math1h3@jane.uh.edu (01/14/90)
What would I like to see in the Mac of the 90's? A better floating point environment, that's what. Here are some obvious improvements I'd like to see: 1. Support for single-precision and double precision arithmetic and transcendental functions in whatever coprocessor is used. It seems terribly wasteful to always do extended precision arithmetic and transcendental functions when the user/programmer might only want or need single precision. How often does one really want 19 digits of accuracy? 2. SANE should recognize when a coprocessor is present and pass through all transcendental function calls. Really, how much more accurate is SANE? It certainly isn't worth the slowdown it produces, not to me it isn't. Apple, please recognize that SOME of us use Macs to do what computers were originally designed for: crunching numbers. The Mac has nice I/O features, but it is still pretty slow at crunching numbers, considering what processors are used. BTW, SOME of us write programs on the Mac for our personal use, or for research. One shouldn't have to be a professional programmer to do that. David Wagner My opinions are completely seperate from my employer's
rcfische@polyslo.CalPoly.EDU (Raymond C. Fischer) (01/18/90)
In article <5338.25af60d1@jane.uh.edu> math1h3@jane.uh.edu writes: > What would I like to see in the Mac of the 90's? A better floating point >environment, that's what. Here are some obvious improvements I'd like to see: > >2. SANE should recognize when a coprocessor is present and pass through all >transcendental function calls. Really, how much more accurate is SANE? It >certainly isn't worth the slowdown it produces, not to me it isn't. Using the 68882 instead of SANE loses 12 bits of precision. The 68882 only computes trancendentals to the accuracy of double precision rather than extended precision. (square root excepted) Ray Fischer rcfische@polyslo.calpoly.edu
ldo@peace.waikato.ac.nz (Lawrence D'Oliveiro) (01/22/90)
In <5338.25af60d1@jane.uh.edu>, math1h3@jane.uh.edu (David Wagner), in a list of wishes for the numeric capabilities of future Macintoshes, asks for "support for single-precision and double precision arithmetic and transcendental functions in whatever coprocessor is used. It seems terribly wasteful to always do extended precision arithmetic and transcendental functions when the user/programmer might only want or need single precision." That's as may be, but if you don't compute the intermediate results to the maximum precision available, the final answers you get will be even less accurate than you're used to. The IEEE 754 standard demands that all arithmetic be carried out to extended precision. Both the numeric coprocessor families for which I have information (68881/68882, 80x87) work this way. Because of this, it's actually *faster* to all the arithmetic in your program in extended precision, and only convert to/from external formats when inputting values, or outputting results.