Metal Pipeline State Caching and Kernel Dispatch Bubbles
Parent: Local Inference Micro-Architectures: Signaling, Quantization Superblocks, and Kernel Pipeline Scheduling · Published reference · snapshot 2026-09-30
↓ Facts as markdownall context files
Source-anchored research for Metal Pipeline State Caching and Kernel Dispatch Bubbles. Run status: COMPLETED.
These notes link each claim to its source. Some citations identify archived research reports that are no longer hosted on this site. A published reference means the content is available; it does not certify independent review or accuracy.Read the editorial policy and follow the sources before relying on a claim.
Measurements and reference values
- Synchronously blocking CPU execution on MTLCommandBuffer waitUntilCompleted() prevents overlapping CPU dispatch with GPU execution, causing up to 25% throughput degradation on small models. (high confidence) [source]
Facts and statements
- Runtime compilation of Metal compute shaders creates hitches during initial token generation; caching compiled pipeline states in an MTLBinaryArchive eliminates shader compilation latency. (high confidence) [source]
Children
- No children recorded.