Skip to content

Commit 0478480

Browse files
author
y-p
committed
VB: cut down worst runtime offenders in vb suite
1 parent 01213e9 commit 0478480

File tree

7 files changed

+40
-40
lines changed

7 files changed

+40
-40
lines changed

‎vb_suite/binary_ops.py

Lines changed: 18 additions & 19 deletions
Original file line numberDiff line numberDiff line change
@@ -13,17 +13,17 @@
1313
# add
1414

1515
setup = common_setup + """
16-
df = DataFrame(np.random.randn(100000, 100))
17-
df2 = DataFrame(np.random.randn(100000, 100))
16+
df = DataFrame(np.random.randn(20000, 100))
17+
df2 = DataFrame(np.random.randn(20000, 100))
1818
"""
1919
frame_add = \
2020
Benchmark("df + df2", setup, name='frame_add',
2121
start_date=datetime(2012, 1, 1))
2222

2323
setup = common_setup + """
2424
import pandas.core.expressions as expr
25-
df = DataFrame(np.random.randn(100000, 100))
26-
df2 = DataFrame(np.random.randn(100000, 100))
25+
df = DataFrame(np.random.randn(20000, 100))
26+
df2 = DataFrame(np.random.randn(20000, 100))
2727
expr.set_numexpr_threads(1)
2828
"""
2929

@@ -33,8 +33,8 @@
3333

3434
setup = common_setup + """
3535
import pandas.core.expressions as expr
36-
df = DataFrame(np.random.randn(100000, 100))
37-
df2 = DataFrame(np.random.randn(100000, 100))
36+
df = DataFrame(np.random.randn(20000, 100))
37+
df2 = DataFrame(np.random.randn(20000, 100))
3838
expr.set_use_numexpr(False)
3939
"""
4040
frame_add_no_ne = \
@@ -45,17 +45,17 @@
4545
# mult
4646

4747
setup = common_setup + """
48-
df = DataFrame(np.random.randn(100000, 100))
49-
df2 = DataFrame(np.random.randn(100000, 100))
48+
df = DataFrame(np.random.randn(20000, 100))
49+
df2 = DataFrame(np.random.randn(20000, 100))
5050
"""
5151
frame_mult = \
5252
Benchmark("df * df2", setup, name='frame_mult',
5353
start_date=datetime(2012, 1, 1))
5454

5555
setup = common_setup + """
5656
import pandas.core.expressions as expr
57-
df = DataFrame(np.random.randn(100000, 100))
58-
df2 = DataFrame(np.random.randn(100000, 100))
57+
df = DataFrame(np.random.randn(20000, 100))
58+
df2 = DataFrame(np.random.randn(20000, 100))
5959
expr.set_numexpr_threads(1)
6060
"""
6161
frame_mult_st = \
@@ -64,8 +64,8 @@
6464

6565
setup = common_setup + """
6666
import pandas.core.expressions as expr
67-
df = DataFrame(np.random.randn(100000, 100))
68-
df2 = DataFrame(np.random.randn(100000, 100))
67+
df = DataFrame(np.random.randn(20000, 100))
68+
df2 = DataFrame(np.random.randn(20000, 100))
6969
expr.set_use_numexpr(False)
7070
"""
7171
frame_mult_no_ne = \
@@ -76,17 +76,17 @@
7676
# multi and
7777

7878
setup = common_setup + """
79-
df = DataFrame(np.random.randn(100000, 100))
80-
df2 = DataFrame(np.random.randn(100000, 100))
79+
df = DataFrame(np.random.randn(20000, 100))
80+
df2 = DataFrame(np.random.randn(20000, 100))
8181
"""
8282
frame_multi_and = \
8383
Benchmark("df[(df>0) & (df2>0)]", setup, name='frame_multi_and',
8484
start_date=datetime(2012, 1, 1))
8585

8686
setup = common_setup + """
8787
import pandas.core.expressions as expr
88-
df = DataFrame(np.random.randn(100000, 100))
89-
df2 = DataFrame(np.random.randn(100000, 100))
88+
df = DataFrame(np.random.randn(20000, 100))
89+
df2 = DataFrame(np.random.randn(20000, 100))
9090
expr.set_numexpr_threads(1)
9191
"""
9292
frame_multi_and_st = \
@@ -95,11 +95,10 @@
9595

9696
setup = common_setup + """
9797
import pandas.core.expressions as expr
98-
df = DataFrame(np.random.randn(100000, 100))
99-
df2 = DataFrame(np.random.randn(100000, 100))
98+
df = DataFrame(np.random.randn(20000, 100))
99+
df2 = DataFrame(np.random.randn(20000, 100))
100100
expr.set_use_numexpr(False)
101101
"""
102102
frame_multi_and_no_ne = \
103103
Benchmark("df[(df>0) & (df2>0)]", setup, name='frame_multi_and_no_ne',cleanup="expr.set_use_numexpr(True)",
104104
start_date=datetime(2012, 1, 1))
105-

‎vb_suite/groupby.py

Lines changed: 2 additions & 2 deletions
Original file line numberDiff line numberDiff line change
@@ -278,8 +278,8 @@ def f(g):
278278
# Transform testing
279279

280280
setup = common_setup + """
281-
n_dates = 1000
282-
n_securities = 500
281+
n_dates = 400
282+
n_securities = 250
283283
n_columns = 3
284284
share_na = 0.1
285285

‎vb_suite/hdfstore_bench.py

Lines changed: 8 additions & 7 deletions
Original file line numberDiff line numberDiff line change
@@ -92,12 +92,13 @@ def remove(f):
9292
# get from a table (mixed)
9393

9494
setup5 = common_setup + """
95-
index = [rands(10) for _ in xrange(25000)]
96-
df = DataFrame({'float1' : randn(25000),
97-
'float2' : randn(25000),
98-
'string1' : ['foo'] * 25000,
99-
'bool1' : [True] * 25000,
100-
'int1' : np.random.randint(0, 250000, size=25000)},
95+
N=10000
96+
index = [rands(10) for _ in xrange(N)]
97+
df = DataFrame({'float1' : randn(N),
98+
'float2' : randn(N),
99+
'string1' : ['foo'] * N,
100+
'bool1' : [True] * N,
101+
'int1' : np.random.randint(0, N, size=N)},
101102
index=index)
102103
103104
remove(f)
@@ -265,7 +266,7 @@ def remove(f):
265266
# write to a table (data_columns)
266267

267268
setup15 = common_setup + """
268-
df = DataFrame(np.random.randn(25000,10),columns = [ 'C%03d' % i for i in xrange(10) ])
269+
df = DataFrame(np.random.randn(10000,10),columns = [ 'C%03d' % i for i in xrange(10) ])
269270
270271
remove(f)
271272
store = HDFStore(f)

‎vb_suite/indexing.py

Lines changed: 6 additions & 6 deletions
Original file line numberDiff line numberDiff line change
@@ -95,17 +95,17 @@
9595
name='indexing_dataframe_boolean_rows_object')
9696

9797
setup = common_setup + """
98-
df = DataFrame(np.random.randn(100000, 100))
99-
df2 = DataFrame(np.random.randn(100000, 100))
98+
df = DataFrame(np.random.randn(50000, 100))
99+
df2 = DataFrame(np.random.randn(50000, 100))
100100
"""
101101
indexing_dataframe_boolean = \
102102
Benchmark("df > df2", setup, name='indexing_dataframe_boolean',
103103
start_date=datetime(2012, 1, 1))
104104

105105
setup = common_setup + """
106106
import pandas.core.expressions as expr
107-
df = DataFrame(np.random.randn(100000, 100))
108-
df2 = DataFrame(np.random.randn(100000, 100))
107+
df = DataFrame(np.random.randn(50000, 100))
108+
df2 = DataFrame(np.random.randn(50000, 100))
109109
expr.set_numexpr_threads(1)
110110
"""
111111

@@ -115,8 +115,8 @@
115115

116116
setup = common_setup + """
117117
import pandas.core.expressions as expr
118-
df = DataFrame(np.random.randn(100000, 100))
119-
df2 = DataFrame(np.random.randn(100000, 100))
118+
df = DataFrame(np.random.randn(50000, 100))
119+
df2 = DataFrame(np.random.randn(50000, 100))
120120
expr.set_use_numexpr(False)
121121
"""
122122

‎vb_suite/io_bench.py

Lines changed: 3 additions & 3 deletions
Original file line numberDiff line numberDiff line change
@@ -47,7 +47,7 @@
4747
#----------------------------------
4848

4949
setup = common_setup + """
50-
df=DataFrame({'A':range(100000)})
50+
df=DataFrame({'A':range(50000)})
5151
df['B'] = df.A + 1.0
5252
df['C'] = df.A + 2.0
5353
df['D'] = df.A + 3.0
@@ -61,8 +61,8 @@
6161
6262
def create_cols(name):
6363
return [ "%s%03d" % (name,i) for i in xrange(5) ]
64-
df_float = DataFrame(np.random.randn(10000, 5),dtype='float64',columns=create_cols('float'))
65-
df_int = DataFrame(np.random.randn(10000, 5),dtype='int64',columns=create_cols('int'))
64+
df_float = DataFrame(np.random.randn(5000, 5),dtype='float64',columns=create_cols('float'))
65+
df_int = DataFrame(np.random.randn(5000, 5),dtype='int64',columns=create_cols('int'))
6666
df_bool = DataFrame(True,index=df_float.index,columns=create_cols('bool'))
6767
df_object = DataFrame('foo',index=df_float.index,columns=create_cols('object'))
6868
df_dt = DataFrame(Timestamp('20010101'),index=df_float.index,columns=create_cols('date'))

‎vb_suite/parser.py

Lines changed: 2 additions & 2 deletions
Original file line numberDiff line numberDiff line change
@@ -55,7 +55,7 @@
5555
KORD,19990127, 21:00:00, 21:18:00, -0.9900, 2.0100, 3.6000, 0.0000, 270.0000
5656
KORD,19990127, 22:00:00, 21:56:00, -0.5900, 1.7100, 5.1000, 0.0000, 290.0000
5757
'''
58-
data = data * 2000
58+
data = data * 200
5959
"""
6060
cmd = ("read_table(StringIO(data), sep=',', header=None, "
6161
"parse_dates=[[1,2], [1,3]])")
@@ -74,7 +74,7 @@
7474
KORD,19990127 21:00:00, 21:18:00, -0.9900, 2.0100, 3.6000, 0.0000, 270.0000
7575
KORD,19990127 22:00:00, 21:56:00, -0.5900, 1.7100, 5.1000, 0.0000, 290.0000
7676
'''
77-
data = data * 2000
77+
data = data * 200
7878
"""
7979
cmd = "read_table(StringIO(data), sep=',', header=None, parse_dates=[1])"
8080
sdate = datetime(2012, 5, 7)

‎vb_suite/timeseries.py

Lines changed: 1 addition & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -162,7 +162,7 @@ def date_range(start=None, end=None, periods=None, freq=None):
162162
# setitem PeriodIndex
163163

164164
setup = common_setup + """
165-
rng = period_range('1/1/1990', freq='S', periods=100000)
165+
rng = period_range('1/1/1990', freq='S', periods=20000)
166166
df = DataFrame(index=range(len(rng)))
167167
"""
168168

0 commit comments

Comments
 (0)