boxes

plots/scalability/hist.py

@@ -0,0 +1,22 @@
+import sys
+import shelve
+import pickle
+
+import numpy as np
+
+if len(sys.argv) != 2:
+  print "usage: %s <interp.shelve>" % sys.argv[0]
+  sys.exit(1)
+
+
+s = shelve.open(sys.argv[1])
+d = dict(s)
+s.close()
+
+for line in (i[1] for i in sorted(d.iteritems(), key = lambda x: x[1]['stats']['participants'])):
+  run = line['stats']
+  print "#", run['participants']
+  x = np.array(run['tasks'].values())
+  for i in x:
+    print i
+  print

plots/scalability/scalability_plot_gen.py

@@ -12,11 +12,15 @@ s.close()
 for line in (i[1] for i in sorted(d.iteritems(), key = lambda x: x[1]['stats']['participants'])):
   run = line['stats']
 
-  velocity = run['count'] / run['receive']
+  thruput = run['count'] / run['receive']
 
+  p = run['participants']
 
   # speedup (http://en.wikipedia.org/wiki/Speedup):
-  S_p = run['receive']/run['count'] / 1148.02904
+  T_1 = 0.114802904391
+  S_p =  T_1 / (run['receive']/run['count'])
 
-  print run['participants'], velocity, S_p
+  E_p = S_p / p
+
+  print p, S_p, E_p
 

plots/scalability/scale.out

@@ -1,19 +1,19 @@
-1 8.71058101972 0.000100000000341
-2 16.9174102062 5.14888564103e-05
-4 33.8342624632 2.57448527477e-05
-5 41.369278595 2.10556754801e-05
-8 70.8631618899 1.22921145728e-05
-11 95.1311787574 9.15638927551e-06
-16 140.277826152 6.20952098301e-06
-22 188.945066816 4.6101129795e-06
-32 275.671657764 3.15976662965e-06
-45 387.071286152 2.25038161213e-06
-64 550.492432027 1.58232530415e-06
-90 774.640026224 1.12446823744e-06
-128 1097.65200754 7.93564899404e-07
-181 1422.36034327 6.12403255659e-07
-196 1500.92599599 5.80347137214e-07
-224 1182.63366931 7.36540931943e-07
-256 1311.56362765 6.64137131117e-07
-384 1214.03638311 7.17489291971e-07
-512 1359.50184047 6.40718591923e-07
+1 0.999999999997 0.999999999997
+2 1.94216782645 0.971083913223
+4 3.8842715987 0.971067899674
+5 4.74931333527 0.949862667054
+8 8.1352967993 ?1.01691209991
+11 10.9213356195 0.992848692681
+16 16.1043018639 ?1.00651886649
+22 21.6914424408 0.985974656402
+32 31.6479069695 0.988997092798
+45 44.4369078566 0.987486841258
+64 63.198130042 0.987470781906
+90 88.930924868 0.988121387422
+128 126.013638476 0.984481550592
+181 163.291098498 0.902160765184
+196 172.310663615 0.879136038854
+224 135.769780068 0.606115089587
+256 150.571313747 0.588169194325
+384 139.374902817 0.362955476086
+512 156.074759811 0.304833515255

plots/scalability/server.out

@@ -1,3 +1,4 @@
+8      2
 16     7
 45    30
 64    40
@@ -7,4 +8,5 @@
 196  175
 224  185
 256  195
-512  195
+384  200
+512  200

plots/scalability/server.plt

@@ -7,5 +7,6 @@ set nokey
 set xlabel "Number of Participating Minions"
 set ylabel "Approximate CPU utilization (% CPU)"
 
-# set log xy
+set yrange [0:210]
+
 plot 'server.out' u 1:2 w lp t 'Speedup'

plots/scalability/speedup.plt

@@ -4,8 +4,7 @@ set output 'speedup.eps'
 
 set nokey
 
-set xlabel "Number of Participating Minions"
 set ylabel "Speedup (S_p)"
+set xlabel "Number of Participating Minions"
 
-set log xy
-plot 'scale.out' u 1:3 w lp t 'Speedup'
+plot 'scale.out' u 1:2 w lp t 'Speedup'

plots/scalability/thruput.plt

@@ -1,10 +0,0 @@
-load '../main.plt'
-
-set output 'thruput.eps'
-
-set nokey
-
-set xlabel "Number of Participating Minions"
-set ylabel "Interpolation Throughput (interpolations / second)"
-
-plot 'scale.out' u 1:2 w lp t 'velocity'

plots/timing/2D/onering.plt

@@ -3,7 +3,7 @@ load '../all.plt'
 
 set output 'timing.2.aggregate.eps'
 
-set title '2-D Interpolation Timing Data'
+# set title '2-D Interpolation Timing Data'
 
 plot 'time.out' u 1:2  t 'Order 2, res 1' w lp, \
      ''         u 1:3  t 'Order 3, res 1' w lp, \

plots/timing/3D/onering.plt

@@ -5,7 +5,7 @@ set output 'timing.3.aggregate.eps'
 
 unset ylabel
 
-set title '3-D Interpolation Timing Data'
+# set title '3-D Interpolation Timing Data'
 
 plot 'time.out' u 1:2  t 'Order 2, Mesh 1' w lp, \
      ''         u 1:3  t 'Order 3, Mesh 1' w lp, \

plots/truthiness/all.plt

@@ -3,6 +3,6 @@ set ylabel "Percent of Improved Interpolations"
 
 set log x
 
-set yrange [0:1]
+# set yrange [0:1]
 
 set key bottom right