• Re: The ultimate measure of a correct simulation

    From Mikko@21:1/5 to immibis on Thu Jan 18 08:53:44 2024
    On 2024-01-18 06:38:10 +0000, immibis said:

    On 1/18/24 05:35, olcott wrote:
    *This is true on the basis of the meaning of its words*

    The ultimate measure [of a correct simulation] is the correct x86
    emulation of the x86 instructions in the order that they are specified.

    The alternative is incorrectly emulating the x86 instructions in some
    other order than they are specified.


    If an instruction specifies to call a function and the simulation does something other than simulating the calling of the function, is it
    correct?

    I would accept as correct if it simulates the effects of the call ( the
    return value and side effects) and continues the simulation at the
    instruction were the call would return.

    If I had to construct a fake halt decider I would do an incorrect
    simulation where the simulated call H(D,D) returns the wrong value.

    Mikko

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)