Queries regarding the library and performance statistics

[Kannamraju P]

Hi All,

  

  As part of evaluating different timezone utilities for a C++ app ran int CCTZ  , looking for some info on  .

1. Is there a standard routine which can take CIVIL time , zone and convert to epoch.

2. Is the Timezone information seeded in the code or is loaded from some file . If its loaded from some file 

does it get loaded every time we invoke an conversion API .

3. Are there any performance statistics for the library API's 

Any help is really appreciated.

Thanks,

Raju. 

   

[Bradley White]

Hi Raju,

On Thursday, August 24, 2017 at 8:48:00 AM UTC-4, Kannamraju P wrote:

1. Is there a standard routine which can take CIVIL time , zone and convert to epoch.

Yes.  Indeed conversion between civil times and absolute times via a time zone is the primary purpose of the library.

In your case, to convert a civil time in a time zone to an absolute time, you'd use something along the lines of:

  cctz::time_zone tz = ...;

  cctz::civil_second cs = ...;

  cctz::time_point<cctz::sys_seconds> tp = cctz::convert(cs, tz);

If you need to fully account for skipped/repeated civil times, look at cctz::time_zone::lookup(const cctz::civil_second&).

 

2. Is the Timezone information seeded in the code or is loaded from some file . If its loaded from some file 

does it get loaded every time we invoke an conversion API .

It is loaded from a file by default, but you can also override how a time-zone name maps to its zoneinfo data.

The zoneinfo data for a named zone is loaded once (for the life of the program), on first reference.

 

3. Are there any performance statistics for the library API's

The src/benchmarks.cc file has a set of micro benchmarks.  You should be able to run them on your own platform (although we only support it using bazel).

For what it's worth, I currently get these numbers:

Run on host (32 X 2600 MHz CPUs); 2017-08-24T22:31:48.73777955-07:00

CPU: Intel Sandybridge with HyperThreading (16 cores) dL1:32KB dL2:256KB dL3:20MB

Benchmark                                    Time(ns)        CPU(ns)     Iterations

-----------------------------------------------------------------------------------

BM_Difference_Days                                 19.8           19.8     35230223

BM_Step_Days                                       14.7           14.6     47679141

BM_Zone_LoadUTCTimeZoneFirst                       33.5           33.5     20559067

BM_Zone_LoadUTCTimeZoneLast                        35.8           35.8     19354916

BM_Zone_LoadTimeZoneFirst                      408455         407232           1704

BM_Zone_LoadTimeZoneCached                         47.6           47.6     14628558

BM_Zone_LoadLocalTimeZoneCached                   137            137        5044375

BM_Zone_LoadAllTimeZonesFirst                  159723         159083           4447

BM_Zone_LoadAllTimeZonesCached                     63.2           63.1     10000000

BM_Zone_TimeZoneImplGetImplicit                     2.81           2.81   250544964

BM_Zone_TimeZoneImplGetExplicit                     1.97           1.97   357661845

BM_Zone_UTCTimeZone                                 2.52           2.52   278138596

BM_Time_ToDateTime_BaseTime                        54.4           54.3     10000000

BM_Time_ToDateTime_Libc                            83.7           83.5      8548525

BM_Time_ToDateTimeUTC_BaseTime                     46.1           46.1     14820691

BM_Time_ToDateTimeUTC_Libc                         45.9           45.9     15042716

BM_Time_FromDateTime_BaseTime                      52.9           52.7     10000000

BM_Time_FromDateTime_Libc                         533            532        1000000

BM_Time_FromDateTimeUTC_BaseTime                   40.6           40.5     17221911

BM_Time_FromDateTimeDay0_BaseTime                  80.9           80.7      8760811

BM_Time_FromDateTimeDay0_Libc                     614            614        1000000

BM_Format_FormatTime/0                            565            564        1000000  %a, %d %b %Y %H:%M:%S %z

BM_Format_FormatTime/1                            449            448        1564818  %d %b %Y %H:%M:%S %z

BM_Format_FormatTime/2                            418            417        1641786  %Y-%m-%dT%H:%M:%E*S%Ez

BM_Format_FormatTime/3                            373            372        1864626  %Y-%m-%dT%H:%M:%S%Ez

BM_Format_FormatTime/4                            340            340        2049963  %Y-%m-%dT%H:%M:%S

BM_Format_FormatTime/5                            232            231        2970344  %Y-%m-%d

BM_Format_ParseTime/0                            1277           1275         560828  %a, %d %b %Y %H:%M:%S %z

BM_Format_ParseTime/1                             830            828         834518  %d %b %Y %H:%M:%S %z

BM_Format_ParseTime/2                             354            353        1969853  %Y-%m-%dT%H:%M:%E*S%Ez

BM_Format_ParseTime/3                             328            328        2131377  %Y-%m-%dT%H:%M:%S%Ez

BM_Format_ParseTime/4                             232            231        3009379  %Y-%m-%dT%H:%M:%S

BM_Format_ParseTime/5                             155            154        4522413  %Y-%m-%d

[Note:  In the To/FromDateTime comparisons with Libc above, the "BaseTime" numbers refer to CCTZ.  I'll fix those names.]

Bradley