Create

Sign in to ReadmeX

or

How we handle your data: Privacy Policy

ReadmeX

Your AI reading companion

What would you like to catch up on?

Pick a briefing, or ask your own question.

Your community and people briefings stay personal to you.

Story

NVIDIA expands DSX use for power-flexible AI factories

AI summary

NVIDIA said companies are using the DSX AI Factory Platform to plan, simulate and operate grids and AI factories as unified systems, while European partners are adopting dynamic power allocation and grid coordination. TechNews reported that tests by Nscale and Lambda produced higher inference or token throughput under fixed power budgets, while NVIDIA claimed DSX MaxLPS can cut power use by up to 30% at the same AI compute output. Preliminary tests with Emerald AI, EPRI, Nebius and National Grid reportedly reduced a Blackwell Ultra cluster’s power demand by up to 40% during simulated grid events while retaining critical workloads.

In-depth insight

DSX connects grid, facility and compute operations from the substation to the rack, allowing AI factories to adjust power demand and allocate resources according to grid conditions.

AI 資料中心缺電怎麼辦?輝達從用電端下手,DSX 最高省電 30%

隨著 AI 資料中心用電需求持續攀升,除了增加電力供應,如何從用電端提高效率也成為重要課題。輝達於德國柏林 GTC 大會宣布,歐洲多家合作夥伴正採用 DSX AI Factory Platform,透過動態電力分配及電網調度技術,提高 AI 工廠能源使用效率,其中 DSX MaxLPS 軟體可在維持相同 AI 運算產出的情況下,最多降低 30% 用電量。

事實上,AI 工廠內的 GPU 並非隨時都需要相同電力,不同運算工作與執行階段,對電力的需求也有所差異。DSX MaxLPS 便是利用這項特性,在既定電力預算內動態分配 GPU 電力,將電力資源配置至更需要的運算工作,使業者不必增加供電容量,也有機會提高整體運算產出。

#NVIDIA#NVIDIA DSX AI Factory Platform#DSX MaxLPS#DSX Flex#DSX Exchange#Nscale#Lambda#Emerald AI#EPRI#Nebius#National Grid#Phaidra

Original · 3 min readRead the original
Group chat

No comments yet. Start the conversation.