Reproduce “AMARA: Compiler-Driven Accelerator Architecture Design for Graphical Probabilistic Sampling” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Zhan Chen, Yihan Fu, Wenshuo Yue, Daijing Shi, Chenyun Yin, Mingrui Wu, Xiangyu Zhang, Yun Liang and 1 more
AMARA compiles discrete, component-wise Markov Chain Monte Carlo models into static programs for a programmable accelerator. Its intermediate representation separates local update computations from dependencies that prevent concurrent updates, enabling compiler-managed scheduling and label-memory reuse. In cycle-accurate RTL simulation, it reports 6.1× average throughput over JAX-GPU on irregular workloads; this is not single-chain convergence speedup, and topology changes require recompilation.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.