Group PJ_TIMESTAMP

group PJ_TIMESTAMP

PJLIB provides High Resolution Timestamp API to access highest resolution timestamp value provided by the platform. The API is usefull to measure precise elapsed time, and can be used in applications such as profiling.

The timestamp value is represented in cycles, and can be related to normal time (in seconds or sub-seconds) using various functions provided.

Examples

For examples, please see:

Functions

pj_status_t pj_gettickcount(pj_time_val *tv)

Get monotonic time since some unspecified starting point.

Parameters:

tv – Variable to store the result.

Returns:

PJ_SUCCESS if successful.

pj_status_t pj_get_timestamp(pj_timestamp *ts)

Acquire high resolution timer value. The time value are stored in cycles.

See also

pj_get_timestamp_freq().

Parameters:

ts – High resolution timer value.

Returns:

PJ_SUCCESS or the appropriate error code.

pj_status_t pj_get_timestamp_freq(pj_timestamp *freq)

Get high resolution timer frequency, in cycles per second.

Parameters:

freq – Timer frequency, in cycles per second.

Returns:

PJ_SUCCESS or the appropriate error code.

void pj_set_timestamp32(pj_timestamp *t, pj_uint32_t hi, pj_uint32_t lo)

Set timestamp from 32bit values.

Parameters:

  • t – The timestamp to be set.

  • hi – The high 32bit part.

  • lo – The low 32bit part.

int pj_cmp_timestamp(const pj_timestamp *t1, const pj_timestamp *t2)

Compare timestamp t1 and t2.

Parameters:

  • t1 – t1.

  • t2 – t2.

Returns:

-1 if (t1 < t2), 1 if (t1 > t2), or 0 if (t1 == t2)

void pj_add_timestamp(pj_timestamp *t1, const pj_timestamp *t2)

Add timestamp t2 to t1.

Parameters:

  • t1 – t1.

  • t2 – t2.

void pj_add_timestamp32(pj_timestamp *t1, pj_uint32_t t2)

Add timestamp t2 to t1.

Parameters:

  • t1 – t1.

  • t2 – t2.

void pj_sub_timestamp(pj_timestamp *t1, const pj_timestamp *t2)

Substract timestamp t2 from t1.

Parameters:

  • t1 – t1.

  • t2 – t2.

void pj_sub_timestamp32(pj_timestamp *t1, pj_uint32_t t2)

Substract timestamp t2 from t1.

Parameters:

  • t1 – t1.

  • t2 – t2.

pj_int32_t pj_timestamp_diff32(const pj_timestamp *t1, const pj_timestamp *t2)

Get the timestamp difference between t2 and t1 (that is t2 minus t1), and return a 32bit signed integer difference.

pj_time_val pj_elapsed_time(const pj_timestamp *start, const pj_timestamp *stop)

Calculate the elapsed time, and store it in pj_time_val. This function calculates the elapsed time using highest precision calculation that is available for current platform, considering whether floating point or 64-bit precision arithmetic is available. For maximum portability, application should prefer to use this function rather than calculating the elapsed time by itself.

See also

pj_elapsed_usec(), pj_elapsed_cycle(), pj_elapsed_nanosec()

Parameters:

  • start – The starting timestamp.

  • stop – The end timestamp.

Returns:

Elapsed time as pj_time_val.

pj_uint32_t pj_elapsed_msec(const pj_timestamp *start, const pj_timestamp *stop)

Calculate the elapsed time as 32-bit miliseconds. This function calculates the elapsed time using highest precision calculation that is available for current platform, considering whether floating point or 64-bit precision arithmetic is available. For maximum portability, application should prefer to use this function rather than calculating the elapsed time by itself.

See also

pj_elapsed_time(), pj_elapsed_cycle(), pj_elapsed_nanosec()

Parameters:

  • start – The starting timestamp.

  • stop – The end timestamp.

Returns:

Elapsed time in milisecond.

pj_uint64_t pj_elapsed_msec64(const pj_timestamp *start, const pj_timestamp *stop)

Variant of pj_elapsed_msec() which returns 64bit value.

pj_uint32_t pj_elapsed_usec(const pj_timestamp *start, const pj_timestamp *stop)

Calculate the elapsed time in 32-bit microseconds. This function calculates the elapsed time using highest precision calculation that is available for current platform, considering whether floating point or 64-bit precision arithmetic is available. For maximum portability, application should prefer to use this function rather than calculating the elapsed time by itself.

See also

pj_elapsed_time(), pj_elapsed_cycle(), pj_elapsed_nanosec()

Parameters:

  • start – The starting timestamp.

  • stop – The end timestamp.

Returns:

Elapsed time in microsecond.

pj_uint32_t pj_elapsed_nanosec(const pj_timestamp *start, const pj_timestamp *stop)

Calculate the elapsed time in 32-bit nanoseconds. This function calculates the elapsed time using highest precision calculation that is available for current platform, considering whether floating point or 64-bit precision arithmetic is available. For maximum portability, application should prefer to use this function rather than calculating the elapsed time by itself.

See also

pj_elapsed_time(), pj_elapsed_cycle(), pj_elapsed_usec()

Parameters:

  • start – The starting timestamp.

  • stop – The end timestamp.

Returns:

Elapsed time in nanoseconds.

pj_uint32_t pj_elapsed_cycle(const pj_timestamp *start, const pj_timestamp *stop)

Calculate the elapsed time in 32-bit cycles. This function calculates the elapsed time using highest precision calculation that is available for current platform, considering whether floating point or 64-bit precision arithmetic is available. For maximum portability, application should prefer to use this function rather than calculating the elapsed time by itself.

See also

pj_elapsed_usec(), pj_elapsed_time(), pj_elapsed_nanosec()

Parameters:

  • start – The starting timestamp.

  • stop – The end timestamp.

Returns:

Elapsed time in cycles.