atomr-accel-cuda 0.10.0

GPU acceleration via the actor model. Wraps NVIDIA CUDA libraries (cuBLAS, cuDNN, cuFFT, cuRAND, cuSOLVER, cuSPARSE, cuTENSOR, cuBLASLt, NVRTC, NCCL) as supervised atomr actors with generation-validated buffers and a uniform async surface.
Documentation 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23

//! Managed (unified) memory + Phase 3 driver-API helpers.
//!
//! - [`managed`] — `cudaMallocManaged` actor (`ManagedAllocatorActor`,
//!   `ManagedRef<T>`, `ManagedFlags`, `ManagedStats`,
//!   `PrefetchTarget`).
//! - [`prefetch`] — `cuMemPrefetchAsync` wrapper.
//! - [`advise`] — `cuMemAdvise` wrapper + the [`advise::MemAdvice`]
//!   typed enum.
//! - [`ipc`] — `cuIpcGetMemHandle` / `cuIpcOpenMemHandle` /
//!   `cuIpcCloseMemHandle` (gated `cuda-ipc`).

pub mod advise;
#[cfg(feature = "cuda-ipc")]
pub mod ipc;
pub mod managed;
pub mod prefetch;

pub use advise::MemAdvice;
#[cfg(feature = "cuda-ipc")]
pub use ipc::{IpcMemHandle, OpenedMem};
pub use managed::{
    ManagedAllocatorActor, ManagedFlags, ManagedMsg, ManagedRef, ManagedStats, PrefetchTarget,
};