from The Register
Posted 28/09/99 12:05pm by Pete Sherriff
Cam in no 820, your time is up
Oh dear, it's almost enough to make you feel sorry for poor old Chipzilla. Betting the farm on the whizz-bang i820 chipset and Rambus is starting to look like being a really bad move for the Great Satan of Chips.
The inability of the 820 to run correctly on mobos with three memory slots is embarrassing enough, but those with longer memories will recall that back in April The Register ran the very first 'What the Hell is...' feature, explaining what Camino was all about.
Even then, Camino was late.
The original spec for the 820 mentions not three, but four memory slots, so it looks very much as if Intel discovered a problem with four slot configurations months ago and quietly cut back the spec to just three, hoping the problem would go away.
Now Chipsetzilla is whimpering on about a measly two memory slots -- and that just isn't enough for the high-end systems the 820 was supposed to support.
There are a number of fundamental differences between the way Camino works (or rather, doesn't work) and conventional chipsets like the trusty BX440, which could explain where the problems lie.
Camino is based on a hub architecture with a Memory Controller Hub (MCH) at its heart. The MCH in turn talks to the CPU through the host bus (at 133MHz), the graphics subsystem through the AGP bus (AGP 4X) and the memory
through Rambus. Everything else is handled by the I/O Controller Hub (ICH) which has a direct link to the MCH.
Unlike regular memory, you need to fit continuity modules in empty memory slots because Rambus daisy chains its memory a bit like SCSI devices. This was supposed to mean that clocking the memory was simpler and was also electrically superior, offering active power management of individual memory modules.
Electrically superior it may be, but if it doesn't work reliably, it's totally useless. It looks as if Intel was overreaching itself by trying to introduce so many new techniques simultaneously. ®
theregister.co.uk |
