AMD ROCm HRX's Loom 0.2 Expands C++ Support, Improved Targeting

AMD ROCm's HRX Loom as the compiler, runtime, and serving infrastructure for supporting AI agent workflows is out today with a new version. Following last month's HRX Loom 0.1 inaugural release, Loom 0.2 is now available. This new version expands the C++ program support, packaged device execution capabilities, native AMD GPU launch and replay support, large graph management, target and JIT efficiency, and various other improvements. In the official release announcement, HRX Loom 0.2 is summed up as: "Loom 0.2.0 expands complete-program C++ authoring, runtime-provided managed objects, stateful finite services, and packaged AMD GPU/NPU and Vulkan execution. It expands the shared target compiler across x86, AIE2P, AMDGPU, SPIR-V, and WebAssembly while removing fixed and superlinear costs from JIT compilation, graph construction and updates, GPU launch and replay, pipeline planning, and long-context serving." Great seeing shared target compiler improvements across different hardware and infrastructure. The release notes go on to elaborate: "Loom tools, examples, serving, sanitizers, and hardware qualification now use one libamdf-backed AMD HAL through the existing --device=amdgpu spelling. The native path supports RDNA and CDNA execution, shared CPU/GPU/NPU memory contracts, explicit compute-resource selections, cooperative execution, virtual mappings, and reusable command buffers. The former ROCr-backed driver, its legacy-amdgpu selector, and its separate runtime and memory ownership paths are removed." More details on this and other features of AMD ROCm's HRX Loom 0.2 release can be found via GitHub.

Original Source

Read the full article at Phoronix →

KhanList aggregates and links to publicly available news content. We do not host full articles from third-party sources. Always verify important information with original sources.