Module numa_gpu::runtime::cuda_wrapper
source · [−]Expand description
Collection of Rust-ified wrappers for commonly-used CUDA functions.
Structs
CUDA memory information
Enums
Constants
Functions
Copy a slice using CUDA’s memcpyAsync function.
Page-lock an existing memory range for efficient GPU transfers.
Unregisters a memory range that was page-locked with host_register().
Advise how the memory range will be used.
Returns the free and total device memory in bytes
Fill a launchable slice using the CUDA memset_async
function
Prefetch memory to the device specified in the current context.