|
|
Re: higher precision floating point computation
Posted:
Nov 17, 2000 10:42 AM
|
|
In article <darcy.974438365@moa.CS.Berkeley.EDU>, darcy@moa.CS.Berkeley.EDU (Joseph D. Darcy) writes:
>>I'm not sure what you mean by calling D floating a "bug".
>
> Well, presumably if the D format was adequate as a double format, the
> G format of the same size wouldn't be introduced. After all, if
> nothing else think of the linguistic difficulties in specifying
> whether you want to use D or G, which one is "double"?
Inadequate and "a bug" are not synonymous in my lexicon. G floating most
certainly did not supplant D floating on VAX.
Contrary to your expectations, D and G floating were used on the same
machine at the same time. There was a compile time switch to control
which type of double would be used by the compiler and link time settings to
control which run time library would be used. The linguistic difficulties
you imagine were not serious.
John Briggs briggs@eisner.decus.org
|
|
