Searcharxiv⌕ Search

arXiv subjects

Daniel Pickard

Publications and source records attributed to Daniel Pickard.

2 recordsLinked to original sources

Learning to Select Sparse Linear Solvers with Convolutional Neural Networks

We investigate convolutional-neural-network (CNN) based automatic solver selection for sparse linear systems, a problem of practical importance where solver performance varies dramatically with matrix sparsity structure. A CNN is trained to regress log timing ratios between candidate solvers using spy plot images of matrix nonzero structure as input, achieving strong prediction accuracy on both held-out SuiteSparse matrices and real-world matrices from high-performance simulation codes.

cs.LG↗

Nonlocal spin transport in single walled carbon nanotube networks

Spin transport in carbon-based materials has stimulated much interest due to their ballistic conductance and a long phase coherence length. While much research has been conducted on individual carbon nanotubes, current growth and placement techniques are incompatible with large-scale fabrication. Here we report nonlocal spin injection and detection in single wall carbon nanotube networks. We observe spin transport over a distance of 1 um, and extract a spin diffusion length of 1.6 - 2.4 um with an injected spin polarization from CoFe into nanotube network of 18 - 41%. Our observations demonstrate that spin transport is possible in carbon nanotube networks due to the formation of natural tunnel barriers between nanotubes and metallic contacts.

cond-mat.mtrl-sci↗