Reusable components for CUDA API development.
This package provides some reusable functionality for programming against CUDA or NVIDIA APIs. It is meant to be used by package developers, and is not expected to export functionality useful for application programmers.
The file src/compatibility.jl
contains hard-coded databases with software and hardware
compatibility information that cannot be queried from APIs.
The file src/discovery.jl
defines helper methods for discovering the NVIDIA driver and
CUDA toolkit, as well as some more generic methods to find libraries and binaries relative
to eg. the location of the driver or toolkit.
When a new version of CUDA is released, CUDAapi.jl needs to be updated accordingly:
discovery.jl
: update thecuda_versions
dictionarycompatibility.jl
: update each_db
variable (refer to the comments for more info)travis.linux
andtravis.osx
: provide a link to the installersappveyor.ps1
: provide a link to the installer, and list the components that need to be installedtravis.yml
andappveyor.yml
: add the version to the CI rosters
Update the gcc_major_versions
and gcc_minor_versions
ranges in
discovery.jl
to cover the new version.