# Metal Pipeline State Caching and Kernel Dispatch Bubbles > Source-anchored research for Metal Pipeline State Caching and Kernel Dispatch Bubbles. Run status: COMPLETED. Parent: [Local Inference Micro-Architectures: Signaling, Quantization Superblocks, and Kernel Pipeline Scheduling](https://llms-explorer.com/tree/inference-microarchitectures-and-kernel-pipelines/) · 2 facets · 2 facts · page: https://llms-explorer.com/tree/metal-pipeline-state-caching-and-kernel-dispatch-bubbles/ ## 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](https://developer.apple.com/documentation/metal/gpu_compilation/building_a_shader_library_offline) ## 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](https://developer.apple.com/documentation/metal/gpu_compilation/building_a_shader_library_offline)