arXiv subjects

Ruijie Gao

Publications and source records attributed to Ruijie Gao.

4 recordsLinked to original sources

arXiv2026

Rethinking Agentic Kernel Generation for Emerging Accelerators

Emerging accelerators often lack mature compiler backends, motivating neural agents that generate and repair kernels from architectural documentation and simulator feedback. This approach repeatedly reconstructs workload-invariant machine semantics--including instruction behavior, legality constraints, synchronization rules, and memory protocols--for every workload. We argue that these semantics should be compiled once into a persistent symbolic artifact, while neural reasoning should focus on workload-dependent mapping decisions. We present Zomboss, a compiler-mediated agentic framework for kernel generation that places neural search within a verified compiler boundary. Zomboss compiles machine semantics and legality constraints into a reusable mapping interface, then uses a neural agent to optimize workload-dependent decisions within the validated mapping space. Across 20 Gemmini and 36 PLENA workload instances, Zomboss returns a correct verified kernel on all 56 instances. Relative to the compiler default, Zomboss achieves geometric-mean speedups of $3.34\times$ on Gemmini and $1.10\times$ on PLENA. Relative to direct agentic generation, it reduces inference tokens by 71.2% on Gemmini and 54.2% on PLENA. These results show that a compiler-defined symbolic interface turns native kernel synthesis into verified design-space exploration: compiler infrastructure preserves legality and correctness, while neural guidance improves workload-specific performance with lower search cost and complete coverage.

cs.AR↗

arXiv2026

TensorLift: Automatic Extraction of Tensor-Level ISA Semantics from Accelerator RTL via MLIR Semantic Lifting

Most proposed tensor accelerators lack well-documented ISAs and compiler backends, and are exercised only through hand-written kernels covering a handful of operators. Recent work (TAIDL, ACT) shows that a tensor-level ISA specification is enough to generate complete software stacks automatically. Writing that specification, however, remains a manual, expert-driven process. We present TensorLift, the first end-to-end MLIR-based pipeline that lifts RTL-extracted accelerator semantics to TAIDL-like tensor ISA specifications. Building on prior architecture-level model extraction that yields bit-level IR, an 8-pass MLIR pipeline progressively recovers tensor structure (MAC idioms, saturation semantics, multi-dimensional buffer organizations, and layout transformations) and emits specifications the ACT ecosystem consumes directly. On Gemmini and VTA, TensorLift captures every hardware instruction the two designs decode across 156 MLIR files, reducing the extracted bit-level MLIR by 24.8% and 41.2% overall and by up to 92.9% on a processing element, with controller modules retaining irreducible control logic. It recovers hardware features the hand-written reference omits (multi-bank DMA configuration, pooling, and im2col) and lifts VTA unmodified. Core compute and data-movement semantics are proven equivalent to the RTL-extracted model by Z3 SMT, the remainder validated against golden simulator data. Fed into ACT, the extracted specification yields a compiler backend at parity with hand-written Gemmini kernels (1.014x geometric mean), giving an automated path from RTL to a working software stack.

cs.AR↗

arXiv2026

Evidence of Long-Lived Powerful Gyrosynchrotron Radio Emission in the Close Binary FF UMa

RS Canum Venaticorum (RS CVn) close binaries, characterized by tidal locking, rapid rotations, and strong magnetic fields, are ideal laboratories for high-resolution radio observations to probe emission processes, magnetic field configurations, and interaction activity. Despite their importance, only a few RS CVn sources have been explored by polarimetric observations of very long baseline interferometry (VLBI). To expand the effort, we have analyzed the existing Very Long Baseline Array (VLBA) astrometric data for the RS CVn binary FF Ursae Majoris (FF UMa). In the 5GHz VLBA experiments conducted between 2021 and 2024, both total intensity and circularly polarized emission were clearly detected at six of seven epochs. The consistently high brightness temperatures (10^7 K) and the moderate fractional circular polarization (10%-30%) over about three years indicate that the radio emission is mainly produced by gyrosynchrotron radiation from mildly relativistic electrons in the highly-ordered magnetic field. The radio luminosities are also comparable to those of previously studied powerful RS CVn binaries and show a significant anti-correlation with fractional circular polarization. A mean centroid offset of 13.4 +/- 3.1 solar radii between the Stokes I and V emission was found across multiple epochs, indicating a possible additional contribution from the secondary star via a magnetically active corona, a giant magnetic loop, or significant interaction activity with the primary star in the quiescent state.

astro-ph.SR↗

arXiv2024

EDGE-LLM: Enabling Efficient Large Language Model Adaptation on Edge Devices via Layerwise Unified Compression and Adaptive Layer Tuning and Voting

Efficient adaption of large language models (LLMs) on edge devices is essential for applications requiring continuous and privacy-preserving adaptation and inference. However, existing tuning techniques fall short because of the high computation and memory overheads. To this end, we introduce a computation- and memory-efficient LLM tuning framework, called Edge-LLM, to facilitate affordable and effective LLM adaptation on edge devices. Specifically, Edge-LLM features three core components: (1) a layer-wise unified compression (LUC) technique to reduce the computation overhead by generating layer-wise pruning sparsity and quantization bit-width policies, (2) an adaptive layer tuning and voting scheme to reduce the memory overhead by reducing the backpropagation depth, and (3) a complementary hardware scheduling strategy to handle the irregular computation patterns introduced by LUC and adaptive layer tuning, thereby achieving efficient computation and data movements. Extensive experiments demonstrate that Edge-LLM achieves a 2.92x speed up and a 4x memory overhead reduction as compared to vanilla tuning methods with comparable task accuracy. Our code is available at https://github.com/GATECH-EIC/Edge-LLM

cs.LG↗