diff options
author | Cristian Cadar <cristic@cs.stanford.edu> | 2012-10-09 14:23:42 +0000 |
---|---|---|
committer | Cristian Cadar <cristic@cs.stanford.edu> | 2012-10-09 14:23:42 +0000 |
commit | a1d4a739609e2a491e846c765c05ddd0b9c79935 (patch) | |
tree | 115a5e469d1a9d8a3cd125c91855eb416d5480b1 | |
parent | a21b6c7cd1395cc50efd89f7dfa582988fd705c6 (diff) | |
download | klee-a1d4a739609e2a491e846c765c05ddd0b9c79935.tar.gz |
Patch by Tomasz Kuchta that adds several useful options (--print-abs-times, --print-rel-times, --precision) to klee-stats.
git-svn-id: https://llvm.org/svn/llvm-project/klee/trunk@165499 91177308-0d34-0410-b5e6-96231b3b80d8
-rwxr-xr-x | tools/klee-stats/klee-stats | 90 |
1 files changed, 68 insertions, 22 deletions
diff --git a/tools/klee-stats/klee-stats b/tools/klee-stats/klee-stats index c98f7000..f20ff14f 100755 --- a/tools/klee-stats/klee-stats +++ b/tools/klee-stats/klee-stats @@ -105,12 +105,17 @@ def sortTable(table, labels, keyName, ascending=False): table[:] = [table[n] for n in indices] -def printTable(table): +def printTable(table, precision): def strOrNone(ob): if ob is None: return '' elif isinstance(ob,float): - return '%.2f'%ob + # prepare format string according to settings of + # floating number + formatString = '%.' + formatString += '%d'%precision + formatString += 'f' + return formatString%ob else: return str(ob) def printRow(row): @@ -139,22 +144,33 @@ def printTable(table): def main(args): from optparse import OptionParser - op = OptionParser(usage="usage: %prog [options] directories", - epilog="""\ + + # inherit from the OptionParser class and override format_epilog + # in order to get newlines in the 'epilog' text + class MyParser(OptionParser): + def format_epilog(self, formatter): + return self.epilog + + op = MyParser(usage="usage: %prog [options] directories", + epilog="""\ + LEGEND ------ -Instrs: Number of executed instructions -Time: Total wall time (s) -ICov: Instruction coverage in the LLVM bitcode (%) -BCov: Branch coverage in the LLVM bitcode (%) -ICount: Total static instructions in the LLVM bitcode -Solver: Time spent in the constraint solver (%) -States: Number of currently active states -Mem: Megabytes of memory currently used -Queries: Number of queries issued to STP -AvgQC: Average number of query constructs per query -Tcex: Time spent in the counterexample caching code (%) -Tfork: Time spent forking (%)""") +Instrs: Number of executed instructions +Time: Total wall time (s) +TUser: Total user time +ICov: Instruction coverage in the LLVM bitcode (%) +BCov: Branch coverage in the LLVM bitcode (%) +ICount: Total static instructions in the LLVM bitcode +TSolver: Time spent in the constraint solver +States: Number of currently active states +Mem: Megabytes of memory currently used +Queries: Number of queries issued to STP +AvgQC: Average number of query constructs per query +Tcex: Time spent in the counterexample caching code +Tfork: Time spent forking +TResolve: Time spent in object resolution +\n""") op.add_option('', '--print-more', dest='printMore', action='store_true', default=False, @@ -162,6 +178,12 @@ Tfork: Time spent forking (%)""") op.add_option('', '--print-all', dest='printAll', action='store_true', default=False, help='Print all available information.') + op.add_option('', '--print-rel-times', dest='printRelTimes', + action='store_true', default=False, + help='Print only values of measured times. Values are relative to the measured system execution time.') + op.add_option('', '--print-abs-times', dest='printAbsTimes', + action='store_true', default=False, + help='Print only values of measured times. Absolute values (in seconds) are printed.') op.add_option('','--sort-by', dest='sortBy', help='key value to sort by, e.g. --sort-by=Instrs') op.add_option('','--ascending', dest='ascending', @@ -171,11 +193,26 @@ Tfork: Time spent forking (%)""") help="key value on which to compare runs to the reference one (which is the first one). E.g., --compare-by=Instrs shows how each run compares to the reference run after executing the same number of instructions as the reference run. If a run hasn't executed as many instructions as the reference one, we simply print the statistics at the end of that run.") op.add_option('','--compare-at', dest='compAt', help='value to compare the runs at. Can be special value \'last\' to compare at the last point which makes sense. Use in conjunction with --compare-by') + op.add_option('','--precision', dest='dispPrecision', default=2, metavar='INTEGER', + help='Floating point numbers display precision. By default it is set to 2.') opts,dirs = op.parse_args() if not dirs: op.error("no directories given.") + # make sure that precision is a positive integer + if not isinstance(opts.dispPrecision, int): + try: + precision = int(opts.dispPrecision) + except ValueError: + op.error("display precision should be an integer") + # make sure that the precision is a positive integer as well + if not (0 <= precision): + op.error("display precision should be positive") + else: + precision = opts.dispPrecision + + actualDirs = [] for dir in dirs: if os.path.exists(os.path.join(dir,'info')): @@ -192,15 +229,19 @@ Tfork: Time spent forking (%)""") ssummary = [] if (opts.printAll): - labels = ('Path','Instrs','Time(s)','ICov(%)','BCov(%)','ICount','Solver(%)', 'States', 'maxStates', 'avgStates', 'Mem(MB)', 'maxMem(MB)', 'avgMem(MB)', 'Queries', 'AvgQC', 'Tcex(%)', 'Tfork(%)') + labels = ('Path','Instrs','Time(s)','ICov(%)','BCov(%)','ICount','TSolver(%)', 'States', 'maxStates', 'avgStates', 'Mem(MB)', 'maxMem(MB)', 'avgMem(MB)', 'Queries', 'AvgQC', 'Tcex(%)', 'Tfork(%)') + elif (opts.printRelTimes): + labels = ('Path', 'Time(s)', 'TUser(%)', 'TSolver(%)', 'Tcex(%)', 'Tfork(%)', 'TResolve(%)') + elif (opts.printAbsTimes): + labels = ('Path', 'Time(s)', 'TUser(s)', 'TSolver(s)', 'Tcex(s)', 'Tfork(s)', 'TResolve(s)') elif (opts.printMore): - labels = ('Path','Instrs','Time(s)','ICov(%)','BCov(%)','ICount','Solver(%)', 'States', 'maxStates', 'Mem(MB)', 'maxMem(MB)') + labels = ('Path','Instrs','Time(s)','ICov(%)','BCov(%)','ICount','TSolver(%)', 'States', 'maxStates', 'Mem(MB)', 'maxMem(MB)') else: - labels = ('Path','Instrs','Time(s)','ICov(%)','BCov(%)','ICount','Solver(%)') + labels = ('Path','Instrs','Time(s)','ICov(%)','BCov(%)','ICount','TSolver(%)') def addRecord(Path,rec, stats): - (I,BFull,BPart,BTot,T,St,Mem,QTot,QCon,NObjs,Treal,SCov,SUnc,QT,Ts,Tcex,Tf) = rec + (I,BFull,BPart,BTot,T,St,Mem,QTot,QCon,NObjs,Treal,SCov,SUnc,QT,Ts,Tcex,Tf,Tr) = rec (maxMem,avgMem,maxStates,avgStates) = stats # special case for straight-line code: report 100% branch coverage if BTot == 0: @@ -211,6 +252,10 @@ Tfork: Time spent forking (%)""") if (opts.printAll): table.append((Path, I, Treal, 100.*SCov/(SCov+SUnc), 100.*(2*BFull+BPart)/(2.*BTot), SCov+SUnc, 100.*Ts/Treal, St, maxStates, avgStates, Mem, maxMem, avgMem, QTot, AvgQC, 100.*Tcex/Treal, 100.*Tf/Treal)) + elif (opts.printRelTimes): + table.append((Path, Treal, 100.*T/Treal, 100.*Ts/Treal, 100.*Tcex/Treal, 100.*Tf/Treal, 100.*Tr/Treal)) + elif (opts.printAbsTimes): + table.append((Path, Treal, T, Ts, Tcex, Tf, Tr)) elif (opts.printMore): table.append((Path, I, Treal, 100.*SCov/(SCov+SUnc), 100.*(2*BFull+BPart)/(2.*BTot), SCov+SUnc, 100.*Ts/Treal, St, maxStates, Mem, maxMem)) @@ -219,7 +264,7 @@ Tfork: Time spent forking (%)""") SCov+SUnc, 100.*Ts/Treal)) def addRow(Path,data,stats): - data = tuple(data[:17]) + (None,)*(17-len(data)) + data = tuple(data[:18]) + (None,)*(18-len(data)) addRecord(Path,data,stats) if not summary: summary[:] = list(data) @@ -268,7 +313,8 @@ Tfork: Time spent forking (%)""") addRecord('Total (%d)'%(len(table)-1,),summary, ssummary) table[0:0] = [None,labels,None] table.append(None) - printTable(table) + + printTable(table, precision) if __name__=='__main__': |