cuda-api-wrappers

Changes since v0.8.1:

CUDA 13.x compatibility & features

• #755 : We can now obtaing graph containing a given node, even when we only have its handle (CUDA 13.1)
• #754 : Support some new CUDA 13.1 graph-related API calls involving IDs
• #748, #749, #746 : Now supporting the use of CUDA-13-versions of some API functio...

Changes since v0.8.0:

CUDA 12.x compatibility

• #711 : Added preliminary information regarding Blackwell cards and micro-architecture *#701 : The --version-ident compilation option to NVRTC was dropped in CUDA 12.2; this is now respected by the wrappers and the option is not exposed for 12.2 and newer versions of CUD...

Changes since v0.7.1:

Support for the nvFatbin library (#681)

• The API wrappers now support NVIDIA's "fat binary" file format creation/marshalling library, nvFatbin. It is supported via a cuda::fatbin_builder_t class: One creates a builder, adds various fragments of...

Changes since v0.7.0:

CUDA Graphs

• #532 Now supporting empty nodes as graph::typed_node's.
• #656 When capturing a graph on a CUDA stream - default to capturing on using the global capture mode task
• #657 Offer stand-alone functions for stream capture begin and end tasks (as opposed to only stream_t class methods)...

Changes since v0.6.9:

Graph support

The wrappers library now supports the creation and manipulation of CUDA graphs - meriting a version number bump.

One can now:

• Construct graphs (= graph templates) directly
• Capture graphs (= graph templates) on streams
• Insta...