Massive VRAM pools on AMD Instinct accelerators drown Linux's hibernation process — 1.5 TB of memory per server creates headaches

AMD AI Linux-powered servers are failing to hibernate due to excessive VRAM and a high number of AMD Instinct accelerators per system.

https://www.tomshardware.com/tech-industry/supercomputers/massive-vram-pools-on-amd-instinct-accelerators-drown-linuxs-hibernation-process-1-5-tb-of-memory-per-server-creates-headaches

Created 2mo | Jun 30, 2025, 6:30:05 PM


Login to add comment

Other posts in this group

 Internet standards body proposes new header field disclosing AI — will make it easier for machines to determine if AI was used on a site

The IETF is recommending adding a header that will tell devices if the page used AI to generate or edit its contents.

Aug 27, 2025, 7:40:09 PM | tomshardware.com
 Anthropic forms new security council to help secure AI's place in government

Anthropic has built a reputation as the safety-first AI lab, but its latest move makes clear that it’s just as serious about chips as cloud capacity.

Aug 27, 2025, 7:40:06 PM | tomshardware.com
 Phison squashes reports of Windows 11 breaking SSDs — says it was unable to reproduce issues despite 4,500 hours of testing, recommends users deploy heatsinks just in case

Following reports of a Windows 11 update breaking SSDs, Phison has just issued a new statement saying they were unable to replicate the issue internally. Phison wants you

Aug 27, 2025, 5:30:22 PM | tomshardware.com
 Huawei to open-source its UB-Mesh data center-scale interconnect soon, details technical aspects — one interconnect to rule them all is designed to replace everything from PCIe to TCP/IP

Huawei unveiled UB-Mesh at Hot Chips 2025 as an open protocol to unify AI datacenter interconnects, enabling million-processor SuperNodes with lower latency, cost, and hi

Aug 27, 2025, 5:30:11 PM | tomshardware.com