Axon: A Synthesizing Superoptimizer for Tensor Programs

2026-06-26T08:52:02Zef8e068be6080ab88fe878c18527fd7c4ee41fff1cdb731482713f60ac1c78af
AI acceleratorsLLMsLeanPCMReRAMRustSMTVeriFastapproximate computingcompilerformal methodsheap buffer overflowhyperdimensional computinginstruction selectionmemory safetyprivacyprogram synthesispurpose-aware programmingseparation logic benchmarksuperoptimizertensor programstheorem provingtilingtypestateuse-after-free

What happened

Collection of recent programming-languages and systems papers (arXiv, 2026-06-26) covering program synthesis and superoptimization for tensor/accelerator kernels (Axon uses SMT over unbounded tensors and empirical kernel search), a typestate system for purpose-aware data usage (PurPL), compiler-driven approximation for hyperdimensional computing (ApproxHDC), mostly-automatic translation of C interpreters to safe Rust that eliminates memory vulnerabilities (Reboot), a core formal model of Go with generics and runtime typing (Welterweight Go), surveys and benchmarks for quantum and auto-formal­

Why it matters

A reviewed impact interpretation has not been published for this record.

Evidence and limitations

Source ID
arxiv_cs_pl
Record identifier
ef8e068be6080ab88fe878c18527fd7c4ee41fff1cdb731482713f60ac1c78af
Enrichment time
2026-06-26T08:52:02Z
AI-assisted enrichment
Yes

This record may overlap with other records. Its enrichment can be incomplete or wrong, and machine assistance was used. Validate consequential decisions against the linked source and your own environment.