Releases: LLNL/CHAI
v2.2.1
Hotfix release for building with CUDA and ENABLE_RAJA_PLUGIN against an externally installed RAJA.
v2.2.0
This release includes a bugfixes and feature improvements, and also requires the following tpl versions:
- RAJA v0.12.0
- Umpire v4.0.0
The submodules have been updated to point to these versions.
v2.1.1
Hotfix release for a bug in pinned memory support.
v2.1.0
This release includes a number of major changes and bugfixes:
- Support for global callbacks
- Update Umpire submodule to version 2.1.0
- Support for pinned host memory
v2.0.0
This release includes a major refactoring of the integration with the RAJA performance portability layer. CHAI now provides an ENABLE_RAJA_PLUGIN
option, that when enabled will build a plugin for RAJA that will enable automatic data migration when CHAI and RAJA are linked in the same application.
v1.2.0
This release contains support for AMD devices using HIP, and updates for the 1.0.0 release of Umpire.
v1.1.0
This release of CHAI switches over to using Umpire for the underlying memory allocation and operations, and allows users to set specific umpire::Allocator
s for each ManagedArray and execution space.
v1.0
Merge branch 'release/1.0'