wrote:wrote:<snip>
A proper 486 will fight you every step of the way.
Hear hear
I've never had any real problems putting together a VLB 486. Try building an XT or 286 with a crusty taiwanese motherboard.
wrote:[...]
Not necessarily. There were a few Socket 4 and Socket 5 boards with VLB slots as well. My understanding is that the VLB implementation was not as good because VLB was closely tied to the 486 architecture.
The issue is that VLB runs at full system speed, supposedly connecting directly to the CPU bus, and was only really reliable at 33MHz - 40 or 50MHz was a stretch, 60 or 66MHz a bridge too far.
I'm not sure how the implementation was on those boards, but basically they either had to run it at full system speed anyway, which would just work with a Pentium 75 (and 50MHz FSB), but not with the old 60/66 (let alone the fact that a Pentium has a 64b bus and VLB is 32b) - or the VLB would have to run at half system speed, bridging it off the PCI bus, in which case it wouldn't have offered any advantage over PCI, and might get a bunch of timing issues as well due to not being designed for asynch operation.
Those early OPTi Pentium chipsets were so awful they pretty much killed off what had been the biggest 486 chipset vendor. By the time they had the at least half-decent Viper-M they were in the wilderness. Apart from a big deal with IBM (most early Pentium Aptiva models had the Viper-M) they hardly managed to sell anything. Still, UMC's Pentium chipsets were just as bad, and they didn't even have the saving grace of supporting VLB 😉
