paper-with-me

홈 › Papers

KForge: Program Synthesis for Diverse AI Hardware Accelerators

2025-11-17 · Taras Sereda, Tom St. John, Burak Bartan, Natalie Serrino, Sachin Katti, Zain Asgar arxiv

GPU kernels are critical for ML performance but difficult to optimize across diverse accelerators. We present KForge, a platform-agnostic framework built on two collaborative LLM-based agents: a generation agent that produces and iteratively refines programs through compilation and correctness feedback, and a performance analysis agent that interprets profiling data to guide optimization. This agent-based architecture requires only a single-shot example to target new platforms. We make three key contributions: (1) introducing an iterative refinement system where the generation agent and performance analysis agent collaborate through functional and optimization passes, interpreting diverse profiling data (from programmatic APIs to GUI-based tools) to generate actionable recommendations that guide program synthesis for arbitrary accelerators; (2) demonstrating that the generation agent effectively leverages cross-platform knowledge transfer, where a reference implementation from one architecture substantially improves generation quality for different hardware targets; and (3) validating the platform-agnostic nature of our approach by demonstrating effective program synthesis across fundamentally different parallel computing platforms: NVIDIA CUDA and Apple Metal.

📄 PDF Abstract BibTeX arXiv:2511.13274

Code (0)

등록된 구현이 없습니다.

Tasks

Program Synthesis

Similar Papers 제목 키워드 기반

KForge: LLM-Driven Cross-Platform Kernel Generation for AI Accelerators

2026-06-01 · Taras Sereda, Burak Bartan, Ankita Nayak, Tom St. John 외 arxiv

Production inference increasingly targets a heterogeneous mix of accelerators. Agentic pipelines interleave reasoning, tool calls, and multi-agent coordination, each with distinct compute and memory profiles. For optimal…

Code Generation

LLM-Driven Design Space Exploration of FPGA-based Accelerators

2026-05-07 · Vinamra Sharma, Xingjian Fu, Jude Haris, José Cano arxiv

Designing field-programmable gate array (FPGA)-based accelerators for modern artificial intelligence workloads requires navigating a large and complex hardware design space encompassing architectural parameters, dataflow…

COGNATE: Acceleration of Sparse Tensor Programs on Emerging Hardware using Transfer Learning

2025-05-31 · Chamika Sudusinghe, Gerasimos Gerogiannis, Damitha Lenadora, Charles Block 외

Sparse tensor programs are essential in deep learning and graph analytics, driving the need for optimized processing. To meet this demand, specialized hardware accelerators are being developed. Optimizing these programs …

Transfer Learning

KernelEvolve: Scaling Agentic Kernel Coding for Heterogeneous AI Accelerators at Meta

2025-12-29 · Gang Liao, Hongsen Qin, Ying Wang, Alicia Golden 외 arxiv

Making deep learning recommendation model (DLRM) training and inference fast and efficient is important. However, this presents three key system challenges - model architecture diversity, kernel primitive diversity, and …

A Graph Deep Learning Framework for High-Level Synthesis Design Space Exploration

2021-11-29 · Lorenzo Ferretti, Andrea Cini, Georgios Zacharopoulos, Cesare Alippi 외

The design of efficient hardware accelerators for high-throughput data-processing applications, e.g., deep neural networks, is a challenging task in computer architecture design. In this regard, High-Level Synthesis (HLS…

Graph Neural NetworkHigh-Level Synthesis