Re: sym53c875E-0: PCI clock seems too high

Allen Brown (abrown@cv.hp.com)
Tue, 28 Sep 1999 15:25:29 -0700


Paul Flinders wrote:
> First off let's remember how CPU's are manufactured. Typically there's
> no difference between, say, a 433Mhz chip and a 333Mhz chip - the
> likelyhood is that they came off the same production line and the only
> difference is that the faster chip passed a (quite short) set of tests
> at that speed.

A little knowledge is a dangerous thing. It is clear to me
that you have never worked for a semiconductor manufacturer.

Semiconductor processes vary from lot to lot. (A lot is defined
a fixed set of wafers that is processed together. Typically
24 or 48.) This variation is different in magnitude for each
process and for each company. Also the magnitude of the variation
is not something most companies will tell you about. But the
variation may result in parts that vary in speed by a factor of
three to one. This is an extreme number, but it is real world.

A company that produces a standard product in several speed
classes will have a test program that tries the part fast.
If it passes, the part is binned as fast. If not, the part
is retested at a slower speed.

What this means is that that 333MHz CPU that you bought really
is only good for 333MHz. OK, so you may be able to run it
faster. But it won't run at its rated temperature. And its
outputs will be out of spec. And its input timing will be
out of spec. And it may get too hot, which will significantly
shorten its life.

The point is that you are playing with fire. Using chips beyond
their manufacturing specs is stupid. If you choose to do this
at home, then you are being stupid at home. All bug reports
against the kernel, any software, or any hardware, reported by
you should be ignored. (Perhaps they already are being ignored,
for the reasons I list here.)

You can flame against my views if you wish. I will probably
ignore you. But hopefully I will convince someone else not to
follow in your footsteps.

--
Allen Brown

The opinions expressed here are my own and may not represent the views of my employer.

- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.rutgers.edu Please read the FAQ at http://www.tux.org/lkml/