Industry79°

OpenAI 模型与 AMD 合作优化 GPU 设计,PPA 提升最高 14%

Why we picked it

OpenAI 的模型帮 AMD 优化旗舰 GPU,PPA 提升最高 14%,部分改动已经进流片了。相当于 5 个月干出一年的活,这种芯片设计玩法 NVIDIA 那边还没见过,做硬件和买卡的都该看看。

OpenAI 的模型被用于芯片设计环节,与 AMD 合作在其旗舰 GPU 设计上实现最高 14% 的功耗、性能与面积(PPA)提升,且这是在已高度优化的基线之上取得的额外增益。

Body

OpenAI 称这一优化速度相当于把约一年的工程进展压缩到不到 5 个月内完成,部分改动已被纳入 AMD 交付制造的设计版本中。这意味着 AI 优化不再停留在软件层面,而是直接进入了芯片设计到流片的物理流程。

对芯片行业而言,在优化饱和的基线上再挤出两位数收益十分罕见,AI 辅助 EDA 可能成为 AMD 对抗 NVIDIA 的一个差异化手段;对开发者来说,未来的 AMD GPU 可能因此带来更好的性价比。

Original · kimmonismus

Recursive self-improvement, chip edition:

OpenAI's models are now helping design the chips they run on. With AMD, an OpenAI model found up to 14% gains in power, performance and area on AMD's flagship GPU design, on top of heavily optimized baselines.

That's roughly a year of progress in under 5 months, OpenAI says. Some of the changes are already in a design AMD sent for manufacturing.

That's pretty nutty and really cool to see.

Industry Coordinates & Source Verification

This report is categorized under AITOP's "Industry" desk, focusing on US frontier AI labs and Silicon Valley technology leaders. Within our cross-regional analytical framework, it directly maps to AI developer toolchains, coding agents, and engineering workflows. Key entities and technical topics under continuous tracking include OpenAI, AMD, chip design, GPU.

Source attribution: curated from kimmonismus (@kimmonismus).