Mathematical Institute (University of Oxford)

Mixed-precision algorithms combine low- and high-precision computations in order to benefit from the performance gains of reduced-precision while retaining good accuracy. In this talk we focus on explicit stabilised Runge-Kutta (ESRK) methods for parabolic PDEs as they are especially amenable to a mixed-precision treatment. However, some of the concepts we present can be extended more generally to Runge-Kutta (RK) methods in general.

Consider the problem $y' = f(t,y)$ and let $u$ be the roundoff unit of the low-precision used. Standard mixed-precision schemes perform all evaluations of $f$ in reduced-precision to improve efficiency. We show that while this approach has many benefits, it harms the convergence order of the method leading to a limiting accuracy of $O(u)$.

In this talk we present a more accurate alternative: a scheme, which we call $q$-order-preserving, that is unaffected by this limiting behaviour. The idea is simple: by using $q$ high-precision evaluations of $f$ we can hope to retain a limiting convergence order of $O(\Delta t^{q})$. However, the practical design of these order-preserving schemes is less straight-forward.

We specifically focus on ESRK schemes as these are low-order schemes that employ a much larger number of stages than dictated by their convergence order so as to maximise stability. As such, these methods require most of the computational effort to be spent for stability rather than for accuracy purposes. We present new $s$-stage order $1$ and $2$ RK-Chebyshev and RK-Legendre methods that are provably full-order preserving. These methods are essentially as cheap as their fully low-precision equivalent and they are as accurate and (almost) as stable as their high-precision counterpart.

--

A link for this talk will be sent to our mailing list a day or two in advance.  If you are not on the list and wish to be sent a link, please contact @email.

Symbol alphabets of n-particle amplitudes in N=4 super-Yang-Mills theory are known to contain certain cluster variables of Gr(4,n) as well as certain algebraic functions of cluster variables. In this talk we suggest an algorithm for computing these symbol alphabets from plabic graphs by solving matrix equations of the form C.Z = 0 to associate functions on Gr(m,n) to parameterizations of certain cells of Gr_+ (k,n) indexed by plabic graphs. For m=4 and n=8 we show that this association precisely reproduces the 18 algebraic symbol letters of the two-loop NMHV eight-point amplitude from four plabic graphs. We further show that it is possible to obtain all rational symbol letters (in fact all cluster variables) by solving C.Z = 0 if one allows C to be an arbitrary cluster parameterization of the top cell of Gr_+ (n-4,n).

A link for this talk will be sent to our mailing list a day or two in advance.  If you are not on the list and wish to be sent a link, please send email to @email.