34 Flaws - Apple Dual Core (so far)
- Thread starter Anthony
- Start date
dgatwood
is out. Leave a message.
Anthony said:
Only 34? Seen the PowerPC G4 (7450) errata? 77.
Of course, 34 is still worse than the current PowerPC G5 errata in sheer number, but not by all that much. 34 is pretty respectable for a first CPU release.
As for the image, I sort of doubt that the comments on the right were really written by Intel. For example, it says, "Also, any OS developer who codes like this deserves this one."
Anyway, I went through the first 9 and didn't see any that look like they should have a severity above "minor annoyance" except #4, though one of them might affect DOS apps. In that first few, there was one whose severity includes the phrase "could be exploited by a virus" which shouldn't even be possible in a modern OS. Might occur in DOS, but usually there are stack pages at one end of the address space and a heap near the other, with no way for code execution to wrap around the top of the address space to the bottom....
That said, I'm somewhat familiar with the PPC down around that level, but not Intel so much, so maybe I'm underestimating the severity of some of them. I rather think the whole thing is seriously overblown, though.... I very much doubt Intel would be shipping chips with truly serious errata. They have too much at stake with AMD right on their heels. Just my gut reaction.
dgatwood
is out. Leave a message.
Havoc said:
ROTFL. I like the way one hardware person described CPU design to me a few years back. It went something like this: anything that can be worked around in the OS isn't a bug, anything that can be worked around in the compiler that doesn't cause the CPU to wedge and isn't exploitable as a security hole isn't a bug, and anything that can be worked around with a motherboard change isn't a bug.