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/

Zendy Plus

Presents the zaia usage as premium feature

ZAIA

Presents the pdf analysis as premium feature

PDF Analysis

Insights

Presents the summarisation as premium feature

Summarisation

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the access of premium content as premium feature

Premium Content

Zendy Tools

Zendy Open

Building efficient algorithms for data-intensive problems requires deep analysis of data access patterns. Random data access patterns exacerbate this process. In this paper, we discuss accelerating a randomized data-intensive machine learning algorithm using multi-core CPUs and several types of GPUs. A thorough analysis of the algorithm’s data dependencies enabled a 75% reduction in its memory footprint. We created custom compute kernels via code generation to identify the optimal set of data placement and computational optimizations per compute device. An empirical evaluation shows up to 245x speedup compared to an optimized sequential version. Another result from this evaluation is that achieving peak performance does not always match intuition: e.g., depending on the GPU architecture, vectorization may increase or hamper performance.

Accelerating Overlapping Community Detection: Performance Tuning a Stochastic Gradient Markov Chain Monte Carlo Algorithm