def show_help():
print """Usage: benchmark.py [OPTION]... [PROJECT]...
Test distcc relative performance building different projects.
By default, all known projects are built.
Options:
--help show brief help message
--list-projects show defined projects
--cc=PATH specify base value of CC to pass to configure
--cxx=PATH specify base value of CXX to pass to configure
--output=FILE print final summary to FILE in addition to stdout
-c, --compiler=COMPILER specify one compiler to use; format is
{local|dist|lzo|pump},h<NUMHOSTS>,j<NUMJOBS>
-n N repeat compilation N times
-a, --actions=ACTIONS comma-separated list of action phases
to perform
-f N, --force=N If set to 0, skip download, unpack, and
configure actions if they've already been
successfully performed; if set to 1 (the
default), only skip the download action;
if set to 2, do not skip any action
The C and C++ compiler versions used can be set with the --cc and --cxx
options.
Use of distcc features is set with the -c/--compiler option. The argument
to -c/--compiler has three components, separated by commas. The first
component specifies which distcc features to enabled: "local" means
run cc locally, "distcc" means use plain distcc, "lzo" means enabled
compression, and "pump" means to enable pump mode and compression.
The second component specifies how many distcc hosts to use. The third
component specifies how many jobs to run (the -j/--jobs option to "make").
Multiple -c/--compiler options specify different scenarios to measure.
The default is to measure a few reasonable scenarios.
"""
actions.action_help()
Options:
--help show brief help message
--list-projects show defined projects
-c, --compiler=COMPILER specify one compiler to use
-n N repeat compilation N times
-a, --actions=ACTIONS comma-separated list of action phases
to perform
Compilers can be specified as either "local,N" to run N copies of gcc,
or dist,N to run N copies of distcc. Multiple -c options specify
different scenarios to measure. The default is to run a nonparallel
local compile and a parallel distributed compile.
"""
actions.action_help()
# -a is for developer use only and not documented; unless you're
# careful the results will just be confusing.
######################################################################
def main():
"""Run the benchmark per arguments"""
sum = Summary()
options, args = getopt(sys.argv[1:], "a:c:n:", ["list-projects", "actions=", "help", "compiler="])
opt_actions = actions.default_actions
set_compilers = []
opt_repeats = 1
Options:
--help show brief help message
--list-projects show defined projects
-c, --compiler=COMPILER specify one compiler to use
-n N repeat compilation N times
-a, --actions=ACTIONS comma-separated list of action phases
to perform
Compilers can be specified as either "local,N" to run N copies of gcc,
or dist,N to run N copies of distcc. Multiple -c options specify
different scenarios to measure. The default is to run a nonparallel
local compile and a parallel distributed compile.
"""
actions.action_help()
# -a is for developer use only and not documented; unless you're
# careful the results will just be confusing.
######################################################################
def main():
"""Run the benchmark per arguments"""
sum = Summary()
options, args = getopt(sys.argv[1:], 'a:c:n:',
['list-projects', 'actions=', 'help', 'compiler='])
opt_actions = actions.default_actions
set_compilers = []
opt_repeats = 1