The Computational Efficacy of Finite Field Arithmetic
<p>We show that there exists an interesting non-uniform model of computational complexity within characteristic-two finite fields. This model regards all problems as families of functions whose domain and co-domain are characteristic-two fields. The model is both a <em>structured</em> and a <em>fully</em> <em>general</em> model of computation.</p><p>We ask if the same is true when the characteristics of the fields are unbounded. We show that this is equivalent to asking whether arithmetic complexity over the prime fields is a fully general measure of complexity.</p><p>We show that this reduces to whether or not a single canonical function is ''easy'' to compute using only modulo <em>p</em> arithmetic.</p><p>We show that the arithmetic complexity of the above function is divided between two other canonical functions, the first to be computed modulo <em>p</em> and the second with modulo p^2 arithmetic.</p><p>We thus have tied the efficacy of finite field arithmetic to specific questions about the arithmetic complexities of some fundamental functions.</p>