<>8088 actually runs for the same clock about 20% slower than the 8086
<>but using significantly fewer glue chips.
<
<Hmm... but it still had the half width bus, right?
You confuse bus width in bit with data rate. the 8088 uses the bus harder
to keep that performance. the 8086 has a log of dead time on the bus
when nothing is happening relevent to the cpu (this is good for DMA).
<Yeah, but it's still a 'downgrade'. But the 386SX was a fairly good
<success, and I take back anything bad I said about it. But once again, th
<386SX didn't give the 386 all it's glory.
Sure it did. it gave you 32bit performance and program execution at far
less power and cost.
<Yeah. I'm not argueing with that. Actually, the 128K's at full clock
<speed, not half, like in a PII, so eventually, you loose an amazingly smal
<amount of performance. When you add in the fact that you can overclock th
the advantage of larger onchip cache is that it's faster as you don't need
external glue logic and the attendant propagation delays (part of the
reason why the PII needs two clock for cache). the difference between a
128k cache and 512k is small anyway unless you have a really intense
application with really large databases.
<Celeron a lot more than a PII (a 333 can go to 450MHz, according to some
<reports, but I don't have that kind of a cooling system. Tropics, and all.
What the point of pushing? In the end it's still wanting for the next
keystroke, mouse click or byte from a slow modem.
Allison
Received on Sun Dec 27 1998 - 21:15:00 GMT