summaryrefslogtreecommitdiff
path: root/util/stats/profile.py
blob: cf946452bd9b82f9b17cdc8f1597a28d14b44667 (plain)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
# Copyright (c) 2005 The Regents of The University of Michigan
# All rights reserved.
#
# Redistribution and use in source and binary forms, with or without
# modification, are permitted provided that the following conditions are
# met: redistributions of source code must retain the above copyright
# notice, this list of conditions and the following disclaimer;
# redistributions in binary form must reproduce the above copyright
# notice, this list of conditions and the following disclaimer in the
# documentation and/or other materials provided with the distribution;
# neither the name of the copyright holders nor the names of its
# contributors may be used to endorse or promote products derived from
# this software without specific prior written permission.
#
# THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
# "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
# LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
# A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
# OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
# SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
# LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
# DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
# THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
# (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
# OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.

from orderdict import orderdict
import output

class RunData(dict):
    def __init__(self, filename=None):
        self.filename = filename

    def __getattr__(self, attr):
        if attr == 'total':
            total = 0.0
            for value in self.itervalues():
                total += value
            return total
        if attr == 'maxsymlen':
            return max([ len(sym) for sym in self.iterkeys() ])

    def display(self, output=None, limit=None, maxsymlen=None):
        if not output:
            import sys
            output = sys.stdout
        elif isinstance(output, str):
            output = file(output, 'w')

        total = float(self.total)

        # swap (string,count) order so we can sort on count
        symbols = [ (count,name) for name,count in self.iteritems() ]
        symbols.sort(reverse=True)
        if limit is not None:
            symbols = symbols[:limit]

        if not maxsymlen:
            maxsymlen = self.maxsymlen

        symbolf = "%-" + str(maxsymlen + 1) + "s %.2f%%"
        for number,name in symbols:
            print >>output, symbolf % (name, 100.0 * (float(number) / total))



class PCData(RunData):
    def __init__(self, filename=None, categorize=None, showidle=True):
        super(PCData, self).__init__(self, filename)
        if filename is None:
            return

        fd = file(filename)

        for line in fd:
            if line.strip() == '>>>PC data':
                break

        for line in fd:
            if line.startswith('>>>'):
                break

            (symbol, count) = line.split()
            if symbol == "0x0":
                continue
            count = int(count)

            if categorize is not None:
                category = categorize(symbol)
                if category is None:
                    category = 'other'
                elif category == 'idle' and not showidle:
                    continue

                self[category] = count

        fd.close()

class FuncNode(object):
    def __new__(cls, filename = None):
        if filename is None:
            return super(FuncNode, cls).__new__(cls)

        fd = file(filename, 'r')
        fditer = iter(fd)
        nodes = {}
        for line in fditer:
            if line.strip() == '>>>function data':
                break

        for line in fditer:
            if line.startswith('>>>'):
                break

            data = line.split()
            node_id = int(data[0], 16)
            node = FuncNode()
            node.symbol = data[1]
            node.count = int(data[2])
            node.children = [ int(child, 16) for child in data[3:] ]
            nodes[node_id] = node

        for node in nodes.itervalues():
            children = []
            for cid in node.children:
                child = nodes[cid]
                children.append(child)
                child.parent = node
            node.children = tuple(children)
        if not nodes:
            print filename
            print nodes
        return nodes[0]

    def __init__(self, filename=None):
        pass

    def total(self):
        total = self.count
        for child in self.children:
            total += child.total()

        return total

    def aggregate(self, dict, categorize, incategory):
        category = None
        if categorize:
            category = categorize(self.symbol)

        total = self.count
        for child in self.children:
            total += child.aggregate(dict, categorize, category or incategory)

        if category:
            dict[category] = dict.get(category, 0) + total
            return 0
        elif not incategory:
            dict[self.symbol] = dict.get(self.symbol, 0) + total

        return total

    def dump(self):
        kids = [ child.symbol for child in self.children]
        print '%s %d <%s>' % (self.symbol, self.count, ', '.join(kids))
        for child in self.children:
            child.dump()

    def _dot(self, dot, threshold, categorize, total):
        from pydot import Dot, Edge, Node
        self.dot_node = None

        value = self.total() * 100.0 / total
        if value < threshold:
            return
        if categorize:
            category = categorize(self.symbol)
            if category and category != 'other':
                return
        label = '%s %.2f%%' % (self.symbol, value)
        self.dot_node = Node(self, label=label)
        dot.add_node(self.dot_node)

        for child in self.children:
            child._dot(dot, threshold, categorize, total)
            if child.dot_node is not None:
                dot.add_edge(Edge(self, child))

    def _cleandot(self):
        for child in self.children:
            child._cleandot()
            self.dot_node = None
            del self.__dict__['dot_node']

    def dot(self, dot, threshold=0.1, categorize=None):
        self._dot(dot, threshold, categorize, self.total())
        self._cleandot()

class FuncData(RunData):
    def __init__(self, filename, categorize=None):
        super(FuncData, self).__init__(filename)
        self.tree = FuncNode(filename)
        self.tree.aggregate(self, categorize, incategory=False)
        self.total = self.tree.total()

    def displayx(self, output=None, maxcount=None):
        if output is None:
            import sys
            output = sys.stdout

        items = [ (val,key) for key,val in self.iteritems() ]
        items.sort(reverse=True)
        for val,key in items:
            if maxcount is not None:
                if maxcount == 0:
                    return
                maxcount -= 1

            percent = val * 100.0 / self.total
            print >>output, '%-30s %8s' % (key, '%3.2f%%' % percent)

class Profile(object):
    # This list controls the order of values in stacked bar data output
    default_categories = [ 'interrupt',
                           'driver',
                           'stack',
                           'buffer',
                           'copy',
                           'syscall',
                           'user',
                           'other',
                           'idle']

    def __init__(self, datatype, categorize=None):
        categories = Profile.default_categories

        self.datatype = datatype
        self.categorize = categorize
        self.data = {}
        self.categories = categories[:]
        self.rcategories = categories[:]
        self.rcategories.reverse()
        self.cpu = 0

    # Read in files
    def inputdir(self, directory):
        import os, os.path, re
        from os.path import expanduser, join as joinpath

        directory = expanduser(directory)
        label_ex = re.compile(r'profile\.(.*).dat')
        for root,dirs,files in os.walk(directory):
            for name in files:
                match = label_ex.match(name)
                if not match:
                    continue

                filename = joinpath(root, name)
                prefix = os.path.commonprefix([root, directory])
                dirname = root[len(prefix)+1:]
                data = self.datatype(filename, self.categorize)
                self.setdata(dirname, match.group(1), data)

    def setdata(self, run, cpu, data):
        if run not in self.data:
            self.data[run] = {}

        if cpu in self.data[run]:
            raise AttributeError, \
                  'data already stored for run %s and cpu %s' % (run, cpu)

        self.data[run][cpu] = data

    def getdata(self, run, cpu):
        try:
            return self.data[run][cpu]
        except KeyError:
            return None

    def alldata(self):
        for run,cpus in self.data.iteritems():
            for cpu,data in cpus.iteritems():
                yield run,cpu,data

    def get(self, job, stat):
        if job.system is None:
            raise AttributeError, 'The job must have a system set'

        data = self.getdata(job.name, '%s.full%d' % (job.system, self.cpu))
        if not data:
            return [ 0.0 for c in self.categories ]

        values = []
        for category in self.categories:
            values.append(data.get(category, 0.0))
        return values

    def dump(self):
        for run,cpu,data in self.alldata():
            print 'run %s, cpu %s' % (run, cpu)
            data.dump()
            print

    def write_dot(self, threshold, jobfile=None, jobs=None):
        import pydot

        if jobs is None:
            jobs = [ job for job in jobfile.jobs() ]

        for job in jobs:
            cpu =  '%s.full%d' % (job.system, self.cpu)
            symbols = self.getdata(job.name, cpu)
            if not symbols:
                continue

            dot = pydot.Dot()
            symbols.tree.dot(dot, threshold=threshold)
            dot.write(symbols.filename[:-3] + 'dot')

    def write_txt(self, jobfile=None, jobs=None):
        if jobs is None:
            jobs = [ job for job in jobfile.jobs() ]

        for job in jobs:
            cpu =  '%s.full%d' % (job.system, self.cpu)
            symbols = self.getdata(job.name, cpu)
            if not symbols:
                continue

            output = file(symbols.filename[:-3] + 'txt', 'w')
            symbols.display(output)

    def display(self, jobfile=None, jobs=None, limit=None):
        if jobs is None:
            jobs = [ job for job in jobfile.jobs() ]

        maxsymlen = 0

        thejobs = []
        for job in jobs:
            cpu =  '%s.full%d' % (job.system, self.cpu)
            symbols = self.getdata(job.name, cpu)
            if symbols:
                thejobs.append(job)
                maxsymlen = max(maxsymlen, symbols.maxsymlen)

        for job in thejobs:
            cpu =  '%s.full%d' % (job.system, self.cpu)
            symbols = self.getdata(job.name, cpu)
            print job.name
            symbols.display(limit=limit, maxsymlen=maxsymlen)
            print


from categories import func_categorize, pc_categorize
class PCProfile(Profile):
    def __init__(self, categorize=pc_categorize):
        super(PCProfile, self).__init__(PCData, categorize)


class FuncProfile(Profile):
    def __init__(self, categorize=func_categorize):
        super(FuncProfile, self).__init__(FuncData, categorize)

def usage(exitcode = None):
    print '''\
Usage: %s [-bc] [-g <dir>] [-j <jobfile>] [-n <num>]

    -c           groups symbols into categories
    -b           dumps data for bar charts
    -d           generate dot output
    -g <d>       draw graphs and send output to <d>
    -j <jobfile> specify a different jobfile (default is Test.py)
    -n <n>       selects number of top symbols to print (default 5)
''' % sys.argv[0]

    if exitcode is not None:
        sys.exit(exitcode)

if __name__ == '__main__':
    import getopt, re, sys
    from os.path import expanduser
    from output import StatOutput
    from jobfile import JobFile

    # default option values
    numsyms = 10
    graph = None
    cpus = [ 0 ]
    categorize = False
    showidle = True
    funcdata = True
    jobfilename = 'Test.py'
    dodot = False
    dotformat = 'raw'
    textout = False
    threshold = 0.01
    inputfile = None

    try:
        opts, args = getopt.getopt(sys.argv[1:], 'C:cdD:f:g:ij:n:pT:t')
    except getopt.GetoptError:
        usage(2)

    for o,a in opts:
        if o == '-C':
            cpus = [ int(x) for x in a.split(',') ]
        elif o == '-c':
            categorize = True
        elif o == '-D':
            dotformat = a
        elif o == '-d':
            dodot = True
        elif o == '-f':
            inputfile = expanduser(a)
        elif o == '-g':
            graph = a
        elif o == '-i':
            showidle = False
        elif o == '-j':
            jobfilename = a
        elif o == '-n':
            numsyms = int(a)
        elif o == '-p':
            funcdata = False
        elif o == '-T':
            threshold = float(a)
        elif o == '-t':
            textout = True

    if args:
        print "'%s'" % args, len(args)
        usage(1)

    if inputfile:
        data = FuncData(inputfile)

        if dodot:
            import pydot
            dot = pydot.Dot()
            data.dot(dot, threshold=threshold)
            #dot.orientation = 'landscape'
            #dot.ranksep='equally'
            #dot.rank='samerank'
            dot.write(dotfile, format=dotformat)
        else:
            data.display(limit=numsyms)

    else:
        jobfile = JobFile(jobfilename)

        if funcdata:
            profile = FuncProfile()
        else:
            profile = PCProfile()

        profile.inputdir(jobfile.rootdir)

        if graph:
            for cpu in cpus:
                profile.cpu = cpu
                if funcdata:
                    name = 'funcstacks%d' % cpu
                else:
                    name = 'stacks%d' % cpu
                output = StatOutput(name, jobfile, info=profile)
                output.graph(graph)

        if dodot:
            for cpu in cpus:
                profile.cpu = cpu
                profile.write_dot(jobfile=jobfile, threshold=threshold)

        if not categorize:
            for cpu in cpus:
                profile.cpu = cpu
                profile.categorize = None

        if textout:
            for cpu in cpus:
                profile.cpu = cpu
                profile.write_txt(jobfile=jobfile)

        if not graph and not textout and not dodot:
            for cpu in cpus:
                profile.cpu = cpu
                profile.display(jobfile=jobfile, limit=numsyms)