Tim Prince wrote:
> As far as I know, PowerPC sup****t for 128-bit floating point has always
> been of the 2 doubles 106 to 107 bit mantissa low exponent range
> variety, not the IEEE variety.
Even so, that is sup****t for a higher precision.
But I do see that I was wrong, and the first bit *is* hidden in the
128-bit format, at least according to the latest draft of IEEE-754r.
John Savard
