How many Intel x86 CISC instructions needed to decode NTP time field (ignoring complexities of decoding full packet, re: NTP broadcasting)? - NTP

This is a discussion on How many Intel x86 CISC instructions needed to decode NTP time field (ignoring complexities of decoding full packet, re: NTP broadcasting)? - NTP ; How many Intel x86 CISC instructions needed to decode NTP time field (ignoring complexities of decoding full packet, re: NTP broadcasting)? From my understanding most NTP code is nearly exclusively compiled C++, efficient -- but not quite assembler. For RISC ...

+ Reply to Thread
Results 1 to 3 of 3

Thread: How many Intel x86 CISC instructions needed to decode NTP time field (ignoring complexities of decoding full packet, re: NTP broadcasting)?

  1. How many Intel x86 CISC instructions needed to decode NTP time field (ignoring complexities of decoding full packet, re: NTP broadcasting)?

    How many Intel x86 CISC instructions needed to decode NTP time field
    (ignoring complexities of decoding full packet, re: NTP broadcasting)?

    From my understanding most NTP code is nearly exclusively compiled C++,
    efficient -- but not quite assembler.

    For RISC microcontroller applications and other CPU designs (like DSPs) --
    NTP decoding may have to be done via assembler.




  2. Re: How many Intel x86 CISC instructions needed to decode NTP time

    Max Power wrote:

    > How many Intel x86 CISC instructions needed to decode NTP time field
    > (ignoring complexities of decoding full packet, re: NTP broadcasting)?
    >

    Write the code and count! IIRC, it's 32 bits of seconds since 1 January
    1970, and 32 bits of fractional seconds.

    > From my understanding most NTP code is nearly exclusively compiled C++,
    > efficient -- but not quite assembler.


    C not C++

    >
    > For RISC microcontroller applications and other CPU designs (like DSPs) --
    > NTP decoding may have to be done via assembler.
    >


    I believe that cross-compilers are available for many microcontrollers.


  3. Re: How many Intel x86 CISC instructions needed to decode NTP time

    In article , Max Power wrote:

    > How many Intel x86 CISC instructions needed to decode NTP time field
    > (ignoring complexities of decoding full packet, re: NTP broadcasting)?


    I wouldn't have thought any were required as the time is already in a form
    that is generally useful, business contexts (science contexts may need
    leap seconds to be preserved). In fact, many applications have to convert
    to something like the NTP format before they do any calculations on dates.

    Obviously some work is needed to convert to typical human time
    representations, but without knowing what you want to convert to,
    it's difficult to say. Modern operating systems all come with libraries
    that will do this sort of conversion for whole seconds.

    I don't understand the broadcasting reference in the subject.

+ Reply to Thread