Remove all binning stuff

--HG--
extra : convert_revision : 6da2b3b0b6c2824f2064d42670fd8383edb7c718

5 files changed, 22 insertions, 171 deletions

diff --git a/util/stats/db.py b/util/stats/db.py
index c0e7796eb..9d876e97f 100644
--- a/util/stats/db.py
+++ b/util/stats/db.py
@@ -135,10 +135,6 @@ class Database(object):
         self.allRunIds = {}
         self.allRunNames = {}
 
-        self.allBins = []
-        self.allBinIds = {}
-        self.allBinNames = {}
-
         self.allFormulas = {}
 
         self.stattop = {}
@@ -147,7 +143,6 @@ class Database(object):
 
         self.mode = 'sum';
         self.runs = None
-        self.bins = None
         self.ticks = None
         self.method = 'sum'
         self._method = type(self).sum
@@ -218,11 +213,6 @@ class Database(object):
             self.allRunIds[run.run] = run
             self.allRunNames[run.name] = run
 
-        self.query('select * from bins')
-        for id,name in self.cursor.fetchall():
-            self.allBinIds[int(id)] = name
-            self.allBinNames[name] = int(id)
-
         self.query('select sd_stat,sd_x,sd_y,sd_name,sd_descr from subdata')
         for result in self.cursor.fetchall():
             subdata = SubData(result)
@@ -245,18 +235,6 @@ class Database(object):
             self.allStatIds[stat.stat] = stat
             self.allStatNames[stat.name] = stat
 
-    # Name: listbins
-    # Desc: Prints all bins matching regex argument, if no argument
-    #       is given all bins are returned
-    def listBins(self, regex='.*'):
-        print '%-50s %-10s' % ('bin name', 'id')
-        print '-' * 61
-        names = self.allBinNames.keys()
-        names.sort()
-        for name in names:
-            id = self.allBinNames[name]
-            print '%-50s %-10d' % (name, id)
-
     # Name: listruns
     # Desc: Prints all runs matching a given user, if no argument
     #       is given all runs are returned
@@ -360,39 +338,10 @@ class Database(object):
                         ret.append(stat)
         return ret
 
-    def getBin(self, bins):
-        if type(bins) is not list:
-            bins = [ bins ]
-
-        ret = []
-        for bin in bins:
-            if type(bin) is int:
-                ret.append(bin)
-            elif type(bin) is str:
-                ret.append(self.allBinNames[bin])
-            else:
-                for name,id in self.allBinNames.items():
-                    if bin.match(name):
-                        ret.append(id)
-
-        return ret
-
-    def getNotBin(self, bin):
-        map = {}
-        for bin in getBin(bin):
-            map[bin] = 1
-
-        ret = []
-        for bin in self.allBinIds.keys():
-            if not map.has_key(bin):
-                ret.append(bin)
-
-        return ret
-
     #########################################
     # get the data
     #
-    def inner(self, op, stat, bins, ticks, group=False):
+    def query(self, op, stat, ticks, group=False):
         sql = 'select '
         sql += 'dt_stat as stat, '
         sql += 'dt_run as run, '
@@ -414,10 +363,6 @@ class Database(object):
             val = ' or '.join([ 'dt_run=%d' % r for r in self.runs ])
             sql += ' and (%s)' % val
 
-        if bins != None and len(bins):
-            val = ' or '.join([ 'dt_bin=%d' % b for b in bins ])
-            sql += ' and (%s)' % val
-
         if ticks != None and len(ticks):
             val = ' or '.join([ 'dt_tick=%d' % s for s in ticks ])
             sql += ' and (%s)' % val
@@ -427,35 +372,21 @@ class Database(object):
             sql += ',dt_tick'
         return sql
 
-    def outer(self, op_out, op_in, stat, bins, ticks):
-        sql = self.inner(op_in, stat, bins, ticks, True)
-        sql = 'select stat,run,x,y,%s(data) from (%s) as tb ' % (op_out, sql)
-        sql += 'group by stat,run,x,y'
-        return sql
-
     # Name: sum
-    # Desc: given a run, a stat and an array of samples and bins,
-    #        sum all the bins and then get the standard deviation of the
-    #        samples for non-binned runs. This will just return the average
-    #        of samples, however a bin array still must be passed
-    def sum(self, stat, bins, ticks):
-        return self.inner('sum', stat, bins, ticks)
+    # Desc: given a run, a stat and an array of samples, total the samples
+    def sum(self, *args, **kwargs):
+        return self.query('sum', *args, **kwargs)
 
     # Name: avg
-    # Desc: given a run, a stat and an array of samples and bins,
-    #        sum all the bins and then average the samples for non-binned
-    #        runs this will just return the average of samples, however
-    #        a bin array still must be passed
-    def avg(self, stat, bins, ticks):
-        return self.outer('avg', 'sum', stat, bins, ticks)
+    # Desc: given a run, a stat and an array of samples, average the samples
+    def avg(self, stat, ticks):
+        return self.query('avg', *args, **kwargs)
 
     # Name: stdev
-    # Desc: given a run, a stat and an array of samples and bins,
-    #        sum all the bins and then get the standard deviation of the
-    #        samples for non-binned runs. This will just return the average
-    #        of samples, however a bin array still must be passed
-    def stdev(self, stat, bins, ticks):
-        return self.outer('stddev', 'sum', stat, bins, ticks)
+    # Desc: given a run, a stat and an array of samples, get the standard
+    #       deviation
+    def stdev(self, stat, ticks):
+        return self.query('stddev', *args, **kwargs)
 
     def __setattr__(self, attr, value):
         super(Database, self).__setattr__(attr, value)
@@ -471,12 +402,10 @@ class Database(object):
         else:
             raise AttributeError, "can only set get to: sum | avg | stdev"
 
-    def data(self, stat, bins=None, ticks=None):
-        if bins is None:
-            bins = self.bins
+    def data(self, stat, ticks=None):
         if ticks is None:
             ticks = self.ticks
-        sql = self._method(self, stat, bins, ticks)
+        sql = self._method(self, stat, ticks)
         self.query(sql)
 
         runs = {}
diff --git a/util/stats/dbinit.py b/util/stats/dbinit.py
index ffae88cdb..203582478 100644
--- a/util/stats/dbinit.py
+++ b/util/stats/dbinit.py
@@ -98,28 +98,6 @@ class MyDB(object):
         ) TYPE=InnoDB''')
 
         #
-        # We keep the bin names separate so that the data table doesn't get
-        # huge since bin names are frequently repeated.
-        #
-        # COLUMNS:
-        #   'id' is the unique bin identifer.
-        #   'name' is the string name for the bin.
-        #
-        # INDEXES:
-        #   'bin' is indexed to get the name of a bin when data is retrieved
-        #       via the data table.
-        #   'name' is indexed to get the bin id for a named bin when you want
-        #       to search the data table based on a specific bin.
-        #
-        self.query('''
-        CREATE TABLE bins(
-            bn_id	SMALLINT UNSIGNED	NOT NULL AUTO_INCREMENT,
-            bn_name	VARCHAR(255)		NOT NULL,
-            PRIMARY KEY(bn_id),
-            UNIQUE (bn_name)
-        ) TYPE=InnoDB''')
-
-        #
         # The stat table gives us all of the data for a particular stat.
         #
         # COLUMNS:
@@ -220,14 +198,12 @@ class MyDB(object):
         #   'run' is the run that the data was generated from.  Details up in
         #       the run table
         #   'tick' is a timestamp generated by the simulator.
-        #   'bin' is the name of the bin that the data was generated in, if
-        #       any.
         #   'data' is the actual stat value.
         #
         # INDEXES:
         #   'stat' is indexed so that a user can find all of the data for a
         #       particular stat. It is not unique, because that specific stat
-        #       can be found in many runs, bins, and samples, in addition to
+        #       can be found in many runs and samples, in addition to
         #       having entries for the mulidimensional cases.
         #   'run' is indexed to allow a user to remove all of the data for a
         #       particular execution run.  It can also be used to allow the
@@ -240,11 +216,10 @@ class MyDB(object):
             dt_y	SMALLINT		NOT NULL,
             dt_run	SMALLINT UNSIGNED	NOT NULL,
             dt_tick	BIGINT UNSIGNED		NOT NULL,
-            dt_bin	SMALLINT UNSIGNED	NOT NULL,
             dt_data	DOUBLE			NOT NULL,
             INDEX (dt_stat),
             INDEX (dt_run),
-            UNIQUE (dt_stat,dt_x,dt_y,dt_run,dt_tick,dt_bin)
+            UNIQUE (dt_stat,dt_x,dt_y,dt_run,dt_tick)
         ) TYPE=InnoDB;''')
 
         #
@@ -396,12 +371,6 @@ class MyDB(object):
         WHERE dt_stat IS NULL''')
 
         self.query('''
-        DELETE bins
-        FROM bins
-        LEFT JOIN data ON bn_id=dt_bin
-        WHERE dt_bin IS NULL''')
-
-        self.query('''
         DELETE events
         FROM events
         LEFT JOIN runs ON ev_run=rn_id
diff --git a/util/stats/info.py b/util/stats/info.py
index 4cb55f564..c1e18bb3f 100644
--- a/util/stats/info.py
+++ b/util/stats/info.py
@@ -174,7 +174,7 @@ def WrapValue(value):
 class Statistic(object):
     def __getattr__(self, attr):
         if attr in ('data', 'x', 'y'):
-            result = self.source.data(self, self.bins, self.ticks)
+            result = self.source.data(self, self.ticks)
             self.data = result.data
             self.x = result.x
             self.y = result.y
@@ -183,7 +183,7 @@ class Statistic(object):
     def __setattr__(self, attr, value):
         if attr == 'stat':
             raise AttributeError, '%s is read only' % stat
-        if attr in ('source', 'bins', 'ticks'):
+        if attr in ('source', 'ticks'):
             if getattr(self, attr) != value:
                 if hasattr(self, 'data'):
                     delattr(self, 'data')
@@ -759,7 +759,6 @@ def NewStat(source, data):
         stat = Formula()
 
     stat.__dict__['source'] = source
-    stat.__dict__['bins'] = None
     stat.__dict__['ticks'] = None
     stat.__dict__.update(data.__dict__)
 
diff --git a/util/stats/output.py b/util/stats/output.py
index abfb8d901..03c100edc 100644
--- a/util/stats/output.py
+++ b/util/stats/output.py
@@ -29,24 +29,16 @@
 from chart import ChartOptions
 
 class StatOutput(ChartOptions):
-    def __init__(self, jobfile, info, stat=None, binstats=None):
+    def __init__(self, jobfile, info, stat=None):
         super(StatOutput, self).__init__()
         self.jobfile = jobfile
         self.stat = stat
-        self.binstats = None
         self.invert = False
         self.info = info
 
-    def printdata(self, name, bin = None, printmode = 'G'):
+    def display(self, name, printmode = 'G'):
         import info
 
-        if bin:
-            print '%s %s stats' % (name, bin)
-
-        if self.binstats:
-            for stat in self.binstats:
-                stat.bins = bin
-
         if printmode == 'G':
             valformat = '%g'
         elif printmode != 'F' and value > 1e6:
@@ -70,16 +62,6 @@ class StatOutput(ChartOptions):
             valstring = ', '.join([ valformat % val for val in value ])
             print '%-50s    %s' % (job.name + ':', valstring)
 
-    def display(self, name, binned = False, printmode = 'G'):
-        if binned and self.binstats:
-            self.printdata(name, 'kernel', printmode)
-            self.printdata(name, 'idle', printmode)
-            self.printdata(name, 'user', printmode)
-            self.printdata(name, 'interrupt', printmode)
-
-            print '%s total stats' % name
-        self.printdata(name, printmode=printmode)
-
     def graph(self, name, graphdir, proxy=None):
         from os.path import expanduser, isdir, join as joinpath
         from barchart import BarChart
diff --git a/util/stats/stats.py b/util/stats/stats.py
index 08281287f..40ef8ac0f 100755
--- a/util/stats/stats.py
+++ b/util/stats/stats.py
@@ -37,7 +37,6 @@ Usage: %s [-E] [-F] [ -G <get> ] [-d <db> ] [-g <graphdir> ] [-h <host>] [-p]
 
        commands    extra parameters   description
        ----------- ------------------ ---------------------------------------
-       bins        [regex]            List bins (only matching regex)
        formula     <formula>          Evaluated formula specified
        formulas    [regex]            List formulas (only matching regex)
        runs        none               List all runs in database
@@ -140,16 +139,6 @@ def commands(options, command, args):
 
         return
 
-    if command == 'bins':
-        if len(args) == 0:
-            source.listBins()
-        elif len(args) == 1:
-            source.listBins(args[0])
-        else:
-            raise CommandException
-
-        return
-
     if command == 'formulas':
         if len(args) == 0:
             source.listFormulas()
@@ -279,7 +268,7 @@ def commands(options, command, args):
             if options.graph:
                 output.graph(stat.name, options.graphdir)
             else:
-                output.display(stat.name, options.binned, options.printmode)
+                output.display(stat.name, options.printmode)
 
         return
 
@@ -299,22 +288,10 @@ def commands(options, command, args):
         if options.graph:
             output.graph(command, options.graphdir, proxy)
         else:
-            output.display(command, options.binned, options.printmode)
-
-    if command == 'usertime':
-        import copy
-        user = copy.copy(system.run0.numCycles)
-        user.bins = 'user'
-
-        output.stat = user / system.run0.numCycles
-        output.ylabel = 'User Fraction'
-
-        display()
-        return
+            output.display(command, options.printmode)
 
     if command == 'ticks':
         output.stat = system.run0.numCycles
-        output.binstats = [ system.run0.numCycles ]
 
         display()
         return
@@ -401,7 +378,6 @@ def commands(options, command, args):
 
     if command == 'mpkb':
         output.stat = misses / (bytes / 1024)
-        output.binstats = [ misses ]
         output.ylabel = 'Misses / KB'
         display()
         return
@@ -409,7 +385,6 @@ def commands(options, command, args):
     if command == 'ipkb':
         interrupts = system.run0.kern.faults[4]
         output.stat = interrupts / kbytes
-        output.binstats = [ interrupts ]
         output.ylabel = 'Interrupts / KB'
         display()
         return
@@ -446,7 +421,6 @@ if __name__ == '__main__':
     options.runs = None
     options.system = 'client'
     options.method = None
-    options.binned = False
     options.graph = False
     options.ticks = False
     options.printmode = 'G'
@@ -454,10 +428,8 @@ if __name__ == '__main__':
     options.jobfile = None
     options.all = False
 
-    opts, args = getopts(sys.argv[1:], '-BEFJad:g:h:j:m:pr:s:u:T:')
+    opts, args = getopts(sys.argv[1:], '-EFJad:g:h:j:m:pr:s:u:T:')
     for o,a in opts:
-        if o == '-B':
-            options.binned = True
         if o == '-E':
             options.printmode = 'E'
         if o == '-F':