Skip to content
NewA calmer, faster ZeroHour

ElevatorII statically translates x86 binaries to AArch64 ahead of time

Original title: ElevatorII: High-Performance Fully-Static Cross-ISA Binary Translation

arXiv cs.CRby Hongyu Chen ()1 min read
AI summary · x-ai/grok-4.7Summaries by Vani AI

On 11 SPEC CPU 2026 Integer benchmarks without C++ exceptions, it runs at 2.5x native AArch64 time, matching Box64 and beating QEMU by 5x.

ElevatorII is a fully static ahead-of-time binary-to-binary compiler from x86 to AArch64. It uses superset disassembly so every possible jump destination in the original text segment has pre-translated code, then observation-guided path selection adds optimized confirmed paths. The authors report output code 51 times the size of source-compiled code, with only 2-3% hot code in the main ELF section. It also translates binaries obfuscated with virtualization and branch functions.

Key facts
  • No run-time support is required on the target, unlike JIT emulators.
  • Superset disassembly builds a control-flow path for every byte interpretation.
  • Cold code is placed in a .text.unlikely ELF section.
  • The reported benchmarks exclude SPEC programs that use C++ exceptions.
Analyze this story with Vani AIImpact, who is affected and what defenders should do, in a new chat

From the original article

1 min read at arxiv.org

The established method for running software compiled for one instruction-set architecture (ISA) on a different ISA is by emulation in conjunction with just-in-time (JIT) compilation. This requires run-time support on the target that needs to be trusted and whose compilation decisions are hard to predict or reconstruct.

arXiv cs.CR

By Hongyu Chen · arxiv.org

Read the full article at arXiv cs.CR

A short excerpt; the article and its images belong to the publisher. Summary, key facts and assessment above are AI-written from the article and can be wrong.