🚀 go-pugleaf

andy the pugh <A.C.Pugh@shef.ac.uk> wrote:

> In article <1d9nnii.sim80gcftt2cN@p38.nas1.is2.u-net.net>, the moving
> finger of real-address-in-sig@rhu.barb.foobar wrote...
> >
> >andy the pugh <A.C.Pugh@shef.ac.uk> wrote:
> 
> >> I tried 240MHz, but it wouldn't boot. Anyone any idea why? 
> >
> >Yes, it's 50% beyond spec.  The only guaranteed speed is 160MHz.
> >Anything faster is down to luck.
> 
> Arguably as the chip was designed to go up to 300MHz in the same process
> size Apple are running them 50% below spec. 

Not really: the maximum design speed for that IC is, you say, 300MHz.
What happens in real life is that the ICs are built, and then tested.
Those that pass the test for 160MHz operation are sold as working at
that speed, and those that pass the test for higher speed operation are
sold as working at that speed.  This is all approximate, since I expect
that faster ICs are sold marked and priced as lower-speed ICs if the
maker has a large order to meet (highly likely, in this case).

But...  The point is that the entire system is built to run at a given
speed.  If you speed up even only one part of it (by using a different
clock multiplier), you run the risk of things failing to work properly.
Not to put too fine a point on it: computers are complicated beasties.

(And just to be irritatingly nit-picking: 160Mhz rather than 300MHz is
46 2/3% under.  I know it doesn't matter.)

> Whilst I see your point, it seems odd that the problem is failure to boot
> at the higher clock speed rather than, say, overheating and crashing some
> time later.

It seems quite sensible to me.  No decently-designed IC should overheat
if used sensibly.  What's happening is that you're trying to run the IC
faster than it can run, so it doesn't work properly and the Mac doesn't
boot.

> Back to the original subject, I suspect the 4400 could be made much quieter
> by clipping a little CPU-mount fan to the existing heatsink and unplugging
> the one in the front of the case. 

Quite possibly.  It's possible that the whole thing could suffer
overheating if the fan doesn't ensure air circulation from inside the
case to the outside world.  My own solution to the fan noise problem was
to sit my Performa 475 on a square of carpet tile with damps down the
annoying high frequency components of the fan noise rather a lot - the
fan sits in the middle of the box, blowing air down onto the desk.  I
did something similar but rather more elaborate to quieten my noisy
external HDD.

Rowland.

-- 
Remove the animal for my email address:   rebecca@astrid.dog.u-net.com
Sorry - the spam got to me.                       PGP pub key A680B89D
UK biker?                    Join MAG and help keep bureaucracy at bay
                          http://dredd.meng.ucl.ac.uk/www/mag/mag.html