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

Sparse matrix-matrix multiplication (SpGEMM) is a critical kernel widelyemployed in machine learning and graph algorithms. However, real-worldmatrices' high sparsity makes SpGEMM memory-intensive. In-situ computing offersthe potential to accelerate memory-intensive applications through highbandwidth and parallelism. Nevertheless, the irregular distribution ofnon-zeros renders SpGEMM a typical unstructured software. In contrast, in-situcomputing platforms follow a fixed calculation manner, making them structuredhardware. The mismatch between unstructured software and structured hardwareleads to sub-optimal performance of current solutions.  In this paper, we propose SPLIM, a novel in-situ computing SpGEMMaccelerator. SPLIM involves two innovations. First, we present a novelcomputation paradigm that converts SpGEMM into structured in-situmultiplication and unstructured accumulation. Second, we develop a uniquecoordinates alignment method utilizing in-situ search operations, effectivelytransforming unstructured accumulation into high parallel searching operations.Our experimental results demonstrate that SPLIM achieves 275.74$\times$performance improvement and 687.19$\times$ energy saving compared to NVIDIA RTXA6000 GPU.

Accelerating Unstructured SpGEMM using Structured In-situ Computing