John, "So educate me. Tell me what the difference is between the way all the different microprocessor companies measure MHz. I claim it's standard, you claim it's not. What's the difference?"
Around the globe, all people measure "MHz" in the same
way - how many ticks/waves it makes per second.
The question is not how to measure, but what and when.
The "difference" is that modern microprocessors contain
blocks that run at different clock speeds. For example,
a significant part of a CPU is Bus Interface Unit,
and it usually runs at input frequency, i.e. 100
or 133MHz. For other internal parts this clock
is multiplied internally by certain number, and it makes
so-called "core clock" - the one which is usually
advertised by marketing. Also, some blocks may
shut the clock down locally after certain period
of inactivity. Now, if a portion of chip that runs
at the advertised speed occupies only a small fraction
of the chip as in P4, what grounds do you have to
claim that the whole chip is rated at that speed?
I have already formulated these problems with "clock speed"
definition:
Message 16448770
If you have problems with what I am talking about,
I am very sorry, I can't help you with the basics.
I also formulated a question you tried to avoid:
"From customer's standpoint, the product metrics
should be measurable. How do you measure the
"clock speed" of P4? Every P4 is fed from [the same] 100MHz
clock. There is no simple measuring procedure
that can distinguish between 1.4GHz P4 and 1.7GHz,
only the manufacturing markings on the top, and
benchmarks.
Tell me, how this is different from AMD markings?"
- Ali
P.S. I still can't get it whether you are kidding
or not with such an elementary question... |
