mycpp Code Generation

Measure the speedup from mycpp, and the resource usage.

Source code: oil/mycpp/examples

User Time (milliseconds)

Lower ratios are better.

example name C++ Python C++ : Python
loops 0 501 0.000
varargs 0 10 0.000
modules 1 323 0.003
fib_iter 3 974 0.004
scoped_resource 10 1,547 0.006
fib_recursive 13 1,491 0.009
asdl_generated 18 710 0.025
parse 43 1,379 0.031
files 9 132 0.068
containers 10 131 0.076
tuple_return_value 26 308 0.086
classes 3 27 0.106
iterators 1 8 0.126
small_str 1 8 0.129
classes_gc 2 14 0.159
length 67 385 0.175
cgi 241 1,063 0.227
arith_ops 1 4 0.259
escape 154 547 0.282
cartesian 157 554 0.283
strings 1 4 0.283
control_flow 316 196 1.616

Max Resident Set Size (MB)

Lower ratios are better. We use MB (powers of 10), not MiB (powers of 2).

example name C++ Python C++ : Python
classes 4.4 10.9 0.40
classes_gc 4.0 9.4 0.43
loops 3.5 7.6 0.46
strings 3.5 7.5 0.47
parse 3.8 8.1 0.47
arith_ops 3.4 7.2 0.48
small_str 3.5 7.3 0.48
iterators 3.6 7.4 0.48
scoped_resource 3.6 7.3 0.49
length 3.7 7.4 0.50
fib_iter 3.7 7.4 0.50
cgi 3.7 7.3 0.50
fib_recursive 3.7 7.3 0.51
varargs 3.8 7.4 0.51
control_flow 3.8 7.5 0.51
asdl_generated 3.7 7.3 0.51
modules 3.8 7.4 0.51
escape 3.7 7.2 0.51
cartesian 3.7 7.3 0.51
tuple_return_value 3.7 7.2 0.52
files 3.8 7.2 0.53
containers 28.5 47.7 0.60

System Time (milliseconds)

Lower ratios are better.

example name C++ Python C++ : Python
arith_ops 0 4 0.000
asdl_generated 0 4 0.000
cgi 0 4 0.000
classes 0 8 0.000
control_flow 0 4 0.000
fib_recursive 0 8 0.000
length 0 4 0.000
parse 0 4 0.000
scoped_resource 0 8 0.000
strings 0 4 0.000
containers 5 24 0.208
fib_iter 3 4 0.851
loops 4 0 inf
varargs 4 0 inf
cartesian 0 0 NA
classes_gc 0 0 NA
escape 0 0 NA
files 0 0 NA
iterators 0 0 NA
modules 0 0 NA
small_str 0 0 NA
tuple_return_value 0 0 NA

raw benchmark files