(j3.2006) Liaison to IFIP WG 2.5

keith bierman khbkhb
Tue Aug 21 22:51:30 EDT 2007


On Aug 21, 2007, at 6:30 PM, Van Snyder wrote:

>
> This is pretty simple for add and subtract, but for multiply there are
> nine cases, and for divide when the divisor doesn't include zero there
> are six.  When the divisor includes zero you just produce NaN (or  
> should
> it be [-Inf,+Inf]?).  So for multiply you'd set the rounding mode down
> and then execute a 9-way where/elsewhere block, then set it up and
> execute another where/elsewhere block.  Maybe speculative execution  
> has
> gotten good enough that the branches don't kill the pipeline.

Branches remain a BadThing for pipelines. But with enough operations  
in flight sometimes one can extract useful work between tapping the  
pipeline. Or with enough cores one could separate out and merge ...  
but one would have to have intimate knowledge of the semantics to do  
efficiently.

---
Keith H. Bierman   khbkhb at gmail.com
<speaking for myself*> Copyright 2007







More information about the J3 mailing list