Python os 模块,times() 实例源码

我们从Python开源项目中,提取了以下50个代码示例,用于说明如何使用os.times()

项目:OSPTF    作者:xSploited    | 项目源码 | 文件源码
def cpu_times(percpu=False):
    """Return system-wide CPU times as a namedtuple.
    Every CPU time represents the seconds the CPU has spent in the given mode.
    The namedtuple's fields availability varies depending on the platform:
     - user
     - system
     - idle
     - nice (UNIX)
     - iowait (Linux)
     - irq (Linux, FreeBSD)
     - softirq (Linux)
     - steal (Linux >= 2.6.11)
     - guest (Linux >= 2.6.24)
     - guest_nice (Linux >= 3.2.0)

    When percpu is True return a list of namedtuples for each CPU.
    First element of the list refers to first CPU, second element
    to second CPU and so on.
    The order of the list is consistent across calls.
    """
    if not percpu:
        return _psplatform.cpu_times()
    else:
        return _psplatform.per_cpu_times()
项目:pupy    作者:ru-faraon    | 项目源码 | 文件源码
def cpu_times(percpu=False):
    """Return system-wide CPU times as a namedtuple.
    Every CPU time represents the seconds the CPU has spent in the given mode.
    The namedtuple's fields availability varies depending on the platform:
     - user
     - system
     - idle
     - nice (UNIX)
     - iowait (Linux)
     - irq (Linux, FreeBSD)
     - softirq (Linux)
     - steal (Linux >= 2.6.11)
     - guest (Linux >= 2.6.24)
     - guest_nice (Linux >= 3.2.0)

    When percpu is True return a list of namedtuples for each CPU.
    First element of the list refers to first CPU, second element
    to second CPU and so on.
    The order of the list is consistent across calls.
    """
    if not percpu:
        return _psplatform.cpu_times()
    else:
        return _psplatform.per_cpu_times()
项目:respeaker_virtualenv    作者:respeaker    | 项目源码 | 文件源码
def test_cmdline(self):
        cmdline = [PYTHON, "-c", "import time; time.sleep(60)"]
        sproc = get_test_subprocess(cmdline)
        try:
            self.assertEqual(' '.join(psutil.Process(sproc.pid).cmdline()),
                             ' '.join(cmdline))
        except AssertionError:
            # XXX - most of the times the underlying sysctl() call on Net
            # and Open BSD returns a truncated string.
            # Also /proc/pid/cmdline behaves the same so it looks
            # like this is a kernel bug.
            if NETBSD or OPENBSD:
                self.assertEqual(
                    psutil.Process(sproc.pid).cmdline()[0], PYTHON)
            else:
                raise
项目:pipenv    作者:pypa    | 项目源码 | 文件源码
def cpu_times(percpu=False):
    """Return system-wide CPU times as a namedtuple.
    Every CPU time represents the seconds the CPU has spent in the given mode.
    The namedtuple's fields availability varies depending on the platform:
     - user
     - system
     - idle
     - nice (UNIX)
     - iowait (Linux)
     - irq (Linux, FreeBSD)
     - softirq (Linux)
     - steal (Linux >= 2.6.11)
     - guest (Linux >= 2.6.24)
     - guest_nice (Linux >= 3.2.0)

    When percpu is True return a list of namedtuples for each CPU.
    First element of the list refers to first CPU, second element
    to second CPU and so on.
    The order of the list is consistent across calls.
    """
    if not percpu:
        return _psplatform.cpu_times()
    else:
        return _psplatform.per_cpu_times()
项目:pipenv    作者:pypa    | 项目源码 | 文件源码
def _cpu_tot_time(times):
    """Given a cpu_time() ntuple calculates the total CPU time
    (including idle time).
    """
    tot = sum(times)
    if LINUX:
        # On Linux guest times are already accounted in "user" or
        # "nice" times, so we subtract them from total.
        # Htop does the same. References:
        # https://github.com/giampaolo/psutil/pull/940
        # http://unix.stackexchange.com/questions/178045
        # https://github.com/torvalds/linux/blob/
        #     447976ef4fd09b1be88b316d1a81553f1aa7cd07/kernel/sched/
        #     cputime.c#L158
        tot -= getattr(times, "guest", 0)  # Linux 2.6.24+
        tot -= getattr(times, "guest_nice", 0)  # Linux 3.2.0+
    return tot
项目:ropi    作者:ThumbGen    | 项目源码 | 文件源码
def cpu_times(percpu=False):
    """Return system-wide CPU times as a namedtuple.
    Every CPU time represents the seconds the CPU has spent in the given mode.
    The namedtuple's fields availability varies depending on the platform:
     - user
     - system
     - idle
     - nice (UNIX)
     - iowait (Linux)
     - irq (Linux, FreeBSD)
     - softirq (Linux)
     - steal (Linux >= 2.6.11)
     - guest (Linux >= 2.6.24)
     - guest_nice (Linux >= 3.2.0)

    When percpu is True return a list of namedtuples for each CPU.
    First element of the list refers to first CPU, second element
    to second CPU and so on.
    The order of the list is consistent across calls.
    """
    if not percpu:
        return _psplatform.cpu_times()
    else:
        return _psplatform.per_cpu_times()
项目:zenchmarks    作者:squeaky-pl    | 项目源码 | 文件源码
def test_cmdline(self):
        cmdline = [PYTHON, "-c", "import time; time.sleep(60)"]
        sproc = get_test_subprocess(cmdline)
        try:
            self.assertEqual(' '.join(psutil.Process(sproc.pid).cmdline()),
                             ' '.join(cmdline))
        except AssertionError:
            # XXX - most of the times the underlying sysctl() call on Net
            # and Open BSD returns a truncated string.
            # Also /proc/pid/cmdline behaves the same so it looks
            # like this is a kernel bug.
            if NETBSD or OPENBSD:
                self.assertEqual(
                    psutil.Process(sproc.pid).cmdline()[0], PYTHON)
            else:
                raise
项目:zenchmarks    作者:squeaky-pl    | 项目源码 | 文件源码
def _cpu_tot_time(times):
    """Given a cpu_time() ntuple calculates the total CPU time
    (including idle time).
    """
    tot = sum(times)
    if LINUX:
        # On Linux guest times are already accounted in "user" or
        # "nice" times, so we subtract them from total.
        # Htop does the same. References:
        # https://github.com/giampaolo/psutil/pull/940
        # http://unix.stackexchange.com/questions/178045
        # https://github.com/torvalds/linux/blob/
        #     447976ef4fd09b1be88b316d1a81553f1aa7cd07/kernel/sched/
        #     cputime.c#L158
        tot -= getattr(times, "guest", 0)  # Linux 2.6.24+
        tot -= getattr(times, "guest_nice", 0)  # Linux 3.2.0+
    return tot
项目:FancyWord    作者:EastonLee    | 项目源码 | 文件源码
def test_cmdline(self):
        cmdline = [PYTHON, "-c", "import time; time.sleep(60)"]
        sproc = get_test_subprocess(cmdline)
        try:
            self.assertEqual(' '.join(psutil.Process(sproc.pid).cmdline()),
                             ' '.join(cmdline))
        except AssertionError:
            # XXX - most of the times the underlying sysctl() call on Net
            # and Open BSD returns a truncated string.
            # Also /proc/pid/cmdline behaves the same so it looks
            # like this is a kernel bug.
            if NETBSD or OPENBSD:
                self.assertEqual(
                    psutil.Process(sproc.pid).cmdline()[0], PYTHON)
            else:
                raise
项目:FancyWord    作者:EastonLee    | 项目源码 | 文件源码
def _cpu_tot_time(times):
    """Given a cpu_time() ntuple calculates the total CPU time
    (including idle time).
    """
    tot = sum(times)
    if LINUX:
        # On Linux guest times are already accounted in "user" or
        # "nice" times, so we subtract them from total.
        # Htop does the same. References:
        # https://github.com/giampaolo/psutil/pull/940
        # http://unix.stackexchange.com/questions/178045
        # https://github.com/torvalds/linux/blob/
        #     447976ef4fd09b1be88b316d1a81553f1aa7cd07/kernel/sched/
        #     cputime.c#L158
        tot -= getattr(times, "guest", 0)  # Linux 2.6.24+
        tot -= getattr(times, "guest_nice", 0)  # Linux 3.2.0+
    return tot
项目:turbo_seti    作者:UCBerkeleySETI    | 项目源码 | 文件源码
def search(self):
        '''Top level search.
        '''
        logger.debug("Start searching...")
        logger.debug(self.get_info())

        self.logwriter = file_writers.LogWriter('%s/%s.log'%(self.out_dir.rstrip('/'), self.data_handle.data_list[0].filename.split('/')[-1].replace('.h5','').replace('.fits','').replace('.fil','')))
        self.filewriter = file_writers.FileWriter('%s/%s.dat'%(self.out_dir.rstrip('/'), self.data_handle.data_list[0].filename.split('/')[-1].replace('.h5','').replace('.fits','').replace('.fil','')),self.data_handle.data_list[0].header)

        logger.info("Start ET search for %s"%self.data_handle.data_list[0].filename)
        self.logwriter.info("Start ET search for %s"%(self.data_handle.data_list[0].filename))

        for ii,target_data_obj in enumerate(self.data_handle.data_list):
            self.search_data(target_data_obj)
##EE-benshmark            cProfile.runctx('self.search_data(target_data_obj)',globals(),locals(),filename='profile_M%2.1f_S%2.1f_t%i'%(self.max_drift,self.snr,int(os.times()[-1])))

            #----------------------------------------
            #Closing instance. Collect garbage.
            self.data_handle.data_list[ii].close()
            gc.collect()
项目:teamspector    作者:wladston    | 项目源码 | 文件源码
def t(s, sinceBegin=False):
    """Pretty-print timing information."""
    global CTIME, CTIMES
    nt = int(time.time())
    ntimes = os.times()
    if not sinceBegin:
        ct = CTIME
        cts = CTIMES
    else:
        ct = BEGIN_TIME
        cts = BEGIN_TIMES
    print '# TIME', s, \
            ': %dmin, %dsec (wall) %dmin, %dsec (user) %dmin, %dsec (system)' \
            % _minSec(nt-ct, ntimes[0]-cts[0], ntimes[1]-cts[1])
    if not sinceBegin:
        CTIME = nt
        CTIMES = ntimes
项目:kinect-2-libras    作者:inessadl    | 项目源码 | 文件源码
def _get_time_times(timer=os.times):
        t = timer()
        return t[0] + t[1]

# Using getrusage(3) is better than clock(3) if available:
# on some systems (e.g. FreeBSD), getrusage has a higher resolution
# Furthermore, on a POSIX system, returns microseconds, which
# wrap around after 36min.
项目:UIP    作者:NITDgpOS    | 项目源码 | 文件源码
def changeCycle(self):
        """Wallpaper change cycle."""
        uold, sold, cold, c, e = os.times()
        while True:
            delta = self.deltaTime()
            if delta >= self.timeout:
                self.change_next()
                self.time = time.time()
            unew, snew, cnew, c, e = os.times()
            start = time.time()
            percentage = get_percentage(unew, uold, start)
            if percentage > 30.0:
                time.sleep(0.1)
项目:hostapd-mana    作者:adde88    | 项目源码 | 文件源码
def _get_time_times(timer=os.times):
        t = timer()
        return t[0] + t[1]

# Using getrusage(3) is better than clock(3) if available:
# on some systems (e.g. FreeBSD), getrusage has a higher resolution
# Furthermore, on a POSIX system, returns microseconds, which
# wrap around after 36min.
项目:OSPTF    作者:xSploited    | 项目源码 | 文件源码
def threads(self):
        """Return threads opened by process as a list of
        (id, user_time, system_time) namedtuples representing
        thread id and thread CPU times (user/system).
        On OpenBSD this method requires root access.
        """
        return self._proc.threads()
项目:OSPTF    作者:xSploited    | 项目源码 | 文件源码
def cpu_times(self):
        """Return a (user, system, children_user, children_system)
        namedtuple representing the accumulated process time, in
        seconds.
        This is similar to os.times() but per-process.
        On OSX and Windows children_user and children_system are
        always set to 0.
        """
        return self._proc.cpu_times()
项目:OSPTF    作者:xSploited    | 项目源码 | 文件源码
def threads(self):
        """Return threads opened by process as a list of
        (id, user_time, system_time) namedtuples representing
        thread id and thread CPU times (user/system).
        On OpenBSD this method requires root access.
        """
        return self._proc.threads()
项目:OSPTF    作者:xSploited    | 项目源码 | 文件源码
def cpu_times(self):
        """Return a (user, system, children_user, children_system)
        namedtuple representing the accumulated process time, in
        seconds.
        This is similar to os.times() but per-process.
        On OSX and Windows children_user and children_system are
        always set to 0.
        """
        return self._proc.cpu_times()
项目:pupy    作者:ru-faraon    | 项目源码 | 文件源码
def threads(self):
        """Return threads opened by process as a list of
        (id, user_time, system_time) namedtuples representing
        thread id and thread CPU times (user/system).
        On OpenBSD this method requires root access.
        """
        return self._proc.threads()
项目:pupy    作者:ru-faraon    | 项目源码 | 文件源码
def cpu_times(self):
        """Return a (user, system, children_user, children_system)
        namedtuple representing the accumulated process time, in
        seconds.
        This is similar to os.times() but per-process.
        On OSX and Windows children_user and children_system are
        always set to 0.
        """
        return self._proc.cpu_times()
项目:pupy    作者:ru-faraon    | 项目源码 | 文件源码
def threads(self):
        """Return threads opened by process as a list of
        (id, user_time, system_time) namedtuples representing
        thread id and thread CPU times (user/system).
        On OpenBSD this method requires root access.
        """
        return self._proc.threads()
项目:pupy    作者:ru-faraon    | 项目源码 | 文件源码
def cpu_times(self):
        """Return a (user, system, children_user, children_system)
        namedtuple representing the accumulated process time, in
        seconds.
        This is similar to os.times() but per-process.
        On OSX and Windows children_user and children_system are
        always set to 0.
        """
        return self._proc.cpu_times()
项目:ipmisim    作者:rhtyd    | 项目源码 | 文件源码
def _monotonic_time():
    return os.times()[4]
项目:zippy    作者:securesystemslab    | 项目源码 | 文件源码
def _get_time_times(timer=os.times):
        t = timer()
        return t[0] + t[1]

# Using getrusage(3) is better than clock(3) if available:
# on some systems (e.g. FreeBSD), getrusage has a higher resolution
# Furthermore, on a POSIX system, returns microseconds, which
# wrap around after 36min.
项目:respeaker_virtualenv    作者:respeaker    | 项目源码 | 文件源码
def test_cpu_times(self):
        times = psutil.Process().cpu_times()
        assert (times.user > 0.0) or (times.system > 0.0), times
        assert (times.children_user >= 0.0), times
        assert (times.children_system >= 0.0), times
        # make sure returned values can be pretty printed with strftime
        for name in times._fields:
            time.strftime("%H:%M:%S", time.localtime(getattr(times, name)))

    # Test Process.cpu_times() against os.times()
    # os.times() is broken on Python 2.6
    # http://bugs.python.org/issue1040026
    # XXX fails on OSX: not sure if it's for os.times(). We should
    # try this with Python 2.7 and re-enable the test.
项目:respeaker_virtualenv    作者:respeaker    | 项目源码 | 文件源码
def test_cpu_times_2(self):
        user_time, kernel_time = psutil.Process().cpu_times()[:2]
        utime, ktime = os.times()[:2]

        # Use os.times()[:2] as base values to compare our results
        # using a tolerance  of +/- 0.1 seconds.
        # It will fail if the difference between the values is > 0.1s.
        if (max([user_time, utime]) - min([user_time, utime])) > 0.1:
            self.fail("expected: %s, found: %s" % (utime, user_time))

        if (max([kernel_time, ktime]) - min([kernel_time, ktime])) > 0.1:
            self.fail("expected: %s, found: %s" % (ktime, kernel_time))
项目:respeaker_virtualenv    作者:respeaker    | 项目源码 | 文件源码
def threads(self):
        """Return threads opened by process as a list of
        (id, user_time, system_time) namedtuples representing
        thread id and thread CPU times (user/system).
        On OpenBSD this method requires root access.
        """
        return self._proc.threads()
项目:respeaker_virtualenv    作者:respeaker    | 项目源码 | 文件源码
def cpu_times(self):
        """Return a (user, system, children_user, children_system)
        namedtuple representing the accumulated process time, in
        seconds.
        This is similar to os.times() but per-process.
        On OSX and Windows children_user and children_system are
        always set to 0.
        """
        return self._proc.cpu_times()
项目:travelling_sales_person_AI    作者:jana-uoft    | 项目源码 | 文件源码
def initStats(self):
        sNode.n = 0
        StateSpace.n = 1    #initial state already generated on call so search
        self.total_search_time = 0
        self.cycle_check_pruned = 0
        self.total_search_time = os.times()[0]
项目:oil    作者:oilshell    | 项目源码 | 文件源码
def _get_time_times(timer=os.times):
        t = timer()
        return t[0] + t[1]

# Using getrusage(3) is better than clock(3) if available:
# on some systems (e.g. FreeBSD), getrusage has a higher resolution
# Furthermore, on a POSIX system, returns microseconds, which
# wrap around after 36min.
项目:oil    作者:oilshell    | 项目源码 | 文件源码
def run_during(duration, func):
    _t = time.time
    n = 0
    start = os.times()
    start_timestamp = _t()
    real_start = start[4] or start_timestamp
    while True:
        func()
        n += 1
        if _t() - start_timestamp > duration:
            break
    end = os.times()
    real = (end[4] if start[4] else time.time()) - real_start
    return n, real, sum(end[0:2]) - sum(start[0:2])
项目:python2-tracer    作者:extremecoders-re    | 项目源码 | 文件源码
def TSTART():
    global t0, t1
    u, s, cu, cs = os.times()
    t0 = u+cu, s+cs, time.time()
项目:python2-tracer    作者:extremecoders-re    | 项目源码 | 文件源码
def TSTOP(*label):
    global t0, t1
    u, s, cu, cs = os.times()
    t1 = u+cu, s+cs, time.time()
    tt = []
    for i in range(3):
        tt.append(t1[i] - t0[i])
    [u, s, r] = tt
    msg = ''
    for x in label: msg = msg + (x + ' ')
    msg = msg + '%r user, %r sys, %r real\n' % (u, s, r)
    sys.stderr.write(msg)
项目:python2-tracer    作者:extremecoders-re    | 项目源码 | 文件源码
def _get_time_times(timer=os.times):
        t = timer()
        return t[0] + t[1]

# Using getrusage(3) is better than clock(3) if available:
# on some systems (e.g. FreeBSD), getrusage has a higher resolution
# Furthermore, on a POSIX system, returns microseconds, which
# wrap around after 36min.
项目:python2-tracer    作者:extremecoders-re    | 项目源码 | 文件源码
def run_during(duration, func):
    _t = time.time
    n = 0
    start = os.times()
    start_timestamp = _t()
    real_start = start[4] or start_timestamp
    while True:
        func()
        n += 1
        if _t() - start_timestamp > duration:
            break
    end = os.times()
    real = (end[4] if start[4] else time.time()) - real_start
    return n, real, sum(end[0:2]) - sum(start[0:2])
项目:sslstrip-hsts-openwrt    作者:adde88    | 项目源码 | 文件源码
def _get_time_times(timer=os.times):
        t = timer()
        return t[0] + t[1]

# Using getrusage(3) is better than clock(3) if available:
# on some systems (e.g. FreeBSD), getrusage has a higher resolution
# Furthermore, on a POSIX system, returns microseconds, which
# wrap around after 36min.
项目:CSC384Project    作者:mariabianchi3    | 项目源码 | 文件源码
def initStats(self):
        sNode.n = 0
        StateSpace.n = 1    #initial state already generated on call so search
        self.cycle_check_pruned = 0
        self.cost_bound_pruned = 0 
        self.total_search_time = 0       
        self.total_search_time = os.times()[0]
项目:Mac-Python-3.X    作者:L1nwatch    | 项目源码 | 文件源码
def main():
    print(os.nice(0))  # get relative process priority
    print(os.nice(1))  # change relative priority
    print(os.times())  # process times: system, user etc...
    print(os.isatty(0))  # is the file descriptor arg a tty?(0 = stdin)
    print(os.isatty(4))  # 4 is just an arbitrary test value
    print(os.getloadavg())  # UNIX only - number of processes in queue
    print(os.cpu_count())  # New in Python 3.4
项目:FOLLOW    作者:adityagilra    | 项目源码 | 文件源码
def sim_run_flush(tFlush,nFlush):
        '''
            Run simulation for nFlush*tFlush seconds,
            Flush probes every tFlush of simulation time,
              (only flush those that don't have 'weights' in their label names)
        '''        
        weighttimeidxold = 0
        #doubledLearningRate = False
        for duration in [tFlush]*nFlush:
            _,_,_,_,realtime = os.times()
            print("Finished till",sim.time,'s, in',realtime-realtimeold,'s')
            sys.stdout.flush()
            # save weights if weightdt or more has passed since last save
            weighttimeidx = int(sim.time/weightdt)
            if weighttimeidx > weighttimeidxold:
                weighttimeidxold = weighttimeidx
                save_current_weights(False,sim.time)
            # flush probes
            for probe in sim.model.probes:
                # except weight probes (flushed in save_current_weights)
                # except error probe which is saved fully in ..._end.shelve
                if probe.label is not None:
                    if 'weights' in probe.label or 'error' in probe.label:
                        break
                del sim._probe_outputs[probe][:]
            ## if time > 1000s, double learning rate
            #if sim.time>1000. and not doubledLearningRate:
            #    changeLearningRate(4.)  # works only if excPESDecayRate = None
            #    doubledLearningRate = True
            # run simulation for tFlush duration
            sim.run(duration,progress_bar=False)

    ###
    ### run the simulation, with flushing for learning simulations ###
    ###
项目:FOLLOW    作者:adityagilra    | 项目源码 | 文件源码
def sim_run_flush(tFlush,nFlush):
        '''
            Run simulation for nFlush*tFlush seconds,
            Flush probes every tFlush of simulation time,
              (only flush those that don't have 'weights' in their label names)
        '''        
        weighttimeidxold = 0
        #doubledLearningRate = False
        for duration in [tFlush]*nFlush:
            _,_,_,_,realtime = os.times()
            print("Finished till",sim.time,'s, in',realtime-realtimeold,'s')
            sys.stdout.flush()
            # save weights if weightdt or more has passed since last save
            weighttimeidx = int(sim.time/weightdt)
            if weighttimeidx > weighttimeidxold:
                weighttimeidxold = weighttimeidx
                save_current_weights(False,sim.time)
            # flush probes
            for probe in sim.model.probes:
                # except weight probes (flushed in save_current_weights)
                # except error probe which is saved fully in ..._end.shelve
                if probe.label is not None:
                    if 'weights' in probe.label or 'error' in probe.label:
                        break
                del sim._probe_outputs[probe][:]
            ## if time > 1000s, double learning rate
            #if sim.time>1000. and not doubledLearningRate:
            #    changeLearningRate(4.)  # works only if excPESDecayRate = None
            #    doubledLearningRate = True
            # run simulation for tFlush duration
            sim.run(duration,progress_bar=False)

    ###
    ### run the simulation, with flushing for learning simulations ###
    ###
项目:FOLLOW    作者:adityagilra    | 项目源码 | 文件源码
def sim_run_flush(tFlush,nFlush):
        '''
            Run simulation for nFlush*tFlush seconds,
            Flush probes every tFlush of simulation time,
              (only flush those that don't have 'weights' in their label names)
        '''        
        weighttimeidxold = 0
        #doubledLearningRate = False
        for duration in [tFlush]*nFlush:
            _,_,_,_,realtime = os.times()
            print("Finished till",sim.time,'s, in',realtime-realtimeold,'s')
            sys.stdout.flush()
            # save weights if weightdt or more has passed since last save
            weighttimeidx = int(sim.time/weightdt)
            if weighttimeidx > weighttimeidxold:
                weighttimeidxold = weighttimeidx
                save_current_weights(False,sim.time)
            # flush probes
            for probe in sim.model.probes:
                # except weight probes (flushed in save_current_weights)
                # except error probe which is saved fully in ..._end.shelve
                if probe.label is not None:
                    if 'weights' in probe.label or 'error' in probe.label:
                        break
                del sim._probe_outputs[probe][:]
            ## if time > 1000s, double learning rate
            #if sim.time>1000. and not doubledLearningRate:
            #    changeLearningRate(4.)  # works only if excPESDecayRate = None
            #    doubledLearningRate = True
            # run simulation for tFlush duration
            sim.run(duration,progress_bar=False)

    ###
    ### run the simulation, with flushing for learning simulations ###
    ###
项目:pefile.pypy    作者:cloudtracer    | 项目源码 | 文件源码
def _get_time_times(timer=os.times):
        t = timer()
        return t[0] + t[1]

# Using getrusage(3) is better than clock(3) if available:
# on some systems (e.g. FreeBSD), getrusage has a higher resolution
# Furthermore, on a POSIX system, returns microseconds, which
# wrap around after 36min.
项目:pipenv    作者:pypa    | 项目源码 | 文件源码
def threads(self):
        """Return threads opened by process as a list of
        (id, user_time, system_time) namedtuples representing
        thread id and thread CPU times (user/system).
        On OpenBSD this method requires root access.
        """
        return self._proc.threads()
项目:pipenv    作者:pypa    | 项目源码 | 文件源码
def cpu_times(self):
        """Return a (user, system, children_user, children_system)
        namedtuple representing the accumulated process time, in
        seconds.
        This is similar to os.times() but per-process.
        On OSX and Windows children_user and children_system are
        always set to 0.
        """
        return self._proc.cpu_times()
项目:pipenv    作者:pypa    | 项目源码 | 文件源码
def _cpu_busy_time(times):
    """Given a cpu_time() ntuple calculates the busy CPU time.
    We do so by subtracting all idle CPU times.
    """
    busy = _cpu_tot_time(times)
    busy -= times.idle
    # Linux: "iowait" is time during which the CPU does not do anything
    # (waits for IO to complete). On Linux IO wait is *not* accounted
    # in "idle" time so we subtract it. Htop does the same.
    # References:
    # https://github.com/torvalds/linux/blob/
    #     447976ef4fd09b1be88b316d1a81553f1aa7cd07/kernel/sched/cputime.c#L244
    busy -= getattr(times, "iowait", 0)
    return busy
项目:appbackendapi    作者:codesdk    | 项目源码 | 文件源码
def get_user_cputime(self):
        return os.times()[0]
项目:appbackendapi    作者:codesdk    | 项目源码 | 文件源码
def burn_cpu(self):
        """Consume REQUEST_CPUTIME_SEC core seconds.
        This method consumes REQUEST_CPUTIME_SEC core seconds. If unable to
        complete within REQUEST_TIMEOUT_SEC walltime seconds, it times out and
        terminates the process.
        """
        start_walltime_sec = self.get_walltime()
        start_cputime_sec = self.get_user_cputime()
        while (self.get_user_cputime() <
               start_cputime_sec + REQUEST_CPUTIME_SEC):
            self.busy_wait()
            if (self.get_walltime() >
                    start_walltime_sec + REQUEST_TIMEOUT_SEC):
                sys.exit(1)
项目:ouroboros    作者:pybee    | 项目源码 | 文件源码
def __init__(self, timer=None, bias=None):
        self.timings = {}
        self.cur = None
        self.cmd = ""
        self.c_func_name = ""

        if bias is None:
            bias = self.bias
        self.bias = bias     # Materialize in local dict for lookup speed.

        if not timer:
            self.timer = self.get_time = time.process_time
            self.dispatcher = self.trace_dispatch_i
        else:
            self.timer = timer
            t = self.timer() # test out timer function
            try:
                length = len(t)
            except TypeError:
                self.get_time = timer
                self.dispatcher = self.trace_dispatch_i
            else:
                if length == 2:
                    self.dispatcher = self.trace_dispatch
                else:
                    self.dispatcher = self.trace_dispatch_l
                # This get_time() implementation needs to be defined
                # here to capture the passed-in timer in the parameter
                # list (for performance).  Note that we can't assume
                # the timer() result contains two values in all
                # cases.
                def get_time_timer(timer=timer, sum=sum):
                    return sum(timer())
                self.get_time = get_time_timer
        self.t = self.get_time()
        self.simulate_call('profiler')

    # Heavily optimized dispatch routine for os.times() timer
项目:ndk-python    作者:gittor    | 项目源码 | 文件源码
def _get_time_times(timer=os.times):
        t = timer()
        return t[0] + t[1]

# Using getrusage(3) is better than clock(3) if available:
# on some systems (e.g. FreeBSD), getrusage has a higher resolution
# Furthermore, on a POSIX system, returns microseconds, which
# wrap around after 36min.