Linux 7.3 Ships Faster Scheduling, I/O

// 1h agoOPENSOURCE RELEASE

Linux 7.3 Ships Faster Scheduling, I/O

Linux 7.3 brings completed sched_ext support, improved cgroup scheduling, faster filesystem I/O, new sandboxing primitives, and broader hardware support. The October 18 release is especially relevant for developers running demanding workloads, virtual machines, and low-end GPUs.

// ANALYSIS

Linux 7.3 is a substantial infrastructure release whose biggest gains are practical rather than flashy, improving responsiveness, storage throughput, and kernel extensibility.

–sched_ext lets cgroups run customizable BPF-based schedulers, opening new workload-tuning possibilities
–Cgroup scheduling improvements can reduce latency when background workloads compete with interactive applications
–Ext4, Btrfs, FUSE, and io_uring receive performance improvements that benefit databases, build systems, and servers
–failfs and fchroot provide new primitives for sandboxing and testing filesystem behavior
–The release set records with 18,765 non-merge commits from 3,021 developers, underscoring Linux’s enormous development velocity

// TAGS

linux-kernelopen-sourceself-hostedsecuritytesting

DISCOVERED

1h ago

2026-10-19

PUBLISHED

2h ago

2026-10-19

RELEVANCE

5/ 10

AUTHOR

cjd8

// KEEP READING

More AI developer news from the feed

EXPLORE FULL FEED

NEWS22m ago

GPT-4o API Retirement Sparks Developer Pushback

OpenAI’s October 14 deprecation notice schedules GPT-4o’s remaining API snapshots for shutdown on January 14, 2027. The campaign calls for preserving or open-sourcing GPT-4o, but OpenAI has announced no open-weights release.

UPDATE1h ago

Rivian Opens Point-to-Point Beta

Rivian is inviting select U.S. owners of Gen 2 R1 and R2 vehicles with paid Autonomy+ subscriptions to test its route-following driving feature. The invite-only rollout begins in late 2026 under NDA, with broader availability targeted for 2027.

INFRA1h ago

Qwen3.6 GGUF Repack Lands on Modelers

Modelers.cn added an amd/Qwen3.6-35B-A3B-gguf artifact, repackaging Alibaba’s existing 35B-total, 3B-active MoE model for local inference. This expands deployment options without introducing new weights or capabilities; the underlying model was already supported in AMD’s official deployment guide.