Device software glossary

Device software glossary#

2026-02-20

4 min read time

Applies to Linux and Windows

This section provides brief definitions of software abstractions and programming models that run on AMD GPUs.

AMDGPU assembly#

AMDGPU assembly (GFX ISA) is the low-level assembly format for programs running on AMD GPUs, generated by the ROCm compiler toolchain. See hip:amdgpu_assembly for instruction set details.

AMDGPU intermediate representation#

AMDGPU IR is an intermediate representation for GPU code, serving as a virtual instruction set between high-level languages and architecture-specific assembly. See hip:amdgpu_ir for compilation details.

Global memory#

Global memory is the device-wide memory accessible to all threads, physically implemented as HBM or GDDR. See hip:memory_hierarchy for global memory details.

Grid#

A grid represents the collection of all work-groups executing a single kernel across the entire GPU. See hip:inherent_thread_hierarchy_grid for grid execution details.

HIP kernel#

A HIP kernel is the unit of GPU code that executes in parallel across many threads, distributed across the GPU’s compute units. See hip:device_program for kernel programming details.

HIP memory hierarchy#

The memory hierarchy pairs each thread hierarchy level with corresponding memory scopes, from private registers to LDS to GPU RAM. See hip:memory_hierarchy for memory architecture details.

HIP thread hierarchy#

The thread hierarchy structures parallel work from individual threads to blocks to grids, mapping onto hardware from SIMD lanes to compute units to the entire GPU. See hip:inherent_thread_model for complete details.

LLVM target name#

The LLVM target name is a string identifier corresponding to a specific GFX IP version that is passed to the HIP compiler toolchain to specify the target GPU architecture for code generation. See llvm-project:reference/rocmcc for details.

ROCm programming model#

The ROCm programming model defines how AMD GPUs execute massively parallel programs using hierarchical work-groups, memory scopes, and barrier synchronization. See hip:programming_model for complete details.