This issue is a blocker because supported PyTorch and vLLM configurations can crash when GPU memory is overcommitted, preventing customers from enabling new models.
-- You received this bug notification because you are a member of Ubuntu Bugs, which is subscribed to Ubuntu. https://bugs.launchpad.net/bugs/2165799 Title: [LNL][WCL][PTL][NVL][BMG] Fix GPU memory overcommit that leads to crash To manage notifications about this bug go to: https://bugs.launchpad.net/intel/+bug/2165799/+subscriptions -- ubuntu-bugs mailing list [email protected] https://lists.ubuntu.com/mailman/listinfo/ubuntu-bugs
