AI芯片与算力分论坛预告 || 第八届北京高校人工智能学术论坛

“通推一体”香山RISC-V AI处理器

微架构及开源IP研发    

生成式AI推理日新月异、蓬勃发展（例如大语言模型、多模态、智能体等），日常生活中的应用软件同步呈现“智能化”趋势。智能化软件是在传统软件基础上增加AI推理功能，因此体现出对传统通用计算和新兴AI推理兼需的趋势。本研究分享了如何为香山高性能RISC-V处理器增加矩阵计算能力，从而将纯粹的通用计算CPU升级为兼顾“通用计算”和“AI推理”的”通推一体”香山AI CPU。与此同时，将详细分享“通推一体”香山RISC-V AI 处理器的微架构、AI推理性能、软件编程模式及开源IP研发进展。    

基于数据流技术的边缘大模型推理加速研究

算电协同背景下构筑分布式算网的一些思考

GCC: A 3DGS Inference Architecture with Gaussian-Wise and Cross-Stage Conditional Processing

3D Gaussian Splatting (3DGS) has emerged as a leading neural rendering technique for high-fidelity view synthesis, prompting the development of dedicated 3DGS accelerators for resource-constrainedplatforms. The conventional decoupled preprocessing-renderingdataflow in existing accelerators has two major limitations: 1) a significant portion of preprocessed Gaussians are not used in rendering, and 2) the same Gaussian gets repeatedly loaded acrossdifferent tile renderings, resulting in substantial computational anddata movement overhead. To address these issues, we propose GCC,a novel accelerator designed for fast and energy-efficient 3DGSinference. GCC introduces a novel dataflow featuring: 1) cross-stageconditional processing, which interleaves preprocessing and rendering to dynamically skip unnecessary Gaussian preprocessing;and 2) Gaussian-wise rendering, ensuring that all rendering opera tions for a given Gaussian are completed before moving to the next,thereby eliminating duplicated Gaussian loading. We also proposean alpha-based boundary identification method to derive compactand accurate Gaussian regions, thereby reducing rendering costs. We implement our GCC accelerator in 28nm technology. Extensive experiments demonstrate that GCC significantly outperformsthe state-of-the-art 3DGS inference accelerator, GSCore, in both performance and energy efficiency.

大模型驱动的处理器电路

自动设计、验证与优化

随着芯片复杂度持续提升，处理器设计需求与设计资源之间的矛盾日益突出。代码大模型在软件工程中已取得显著进展，但在硬件设计领域距离实际应用仍有较大差距，基础模型能力和应用方法仍需进一步发展。

大模型用于芯片设计主要面临四方面挑战：高质量数据不足、设计抽象层级跨度大、模态多样以及优化目标复杂。围绕这些问题，本报告将介绍实验室围绕处理器电路逻辑设计、形式验证和PPA优化开展的一系列工作，包括基础模型、测试基准和Agent框架。这些工作对大模型驱动的处理器设计新流程进行了初步探索。

Hardwired-Neuron Language Processing Units as General-Purpose Cognitive Substrates

The rapid advancement of Large Language Models (LLMs) has established language as a core general-purpose cognitive substrate, driving the demand for specialized Language Processing Units (LPUs) tailored for LLM inference. To overcome the growing energy consumption of LLM inference systems, this paper proposes a Hardwired-Neurons Language Processing Unit (HNLPU), which physically hardwires LLM weight parameters into the computational fabric, achieving several orders of magnitude computational efficiency improvement by extreme specialization. However, a significant challenge still lies in the scale of modern LLMs. A straightforward hardwiring of GPT-OSS-120B would require fabricating photomask sets valued at over 6 billion dollars, rendering this straightforward solution economically impractical.

Addressing this challenge, we propose the novel Metal-Embedding methodology. Instead of embedding weights in a 2D grid of silicon device cells, Metal-Embedding embeds weight parameters into the 3D topology of metal wires. This brings two benefits: (1) a 15× increase in density, and (2) 60 out of 70 photomask layers are homogeneous across chips, including all EUV photomasks. In total, Metal-Embedding reduced the photomask cost by 112×, bringing the Non-Recurring Engineering (NRE) cost of HNLPU into an economically viable range. Experimental results show that HNLPU achieved 249,960 tokens/s (5,555×/85× that of GPU/WSE), 36 tokens/J (1,047×/283× that of GPU/WSE), 13,232 mm2 total die area, 59.46 M-123.5 M estimated NRE at 5 nm technology. Analysis shows that HNLPU achieved 41.7-80.4× improvement in cost-effectiveness and 357× reduction in carbon footprint compared to OpenAI-scale H100 clusters, under an annual weight updating assumption.