Re: File I/O

From: wolfgang kern (nowhere_at_nevernet.at)
Date: 01/12/05

  • Next message: Tim Roberts : "Re: Beginner q: plugging into an IRQ handler" 
    Date: Wed, 12 Jan 2005 22:11:20 +0000 (UTC)

    Yes, correct Frank,
     
    | > DAA -method is shortest but very slow.

    | ; nibble masked off
    | cmp al, 0Ah
    | sbb al, 69h
    | das
     
    | shorter? possibly even slower?

    your code (5 bytes,10 cycles) is shorter and faster than
    the DAA-version, but still much slower than 'my' 3..4
    cycles with 8 bytes.

    All this anchient multifunctional ACCU-instructions
    [aaa aad aam aas daa das] which need
    [ 6 6 16 6 8 8 ] vectored clock-cycles on K7,
    (not any better on Intel-CPU's)
    can be considered as extremely slow and are
    just usable if code-size is the main matter.

    __
    wolfgang

  • Next message: Tim Roberts : "Re: Beginner q: plugging into an IRQ handler"

    Relevant Pages

    • Re: Rework [Was: Static vs. Dynamic typing...]
      ... flabbergasted at how tiny the cycles were. ... getting shorter and shorter; and I don't find that it takes any longer ... Object Mentor Inc. ... "The aim of science is not to open the door to infinite wisdom, ...
      (comp.object)
    • Re: Moon hopping around?
      ... how long cycles such as this were learned by observation. ... explain how religions started. ... The average life span may have been shorter, ...
      (sci.physics)
    • Re: lost in upgrade from dapper to edgy
      ... Due mostly to the shorter release cycles, the unstable Ubuntu repos are ... in a much greater disarray than Debian's during some parts of the cycle. ...
      (Ubuntu)
    • Re: File I/O
      ... sbb al, 69h ... possibly even slower? ...
      (comp.lang.asm.x86)
    • Re: from elsewhere, an assembler
      ... sbb al, 69h ... Shorter, and eliminates the conditional jump... ... but "das" is so slow ...
      (alt.lang.asm)