English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

Leveraging spatial sparsity has become a popular approach to accelerate 3Dcomputer graphics applications. Spatially sparse data structures and efficientsparse kernels (such as parallel stencil operations on active voxels), are keyto achieve high performance. Existing work focuses on improving performancewithin a single sparse computational kernel. We show that a system that looksbeyond a single kernel, plus additional domain-specific sparse data structureanalysis, opens up exciting new space for optimizing sparse computations.Specifically, we propose a domain-specific data-flow graph model of imperativeand sparse computation programs, which describes kernel relationships andenables easy analysis and optimization. Combined with an asynchronous executionengine that exposes a wide window of kernels, the inter-kernel optimizer canthen perform effective sparse computation optimizations, such as eliminatingunnecessary voxel list generations and removing voxel activation checks. Thesedomain-specific optimizations further make way for classical general-purposeoptimizations that are originally challenging to directly apply to computationswith sparse data structures. Without any computational code modification, ournew system leads to $4.02\times$ fewer kernel launches and $1.87\times$ speedup on our GPU benchmarks, including computations on Eulerian grids, Lagrangianparticles, meshes, and automatic differentiation.

AsyncTaichi: On-the-fly Inter-kernel Optimizations for Imperative and Spatially Sparse Programming