# GGML Metal Kernel Dispatch and Command Buffer Offload > Source-anchored research for GGML Metal Kernel Dispatch and Command Buffer Offload. Run status: COMPLETED. Parent: [Local Inference Acceleration: Memory Architectures, Metal Kernels, and Latency Optimization](https://llms-explorer.com/tree/local-inference-acceleration-and-kernels/) · 2 facets · 2 facts · page: https://llms-explorer.com/tree/ggml-metal-kernel-dispatch-and-command-buffer-offload/ ## How it works - ggml-metal batches matrix multiplication and activation kernels into Metal command buffers, but host-device barrier synchronization causes pipeline stalls if intermediate outputs are queried synchronously. (high confidence) — [source](https://github.com/ggerganov/llama.cpp/blob/master/ggml/src/ggml-metal.metal) ## Facts and statements - Pre-compiling Metal pipeline state objects and caching shader functions avoids runtime compilation pauses during initial context evaluation. (high confidence) — [source](https://github.com/ggerganov/llama.cpp/blob/master/ggml/src/ggml-metal.metal)