Use #!/usr/bin/env to allow for interpreters in different directories.
[fio.git] / tools / plot / fio2gnuplot
... / ...
CommitLineData
1#!/usr/bin/env python
2#
3# Copyright (C) 2013 eNovance SAS <licensing@enovance.com>
4# Author: Erwan Velu <erwan@enovance.com>
5#
6# The license below covers all files distributed with fio unless otherwise
7# noted in the file itself.
8#
9# This program is free software; you can redistribute it and/or modify
10# it under the terms of the GNU General Public License version 2 as
11# published by the Free Software Foundation.
12#
13# This program is distributed in the hope that it will be useful,
14# but WITHOUT ANY WARRANTY; without even the implied warranty of
15# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
16# GNU General Public License for more details.
17#
18# You should have received a copy of the GNU General Public License
19# along with this program; if not, write to the Free Software
20# Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
21
22import os
23import fnmatch
24import sys
25import getopt
26import re
27import math
28import shutil
29
30def find_file(path, pattern):
31 fio_data_file=[]
32 # For all the local files
33 for file in os.listdir(path):
34 # If the file math the regexp
35 if fnmatch.fnmatch(file, pattern):
36 # Let's consider this file
37 fio_data_file.append(file)
38
39 return fio_data_file
40
41def generate_gnuplot_script(fio_data_file,title,gnuplot_output_filename,gnuplot_output_dir,mode,disk_perf,gpm_dir):
42 if verbose: print "Generating rendering scripts"
43 filename=gnuplot_output_dir+'mygraph'
44 temporary_files.append(filename)
45 f=open(filename,'w')
46
47 # Plotting 3D or comparing graphs doesn't have a meaning unless if there is at least 2 traces
48 if len(fio_data_file) > 1:
49 f.write("call \'%s/graph3D.gpm\' \'%s' \'%s\' \'\' \'%s\' \'%s\'\n" % (gpm_dir,title,gnuplot_output_filename,gnuplot_output_filename,mode))
50
51 # Setting up the compare files that will be plot later
52 compare=open(gnuplot_output_dir + 'compare.gnuplot','w')
53 compare.write('''
54set title '%s'
55set terminal png size 1280,1024
56set ytics axis out auto
57set key top left reverse
58set xlabel "Time (Seconds)"
59set ylabel '%s'
60set yrange [0:]
61set style line 1 lt 1 lw 3 pt 3 linecolor rgb "green"
62'''% (title,mode))
63 compare.close()
64 #Copying the common file for all kind of graph (raw/smooth/trend)
65 compare_raw_filename="compare-%s-2Draw" % (gnuplot_output_filename)
66 compare_smooth_filename="compare-%s-2Dsmooth" % (gnuplot_output_filename)
67 compare_trend_filename="compare-%s-2Dtrend" % (gnuplot_output_filename)
68
69 shutil.copy(gnuplot_output_dir+'compare.gnuplot',gnuplot_output_dir+compare_raw_filename+".gnuplot")
70 shutil.copy(gnuplot_output_dir+'compare.gnuplot',gnuplot_output_dir+compare_smooth_filename+".gnuplot")
71 shutil.copy(gnuplot_output_dir+'compare.gnuplot',gnuplot_output_dir+compare_trend_filename+".gnuplot")
72 temporary_files.append(gnuplot_output_dir+compare_raw_filename+".gnuplot")
73 temporary_files.append(gnuplot_output_dir+compare_smooth_filename+".gnuplot")
74 temporary_files.append(gnuplot_output_dir+compare_trend_filename+".gnuplot")
75
76 #Setting up a different output filename for each kind of graph
77 compare_raw=open(gnuplot_output_dir+compare_raw_filename + ".gnuplot",'a')
78 compare_raw.write("set output '%s.png'\n" % compare_raw_filename)
79 compare_smooth=open(gnuplot_output_dir+compare_smooth_filename+".gnuplot",'a')
80 compare_smooth.write("set output '%s.png'\n" % compare_smooth_filename)
81 compare_trend=open(gnuplot_output_dir+compare_trend_filename+".gnuplot",'a')
82 compare_trend.write("set output '%s.png'\n" % compare_trend_filename)
83
84 # Let's plot the average value for all the traces
85 global_disk_perf = sum(disk_perf, [])
86 global_avg = average(global_disk_perf)
87 compare_raw.write("plot %s w l ls 1 ti 'Global average value (%.2f)'" % (global_avg,global_avg));
88 compare_smooth.write("plot %s w l ls 1 ti 'Global average value (%.2f)'" % (global_avg,global_avg));
89 compare_trend.write("plot %s w l ls 1 ti 'Global average value (%.2f)'" % (global_avg,global_avg));
90
91 pos=0
92 # Let's create a temporary file for each selected fio file
93 for file in fio_data_file:
94 tmp_filename = "gnuplot_temp_file.%d" % pos
95
96 # Plotting comparing graphs doesn't have a meaning unless if there is at least 2 traces
97 if len(fio_data_file) > 1:
98 # Adding the plot instruction for each kind of comparing graphs
99 compare_raw.write(",\\\n'%s' using 2:3 with linespoints title '%s'" % (tmp_filename,fio_data_file[pos]))
100 compare_smooth.write(",\\\n'%s' using 2:3 smooth csplines title '%s'" % (tmp_filename,fio_data_file[pos]))
101 compare_trend.write(",\\\n'%s' using 2:3 smooth bezier title '%s'" % (tmp_filename,fio_data_file[pos]))
102
103 png_file=file.replace('.log','')
104 raw_filename = "%s-2Draw" % (png_file)
105 smooth_filename = "%s-2Dsmooth" % (png_file)
106 trend_filename = "%s-2Dtrend" % (png_file)
107 avg = average(disk_perf[pos])
108 f.write("call \'%s/graph2D.gpm\' \'%s' \'%s\' \'%s\' \'%s\' \'%s\' \'%s\' \'%s\' \'%f\'\n" % (gpm_dir,title,tmp_filename,fio_data_file[pos],raw_filename,mode,smooth_filename,trend_filename,avg))
109 pos = pos +1
110
111 # Plotting comparing graphs doesn't have a meaning unless if there is at least 2 traces
112 if len(fio_data_file) > 1:
113 os.remove(gnuplot_output_dir+"compare.gnuplot")
114 compare_raw.close()
115 compare_smooth.close()
116 compare_trend.close()
117 f.close()
118
119def generate_gnuplot_math_script(title,gnuplot_output_filename,mode,average,gnuplot_output_dir,gpm_dir):
120 filename=gnuplot_output_dir+'mymath';
121 temporary_files.append(filename)
122 f=open(filename,'a')
123 f.write("call \'%s/math.gpm\' \'%s' \'%s\' \'\' \'%s\' \'%s\' %s\n" % (gpm_dir,title,gnuplot_output_filename,gnuplot_output_filename,mode,average))
124 f.close()
125
126def compute_aggregated_file(fio_data_file, gnuplot_output_filename, gnuplot_output_dir):
127 if verbose: print "Processing data file 2/2"
128 temp_files=[]
129 pos=0
130
131 # Let's create a temporary file for each selected fio file
132 for file in fio_data_file:
133 tmp_filename = "%sgnuplot_temp_file.%d" % (gnuplot_output_dir, pos)
134 temp_files.append(open(tmp_filename,'r'))
135 pos = pos +1
136
137 f = open(gnuplot_output_dir+gnuplot_output_filename, "w")
138 temporary_files.append(gnuplot_output_dir+gnuplot_output_filename)
139 index=0
140 # Let's add some information
141 for tempfile in temp_files:
142 f.write("# Disk%d was coming from %s\n" % (index,fio_data_file[index]))
143 f.write(tempfile.read())
144 f.write("\n")
145 tempfile.close()
146 index = index + 1
147 f.close()
148
149def average(s): return sum(s) * 1.0 / len(s)
150
151def compute_temp_file(fio_data_file,disk_perf,gnuplot_output_dir, min_time, max_time):
152 end_time=max_time
153 if end_time == -1:
154 end_time="infinite"
155 if verbose: print "Processing data file 1/2 with %s<time<%s" % (min_time,end_time)
156 files=[]
157 temp_outfile=[]
158 blk_size=0
159 for file in fio_data_file:
160 files.append(open(file))
161 pos = len(files) - 1
162 tmp_filename = "%sgnuplot_temp_file.%d" % (gnuplot_output_dir,pos)
163 temporary_files.append(tmp_filename)
164 gnuplot_file=open(tmp_filename,'w')
165 temp_outfile.append(gnuplot_file)
166 gnuplot_file.write("#Temporary file based on file %s\n" % file)
167 disk_perf.append([])
168
169 shall_break = False
170 while True:
171 current_line=[]
172 nb_empty_files=0
173 nb_files=len(files)
174 for myfile in files:
175 s=myfile.readline().replace(',',' ').split()
176 if not s:
177 nb_empty_files+=1
178 s="-1, 0, 0, 0".replace(',',' ').split()
179
180 if (nb_empty_files == nb_files):
181 shall_break=True
182 break;
183
184 current_line.append(s);
185
186 if shall_break == True:
187 break
188
189 last_time = -1
190 index=-1
191 perfs=[]
192 for line in enumerate(current_line):
193 # Index will be used to remember what file was featuring what value
194 index=index+1
195
196 time, perf, x, block_size = line[1]
197 if (blk_size == 0):
198 try:
199 blk_size=int(block_size)
200 except:
201 print "Error while reading the following line :"
202 print line
203 sys.exit(1);
204
205 # We ignore the first 500msec as it doesn't seems to be part of the real benchmark
206 # Time < 500 usually reports BW=0 breaking the min computing
207 if (min_time == 0):
208 min_time==0.5
209
210 # Then we estimate if the data we got is part of the time range we want to plot
211 if ((float(time)>(float(min_time)*1000)) and ((int(time) < (int(max_time)*1000)) or max_time==-1)):
212 disk_perf[index].append(int(perf))
213 perfs.append("%d %s %s"% (index, time, perf))
214
215 # If we reach this point, it means that all the traces are coherent
216 for p in enumerate(perfs):
217 index, perf_time,perf = p[1].split()
218 temp_outfile[int(index)].write("%s %.2f %s\n" % (index, float(float(perf_time)/1000), perf))
219
220
221 for file in files:
222 file.close()
223 for file in temp_outfile:
224 file.close()
225 return blk_size
226
227def compute_math(fio_data_file, title,gnuplot_output_filename,gnuplot_output_dir,mode,disk_perf,gpm_dir):
228 if verbose: print "Computing Maths"
229 global_min=[]
230 global_max=[]
231 average_file=open(gnuplot_output_dir+gnuplot_output_filename+'.average', 'w')
232 min_file=open(gnuplot_output_dir+gnuplot_output_filename+'.min', 'w')
233 max_file=open(gnuplot_output_dir+gnuplot_output_filename+'.max', 'w')
234 stddev_file=open(gnuplot_output_dir+gnuplot_output_filename+'.stddev', 'w')
235 global_file=open(gnuplot_output_dir+gnuplot_output_filename+'.global','w')
236 temporary_files.append(gnuplot_output_dir+gnuplot_output_filename+'.average')
237 temporary_files.append(gnuplot_output_dir+gnuplot_output_filename+'.min')
238 temporary_files.append(gnuplot_output_dir+gnuplot_output_filename+'.max')
239 temporary_files.append(gnuplot_output_dir+gnuplot_output_filename+'.stddev')
240 temporary_files.append(gnuplot_output_dir+gnuplot_output_filename+'.global')
241
242 min_file.write('DiskName %s\n' % mode)
243 max_file.write('DiskName %s\n'% mode)
244 average_file.write('DiskName %s\n'% mode)
245 stddev_file.write('DiskName %s\n'% mode )
246 for disk in xrange(len(fio_data_file)):
247# print disk_perf[disk]
248 min_file.write("# Disk%d was coming from %s\n" % (disk,fio_data_file[disk]))
249 max_file.write("# Disk%d was coming from %s\n" % (disk,fio_data_file[disk]))
250 average_file.write("# Disk%d was coming from %s\n" % (disk,fio_data_file[disk]))
251 stddev_file.write("# Disk%d was coming from %s\n" % (disk,fio_data_file[disk]))
252 avg = average(disk_perf[disk])
253 variance = map(lambda x: (x - avg)**2, disk_perf[disk])
254 standard_deviation = math.sqrt(average(variance))
255# print "Disk%d [ min=%.2f max=%.2f avg=%.2f stddev=%.2f \n" % (disk,min(disk_perf[disk]),max(disk_perf[disk]),avg, standard_deviation)
256 average_file.write('%d %d\n' % (disk, avg))
257 stddev_file.write('%d %d\n' % (disk, standard_deviation))
258 local_min=min(disk_perf[disk])
259 local_max=max(disk_perf[disk])
260 min_file.write('%d %d\n' % (disk, local_min))
261 max_file.write('%d %d\n' % (disk, local_max))
262 global_min.append(int(local_min))
263 global_max.append(int(local_max))
264
265 global_disk_perf = sum(disk_perf, [])
266 avg = average(global_disk_perf)
267 variance = map(lambda x: (x - avg)**2, global_disk_perf)
268 standard_deviation = math.sqrt(average(variance))
269
270 global_file.write('min=%.2f\n' % min(global_disk_perf))
271 global_file.write('max=%.2f\n' % max(global_disk_perf))
272 global_file.write('avg=%.2f\n' % avg)
273 global_file.write('stddev=%.2f\n' % standard_deviation)
274 global_file.write('values_count=%d\n' % len(global_disk_perf))
275 global_file.write('disks_count=%d\n' % len(fio_data_file))
276 #print "Global [ min=%.2f max=%.2f avg=%.2f stddev=%.2f \n" % (min(global_disk_perf),max(global_disk_perf),avg, standard_deviation)
277
278 average_file.close()
279 min_file.close()
280 max_file.close()
281 stddev_file.close()
282 global_file.close()
283 try:
284 os.remove(gnuplot_output_dir+'mymath')
285 except:
286 True
287
288 generate_gnuplot_math_script("Average values of "+title,gnuplot_output_filename+'.average',mode,int(avg),gnuplot_output_dir,gpm_dir)
289 generate_gnuplot_math_script("Min values of "+title,gnuplot_output_filename+'.min',mode,average(global_min),gnuplot_output_dir,gpm_dir)
290 generate_gnuplot_math_script("Max values of "+title,gnuplot_output_filename+'.max',mode,average(global_max),gnuplot_output_dir,gpm_dir)
291 generate_gnuplot_math_script("Standard Deviation of "+title,gnuplot_output_filename+'.stddev',mode,int(standard_deviation),gnuplot_output_dir,gpm_dir)
292
293def parse_global_files(fio_data_file, global_search):
294 max_result=0
295 max_file=''
296 for file in fio_data_file:
297 f=open(file)
298 disk_count=0
299 search_value=-1
300
301 # Let's read the complete file
302 while True:
303 try:
304 # We do split the name from the value
305 name,value=f.readline().split("=")
306 except:
307 f.close()
308 break
309 # If we ended the file
310 if not name:
311 # Let's process what we have
312 f.close()
313 break
314 else:
315 # disks_count is not global_search item
316 # As we need it for some computation, let's save it
317 if name=="disks_count":
318 disks_count=int(value)
319
320 # Let's catch the searched item
321 if global_search in name:
322 search_value=float(value)
323
324 # Let's process the avg value by estimated the global bandwidth per file
325 # We keep the biggest in memory for reporting
326 if global_search == "avg":
327 if (disks_count > 0) and (search_value != -1):
328 result=disks_count*search_value
329 if (result > max_result):
330 max_result=result
331 max_file=file
332 # Let's print the avg output
333 if global_search == "avg":
334 print "Biggest aggregated value of %s was %2.f in file %s\n" % (global_search, max_result, max_file)
335 else:
336 print "Global search %s is not yet implemented\n" % global_search
337
338def render_gnuplot(fio_data_file, gnuplot_output_dir):
339 print "Running gnuplot Rendering"
340 try:
341 # Let's render all the compared files if some
342 if len(fio_data_file) > 1:
343 if verbose: print " |-> Rendering comparing traces"
344 os.system("cd %s; for i in *.gnuplot; do gnuplot $i; done" % gnuplot_output_dir)
345 if verbose: print " |-> Rendering math traces"
346 os.system("cd %s; gnuplot mymath" % gnuplot_output_dir)
347 if verbose: print " |-> Rendering 2D & 3D traces"
348 os.system("cd %s; gnuplot mygraph" % gnuplot_output_dir)
349
350 name_of_directory="the current"
351 if gnuplot_output_dir != "./":
352 name_of_directory=gnuplot_output_dir
353 print "\nRendering traces are available in %s directory" % name_of_directory
354 global keep_temp_files
355 keep_temp_files=False
356 except:
357 print "Could not run gnuplot on mymath or mygraph !\n"
358 sys.exit(1);
359
360def print_help():
361 print 'fio2gnuplot -ghbiodvk -t <title> -o <outputfile> -p <pattern> -G <type> -m <time> -M <time>'
362 print
363 print '-h --help : Print this help'
364 print '-p <pattern> or --pattern <pattern> : A pattern in regexp to select fio input files'
365 print '-b or --bandwidth : A predefined pattern for selecting *_bw.log files'
366 print '-i or --iops : A predefined pattern for selecting *_iops.log files'
367 print '-g or --gnuplot : Render gnuplot traces before exiting'
368 print '-o or --outputfile <file> : The basename for gnuplot traces'
369 print ' - Basename is set with the pattern if defined'
370 print '-d or --outputdir <dir> : The directory where gnuplot shall render files'
371 print '-t or --title <title> : The title of the gnuplot traces'
372 print ' - Title is set with the block size detected in fio traces'
373 print '-G or --Global <type> : Search for <type> in .global files match by a pattern'
374 print ' - Available types are : min, max, avg, stddev'
375 print ' - The .global extension is added automatically to the pattern'
376 print '-m or --min_time <time> : Only consider data starting from <time> seconds (default is 0)'
377 print '-M or --max_time <time> : Only consider data ending before <time> seconds (default is -1 aka nolimit)'
378 print '-v or --verbose : Increasing verbosity'
379 print '-k or --keep : Keep all temporary files from gnuplot\'s output dir'
380
381def main(argv):
382 mode='unknown'
383 pattern=''
384 pattern_set_by_user=False
385 title='No title'
386 gnuplot_output_filename='result'
387 gnuplot_output_dir='./'
388 gpm_dir="/usr/share/fio/"
389 disk_perf=[]
390 run_gnuplot=False
391 parse_global=False
392 global_search=''
393 min_time=0
394 max_time=-1
395 global verbose
396 verbose=False
397 global temporary_files
398 temporary_files=[]
399 global keep_temp_files
400 keep_temp_files=True
401 force_keep_temp_files=False
402
403 if not os.path.isfile(gpm_dir+'math.gpm'):
404 gpm_dir="/usr/local/share/fio/"
405 if not os.path.isfile(gpm_dir+'math.gpm'):
406 print "Looks like fio didn't got installed properly as no gpm files found in '/usr/share/fio' or '/usr/local/share/fio'\n"
407 sys.exit(3)
408
409 try:
410 opts, args = getopt.getopt(argv[1:],"ghkbivo:d:t:p:G:m:M:",['bandwidth', 'iops', 'pattern', 'outputfile', 'outputdir', 'title', 'min_time', 'max_time', 'gnuplot', 'Global', 'help', 'verbose','keep'])
411 except getopt.GetoptError:
412 print "Error: One of the option passed to the cmdline was not supported"
413 print "Please fix your command line or read the help (-h option)"
414 sys.exit(2)
415
416 for opt, arg in opts:
417 if opt in ("-b", "--bandwidth"):
418 pattern='*_bw.log'
419 elif opt in ("-i", "--iops"):
420 pattern='*_iops.log'
421 elif opt in ("-v", "--verbose"):
422 verbose=True
423 elif opt in ("-k", "--keep"):
424 #User really wants to keep the temporary files
425 force_keep_temp_files=True
426 elif opt in ("-p", "--pattern"):
427 pattern_set_by_user=True
428 pattern=arg
429 pattern=pattern.replace('\\','')
430 elif opt in ("-o", "--outputfile"):
431 gnuplot_output_filename=arg
432 elif opt in ("-d", "--outputdir"):
433 gnuplot_output_dir=arg
434 if not gnuplot_output_dir.endswith('/'):
435 gnuplot_output_dir=gnuplot_output_dir+'/'
436 if not os.path.exists(gnuplot_output_dir):
437 os.makedirs(gnuplot_output_dir)
438 elif opt in ("-t", "--title"):
439 title=arg
440 elif opt in ("-m", "--min_time"):
441 min_time=arg
442 elif opt in ("-M", "--max_time"):
443 max_time=arg
444 elif opt in ("-g", "--gnuplot"):
445 run_gnuplot=True
446 elif opt in ("-G", "--Global"):
447 parse_global=True
448 global_search=arg
449 elif opt in ("-h", "--help"):
450 print_help()
451 sys.exit(1)
452
453 # Adding .global extension to the file
454 if parse_global==True:
455 if not gnuplot_output_filename.endswith('.global'):
456 pattern = pattern+'.global'
457
458 fio_data_file=find_file('.',pattern)
459 if len(fio_data_file) == 0:
460 print "No log file found with pattern %s!" % pattern
461 sys.exit(1)
462 else:
463 print "%d files Selected with pattern '%s'" % (len(fio_data_file), pattern)
464
465 fio_data_file=sorted(fio_data_file, key=str.lower)
466 for file in fio_data_file:
467 print ' |-> %s' % file
468 if "_bw.log" in file :
469 mode="Bandwidth (KB/sec)"
470 if "_iops.log" in file :
471 mode="IO per Seconds (IO/sec)"
472 if (title == 'No title') and (mode != 'unknown'):
473 if "Bandwidth" in mode:
474 title='Bandwidth benchmark with %d fio results' % len(fio_data_file)
475 if "IO" in mode:
476 title='IO benchmark with %d fio results' % len(fio_data_file)
477
478 print
479 #We need to adjust the output filename regarding the pattern required by the user
480 if (pattern_set_by_user == True):
481 gnuplot_output_filename=pattern
482 # As we do have some regexp in the pattern, let's make this simpliest
483 # We do remove the simpliest parts of the expression to get a clear file name
484 gnuplot_output_filename=gnuplot_output_filename.replace('-*-','-')
485 gnuplot_output_filename=gnuplot_output_filename.replace('*','-')
486 gnuplot_output_filename=gnuplot_output_filename.replace('--','-')
487 gnuplot_output_filename=gnuplot_output_filename.replace('.log','')
488 # Insure that we don't have any starting or trailing dash to the filename
489 gnuplot_output_filename = gnuplot_output_filename[:-1] if gnuplot_output_filename.endswith('-') else gnuplot_output_filename
490 gnuplot_output_filename = gnuplot_output_filename[1:] if gnuplot_output_filename.startswith('-') else gnuplot_output_filename
491
492 if parse_global==True:
493 parse_global_files(fio_data_file, global_search)
494 else:
495 blk_size=compute_temp_file(fio_data_file,disk_perf,gnuplot_output_dir,min_time,max_time)
496 title="%s @ Blocksize = %dK" % (title,blk_size/1024)
497 compute_aggregated_file(fio_data_file, gnuplot_output_filename, gnuplot_output_dir)
498 compute_math(fio_data_file,title,gnuplot_output_filename,gnuplot_output_dir,mode,disk_perf,gpm_dir)
499 generate_gnuplot_script(fio_data_file,title,gnuplot_output_filename,gnuplot_output_dir,mode,disk_perf,gpm_dir)
500
501 if (run_gnuplot==True):
502 render_gnuplot(fio_data_file, gnuplot_output_dir)
503
504 # Shall we clean the temporary files ?
505 if keep_temp_files==False and force_keep_temp_files==False:
506 # Cleaning temporary files
507 if verbose: print "Cleaning temporary files"
508 for f in enumerate(temporary_files):
509 if verbose: print " -> %s"%f[1]
510 try:
511 os.remove(f[1])
512 except:
513 True
514
515#Main
516if __name__ == "__main__":
517 sys.exit(main(sys.argv))