Neuron Custom C++ Library Release Notes — AWS Neuron Documentation (original) (raw)

Contents

• aws-neuronx-gpsimd-customop-lib [0.13]
• aws-neuronx-gpsimd-customop-lib [0.3]
• aws-neuronx-gpsimd-customop [0.1]

This document is relevant for: Inf1, Inf2, Trn1, Trn2

Neuron Custom C++ Library Release Notes#

Note

Neuron Custom C++ Operators feature is currently supported on NeuronCore-v2 architecture only, which is found in Trainium (Trn1) and second-generation Inferentia (Inf2) chips.

aws-neuronx-gpsimd-customop-lib [0.13]#

Date: 12/12/2024

• Neuron Custom C++ Operators feature is currently supported on NeuronCore-v2 architecture only, which is found in Trainium (Trn1) and second-generation Inferentia (Inf2) chips.

aws-neuronx-gpsimd-customop-lib [0.3]#

Date: 04/28/2023

• Add initial support for using Multiple GPSIMD Cores for Custom C++ Operators
• Package name was changed to aws-neuronx-gpsimd-customop-lib

aws-neuronx-gpsimd-customop [0.1]#

Date: 02/08/2023

• First release of aws-neuronx-gpsimd-customop. This release provides tensor library support required for building Neuron Custom C++ operators.

This document is relevant for: Inf1, Inf2, Trn1, Trn2