[Prev][Next][Index][Thread]

Re: 68040 clock doubling



>>>>> "Matthew" == Matthew  <Matthew@ntr.net> writes:
In article <Matthew-1702962219030001@rmta036.ntr.net> Matthew@ntr.net (Matthew) writes:


    Matthew> No. As I understand it, Apple's marketing department
    Matthew> decided to response to what had been standard practice in
    Matthew> the WinTel world; that is, quoting the processor speed,
    Matthew> rather than the bus speed. Apple had been using the more
    Matthew> conservative rating, but decided to change to what had
    Matthew> become an industry standard method of quoting
    Matthew> speed. Thus, consumers that had been comparing, say, a
    Matthew> 33mhz 68040 to a 66mhz 80486 and concluding that the 486
    Matthew> was faster & therefor better ('taint so,) would be more
    Matthew> inclined to compare Apples & oh, say, prunes and thus
    Matthew> arrive at a more enlightened conclusion.

You are correct in that the PPC upgrade speed explanation is incorrect.

However, A 66 MHz 80486 *is* faster than a 33 MHz 68040, and Motorola,
(makers of the 68040) themselves call it a 33 MHz part.  It is only
because it *is* so much faster that we can get a 75 MHz DX4 to run
68040 code at about the same speed that a 25 MHz 68040 can -- after
all, there is a noticable amount of overhead in Syn68k.  If you were
to call the 25 MHz 68040 a 50 MHz CPU, then Syn68k would look better
than it is (we could claim that a 75 MHz DX4 runs code approximately
as fast as a 50 MHz 68040).  It is because we are more honest than we
are opportunistic that we don't join the people who retroactively
redesignated the speeds of various 68040 CPUs.

--Cliff
ctm@ardi.com

    Matthew> Here's a quote from the Executor FAQ.

    Matthew> " Question 1.8. How fast is Executor?

    Matthew> Executor converts mc680x0 instructions into 80x86
    Matthew> instructions and then runs the new instructions. There is
    Matthew> some overhead associated with this process, but for cpu
    Matthew> intensive tasks, a 75 MHz 486DX4 will run approximately
    Matthew> as quickly as a 25 MHz 68040. NOTE: Lately some people
    Matthew> have begun calling 25 MHz 68040s "50 MHz 68040s", but
    Matthew> we're not using that trickery in our description. The
    Matthew> paper /pub/SynPaper available on ftp.ardi.com describes
    Matthew> how we can run mc68040 code so quickly on an 80x86.
    Matthew> SynPaper compares a few different systems and shows that
    Matthew> a 90 MHz Pentium runs almost as fast as a 50 MHz 68040. "



    Matthew> In article
    Matthew> <edkemp-1602961304180001@ekemp.consult.csc.com>,
    Matthew> edkemp@tiac.net (Eric Kemp) wrote:

    -> > > Now, when Apple decided to start calling its, for example,
    -> powerbooks > "33/66-MHz" PowerBooks, I just had to laugh. This
    -> shows just how > desparate they are. The 68040 does indeed use
    -> two different clocks signals.  > This is an implementation
    -> choice that does not necessarily have > anything to do with
    -> exactly how the internal processor core operates.
    -> 
    -> While I agree that this is misleading, I believe it was done to
    -> indicate the clock speed if the machine is upgraded to a
    -> PowerPC chip.  The Quadra is used to own was 33 Mhz base, 66Mhz
    -> with upgrade.  I would not call this an act desperation.
    -> Unfortunately, the general public views megahertz as the
    -> ubiquitous definition of speed (like MPH) when it is accurate
    -> only in comparison to the same processor.


References: