| Thread Links | Date Links | ||||
|---|---|---|---|---|---|
| Thread Prev | Thread Next | Thread Index | Date Prev | Date Next | Date Index |
Ralph Baker Kearfott wrote:
I'd like to remind people that Ulrich and others have argued in the past that accurate dot product should be implemented in hardware because it is otherwise too slow. On the other hand, for what it's worth, Rump, Oishi et al have developed "almost accurate" dot product algorithms that can be implemented efficiently in current IEEE-754 conforming hardware.
I think that the optimally enclosed dot product (which is what Rump et al. provide) is good enough for all applications of the exact dot product I an aware of. It is much faster in software than an exact dot product.I had therefore asked for providing evidence for applications that really need the exact dot product, but this hasn't generated any response.
Therefore I still maintain that only the optimally enclosed dot product is needed, and recommend to vote accordingly. Arold Neumaier