1 مستودع
Mechanisms for selecting the most efficient low-level implementation of a mathematical operation based on performance and accuracy targets.
Distinct from Algorithmic Performance Optimizations: Focuses on switching between specific kernel implementations (e.g., Winograd vs GEMM) rather than general complexity analysis.
Explore 1 awesome GitHub repository matching software engineering & architecture · Kernel Implementation Selection. Refine with filters or upvote what's useful.
oneDNN is a library for deep learning acceleration that provides optimized building blocks for neural network training and inference. It manages tensor computation across CPU and GPU hardware, enabling the execution of high-performance primitives for model training and neural network inference optimization. The project distinguishes itself through hardware-specific kernel optimization and the use of just-in-time compilation to target specific processor instruction sets. It supports quantized neural network execution using both static and dynamic quantization to reduce memory usage and increas
Selects between direct, Winograd, or implicit GEMM implementations to balance performance, memory, and numerical accuracy.