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

MLIR is an emerging compiler infrastructure for modern hardware, but existingprograms cannot take advantage of MLIR's high-performance compilation if theyare described in lower-level general purpose languages. Consequently, to avoidprograms needing to be rewritten manually, this has led to efforts toautomatically raise lower-level to higher-level dialects in MLIR. However,current methods rely on manually-defined raising rules, which limit theirapplicability and make them challenging to maintain as MLIR dialects evolve.  We present mlirSynth -- a novel approach which translates programs fromlower-level MLIR dialects to high-level ones without manually defined rules.Instead, it uses available dialect definitions to construct a program space andsearches it effectively using type constraints and equivalences. We demonstrateits effectiveness \revi{by raising C programs} to two distinct high-level MLIRdialects, which enables us to use existing high-level dialect specificcompilation flows. On Polybench, we show a greater coverage than previousapproaches, resulting in geomean speedups of 2.5x (Intel) and 3.4x (AMD) overstate-of-the-art compilation flows for the C programming language. mlirSynthalso enables retargetability to domain-specific accelerators, resulting in ageomean speedup of 21.6x on a TPU.

mlirSynth: Automatic, Retargetable Program Raising in Multi-Level IR using Program Synthesis