Content deleted Content added
more general description of the half-iteration exit condition. e.g. matlab uses norm(b-Ax) instead of the less stable norm(s) for deciding. |
|||
Line 83:
:{{math|<var>P<sub>i</sub></var>('''<var>A</var>''') {{=}} <var>P</var><sub><var>i</var>−1</sub>('''<var>A</var>''') − <var>α<sub>i</sub>'''A'''T</var><sub><var>i</var>−1</sub>('''<var>A</var>''')}},
:{{math|<var>T<sub>i</sub></var>('''<var>A</var>''') {{=}} <var>P<sub>i</sub></var>('''<var>A</var>''')
===Derivation of BiCGSTAB from BiCG===
| |||