PACO 2019: 3rd Workshop on Power-Aware Computing

Material Package

Added Since

Include only attachments uploaded after this date

Include*

Everything

Specific sessions

Specific contributions

Specific days

Sessions*

Day II

Day I

Include materials from selected sessions

Contributions*

Automatic selection of GPU sparse triangular solvers based on energy consumption (in session "Day I")

Convolutional Neural Nets for the Run Time and Energy Consumption Estimation of the Sparse Matrix–Vector Product (in session "Day I")

Domain decomposition methods in FreeFEM with ffddm (in session "Day II")

Energy Efficiency of Nonlinear Domain Decomposition Methods (in session "Day II")

Energy-Time Analysis of Heterogeneous Clusters for EEG Classification (in session "Day I")

Evaluating asynchronous Schwarz solvers for Exascale. (in session "Day II")

Exploiting Nested Task-Parallelism in the LU Factorization of Hierarchical Matrices (in session "Day I")

Gingko's load-balancing COO SpMV on NVIDIA and AMD GPU architectures (in session "Day I")

In-application energy measurement on Megware SlideSX systems (in session "Day I")

Iterative Refinement in Three Precisions

Massively Parallel & Low Precision Accelerator Hardware as Trends in HPC - How to use it for large scale simulations allowing high computational, numerical and energy efficiency with application to CFD

Opening

Parallel Algorithms for CP, Tucker, and Tensor Train Decompositions

Parallel multiprecision iterative Krylov subspace solver (in session "Day I")

Parallel solution of large sparse systems by direct and hybrid methods

Revisiting the idea of multiprecision block-Jacobi preconditioning - where do we stand 2 years later? (in session "Day I")

S-Step Enlarged Conjugate Gradient Methods (in session "Day II")

Towards an efficient many-core implementation of the IRKA (in session "Day I")

Unleashing the sptrsv method in FPGAs (in session "Day I")

Include materials from selected contributions

Events scheduled on*

05/11/2019

06/11/2019

Include materials from sessions/contributions scheduled on the selected dates