Backend
DexAttention: Distributed Sparse Attention through Rapid Online Planning and Persistent GPU Execution
arXiv cs.DC0 views
arXiv:2611.23512v1 Announce Type: new Abstract: Sparse attention reduces computation, but distributed execution may still transfer unnecessary data and incur substantial per-operation overhead. Mask-aware assignment and scheduling can reduce these costs, but replanning for dynamic masks must be cheaper than the execution time it saves. We present DexAttention, a distributed sparse attention system that co-designs fast online planning with persistent execution. Its assigner uses GPU-accelerated, topology-aware clustering to balance computational load and reduce communication. A lightweight greedy scheduler orders coarse communication and computation operations, while a persistent kernel on each GPU enforces fine-grained dependencies and dispatches ready tasks. This division keeps detailed synchronization out of planning while enabling overlap without repeated attention launches. The framework supports static and dynamic per-head masks across multiple attention compute kernels. On a four-node, 16-GPU cluster, DexAttention achieves a $1.39\times$ geometric-mean end-to-end training speedup over the fastest applicable baseline across 11 configurations. For dynamic video attention, it achieves $2.11$--$2.37\times$ speedups over Ring Attention including planning.
Source:https://arxiv.org/abs/2611.23512PreviousEnd-to-End Multi-Task Learning for Adjustable Joint Noise Reduction and Hearing Loss CompensationNextStephen: Verifying Sequencer-Aided Coordination Systems with Cross-View Prophecy
Related posts
0.58.0 and v0.58.0 Are the Same Version. Our Code Disagreed.
0dev.to
How to pull a Substack newsletter's posts as Markdown
0dev.to
Yup for Beginners: Schema Validation That Reads Like English
0dev.to
EU Digital Fairness Act Proposal Targets Dark Patterns in Online Shopping and Marketing
0dev.to
I self-hosted a CRM and Claude Code set up a sales pipeline in 58 seconds. Here's what MCP got right and wrong
0dev.to
Integrating Sentry into Tauri 2 without losing the local crash log
0dev.to