This document is relevant for: Inf1
, Inf2
, Trn1
, Trn2
AWS Neuron SDK 2.25.0: Neuron Compiler release notes#
Date of release: July 31, 2025
Go back to the AWS Neuron 2.25.0 release notes home
Announcements#
The Neuron Compiler default for the --auto-cast
option will change from --auto-cast=matmult
to --auto-cast=none
in a future release.
Behavioral changes#
Behavioral changes are small, user-facing changes that you may notice after upgrading to this version.
Minor bug fixes and performance enhancements for both the
trn1
andtrn2
platforms.
Known issues#
The Llama3 70B test has a compile time increase of 16% and 18%, for 16 and 32 nodes respectively. We are investigating the cause of this increase and will provide an update in the future.
This document is relevant for: Inf1
, Inf2
, Trn1
, Trn2