jcallen@Encore.COM (Jerry Callen) (02/09/90)
In article <7356@pdn.paradyne.com> alan@oz.paradyne.com (Alan Lovejoy) writes: >In article <2100@crdos1.crd.ge.COM> davidsen@crdos1.crd.ge.com (bill davidsen) writes: >> Having explained why vendors were interested in the 860, and what it >>offers relative to SPARC, would someone tell me why DG thinks the 88k is >>better than SPARC? Serious question, what has the 88k got that SPARC >>doesn't? > > [stuff deleted] > >Hardware multiply with a latency of only four cycles, and the ability to >start a new multiply thru the pipe each clock cycle. Grrrrrr..... for the most part, I really like the 88K, but the $#@ MUL instruction is totally brain-damaged. It's a 32x32 giving 32 multiply, and it DOESN'T TELL YOU if it overflows! C hackers probably don't care, but this is useless for languages that DO care about overflow (like Ada). I was really delighted that Moto put fixed point multiply on the chip until I discovered this lossage. And then there is DIV. For some unknown reason, the $#@ chip faults if you divide two numbers with unlike signs! The 88open folks recognize the absurdity of this and specify that a BCS compliant OS has to fix this up transparently. Don't get me wrong; I like the chip. But I'm really annoyed that these little stupidities snuck in. -- Jerry Callen jcallen@encore.com (508) 460-0500