Abstract
Local roundoff error in floating point addition is characterized and its mean and variance are approximated. These results are applied to finding distributions for the roundoff error accumulated in sums and long inner products. Theorems are stated which resolve some open questions.
| Original language | English (US) |
|---|---|
| Pages | 100-105 |
| Number of pages | 6 |
| State | Published - 1981 |
| Event | Proc - Symp on Comput Arith, 5th - Ann Arbor, Mich Duration: May 18 1981 → May 19 1981 |
Other
| Other | Proc - Symp on Comput Arith, 5th |
|---|---|
| City | Ann Arbor, Mich |
| Period | 5/18/81 → 5/19/81 |
All Science Journal Classification (ASJC) codes
- General Engineering