AlgoTimeRegister#

This is a Python binding to the C++ class Mantid::Instrumentation::AlgoTimeRegister.

bases: mantid.api.AlgoTimeRegisterImpl

class mantid.api.AlgoTimeRegisterImpl#

static Instance() → AlgoTimeRegisterImpl :#: Returns a reference to the AlgoTimeRegister

static addTime((str)name, (int)begin, (int)end) → None :#: Adds a time entry in the file for a function with <name> that starts at <begin> time_ns and ends at <end> time_ns relative to the <START_POINT> clock

Usage#

In the beggining of the script, initialize AlgoTimeRegister to start the clock (START_POINT). Then use python time (time.time_ns()) to measure time (start_time and end_time in nanoseconds) and a function (name). Afterwards, call addTime() to store them in the performancelog.filename log file when performancelog.write is activated (defined at Mantid.user.properties file) in the following format:

ThreadID=<thread_id>, AlgorithmName=<name>, StartTime=<start_time>, EndTime=<end_time>

Example 1 - Use AlgoTimeRegister to measure time - default case: This example shows the initialization process and adding a time entry

# use python time module to measure time in nanoseconds
import time

#initialize the clock *once*
AlgoTimeRegister.Instance()

# get start time in nanoseconds before calling the function of interest
start_time = time.time_ns()

# execute the function
demo_function()

# get end time in nanoseconds after it is completed
end_time = time.time_ns()

# store the time log
name = "demo_function"
AlgoTimeRegister.addTime(name,start_time,end_time)

Example Output at Mantid.user.properties.performance.filename file:

START_POINT: 1728507408768867137 MAX_THREAD: 12
ThreadID=124522639565888, AlgorithmName=demo_function, StartTime=144539, EndTime=1346405816

Example 2 - Use AlgoTimeRegister to measure time - overwrite the performance.log file: The performance configuration can be overwritten in the script, by defining a custom performancelog.filename and calling evertyhing within the amend_config with block.

#define `performancelog.filenamer` filepath for ta particular script
performance_config = {"performancelog.write": "On", "performancelog.filename": <custom filepath>}
with amend_config(**performance_config):
    # use python time module to measure time in nanoseconds
    import time

    #initialize the clock *once*
    AlgoTimeRegister.Instance()

    # get start time in nanoseconds before calling the function of interest
    start_time = time.time_ns()

    # execute the function
    demo_function()

    # get end time in nanoseconds after it is completed
    end_time = time.time_ns()

    # store the time log
    name = "demo_function"
    AlgoTimeRegister.addTime(name,start_time,end_time)

Example Output at performance_config.performance.filename file:

START_POINT: 1728507555711066754 MAX_THREAD: 12
ThreadID=139238466401344, AlgorithmName=demo_function, StartTime=144981, EndTime=4115162290

Example 3 - Use AlgoTimeRegister to measure time - initialization omitted: The AlgoTimeRegister initialization can be omitted, if a Mantid algorithm is the first command in the script and the time before this is not needed. In any other case, initialization should be specified very eary on, in the script.

# use python time module to measure time in nanoseconds
import time

#call algorithm, clock initialization
#internally it calls AlgoTimeRegister.addTime
ws = CreateSampleWorkspace()

# get start time in nanoseconds before calling the function of interest
start_time = time.time_ns()

# execute the function
demo_function()

# get end time in nanoseconds after it is completed
end_time = time.time_ns()

# store the time log
name = "demo_function"
AlgoTimeRegister.addTime(name,start_time,end_time)

Example Output at Mantid.user.properties.performance.filename file:

START_POINT: 1728507650978221781 MAX_THREAD: 12
ThreadID=138583991714880, AlgorithmName=CreateSampleWorkspace, StartTime=154046, EndTime=1818067
ThreadID=138583991714880, AlgorithmName=demo_function, StartTime=2044656, EndTime=4117415173

AlgoTimeRegister#

Usage#

This Page