P4 chipset bug still crawling
Saw this over at Arstechnica:
The flaw in the P4 chipset's PCI handling hasn't gone away yet. The bug, which was originally uncovered before the launch of the P4 and
ultimately delayed the rollout of the CPU, is apparently not considered serious enough to warrant immediate attention. Note that the bug is
in the chipset, and not the P4 CPU itself (as some sites mistakenly reported). This bug seems to be minor, but I get a sense that we're not
getting the full story. The official explanation is that video performance via the PCI bus (i.e., not via AGP) is flawed. As a result, the
"solution" is simply not to put a video card in a PCI slot, which means that dual-monitor solutions need to be sought in a 2-in-1 AGP
solution, such as the Matrox G450. But doesn't stand to reason that other high-bandwidth devices may suffer problems as well?
PCI-based FireWire cards, DVD decoders, etc.?
