IndustrySelected76°

TACC 将 SambaStack 部署规模扩大三倍

Why we picked it

SambaNova 老大 Rodrigo Liang 接受 HPCwire 采访,聊 TACC 为什么把 SambaStack 扩到三倍:训练留给超算,推理跑 SambaNova。做 HPC 或科研计算的人可以看看这种分工思路。

德州先进计算中心(TACC)宣布将其 SambaNova 的 SambaStack 部署规模扩大三倍。SambaNova CEO Rodrigo Liang 在 HPCwire 的采访中解释了这一决策:科学计算仍依赖超算,但越来越多的环节开始由 AI 承担。

Body

新的分工模式是:模型训练放在传统超算上完成,推理负载则交给 SambaNova 的 AI 推理硬件运行,让两类系统各自发挥所长。这种"训练用超算、推理用专用 AI 芯片"的架构正在成为高性能计算中心的常见配置。

TACC 是美国主要学术超算中心之一,此次扩容三倍表明科研机构对推理算力的需求在快速增长,也为其他 HPC 中心提供了可参考的部署路径。

Original · SambaNova

"Science still runs on supercomputers, and more of it now runs on AI." — @RodrigoLiang in @HPCwire on why @TACC is tripling its SambaStack deployment: train on the supercomputer, run inference on SambaNova, and let each do what it's best at.

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 SambaNova, TACC, HPC, AI inference.

Source attribution: curated from SambaNova (@SambaNovaAI).