Crate onnx_runtime_ep_cpu 

onnx-runtime-ep-cpu

The CPU execution provider for the ORT 2.0 runtime (see docs/ORT2.md §4.4 and §54 Phase 1). It implements onnx_runtime_ep_api::ExecutionProvider and hosts pure-Rust reference kernels for the Phase-1 op set (MatMul, Add, Relu, Reshape, Transpose, Gather, LayerNormalization).

Backends: correctness baseline + optional oneDNN

The GEMM hot spot is served through backend::CpuBackend (docs/ORT2.md §25.2). The default backend is a pure-Rust blocked, register-tiled, rayon-parallelized f32 GEMM — the portable, offline correctness baseline that compiles anywhere with no C++/FFI. The non-default onednn cargo feature statically links oneDNN and routes the 2-D tile GEMM through dnnl_sgemm ([kernels::onednn]). Every backend lives behind the onnx_runtime_ep_api::Kernel trait, so neither the EP contract nor the session observes which one ran. See kernels::matmul for the hot spot.

unsafe

The default (Generic) path is unsafe-minimal: the only unsafe is the raw device-buffer access the ep-api contract forces (aligned host alloc/dealloc, memcpy, and strided element reads/writes), each isolated and SAFETY-documented, plus — only under the onednn feature — the dnnl_sgemm FFI call, confined to [kernels::onednn]. The blocked rayon GEMM itself contains no unsafe; all kernel arithmetic is safe Rust operating on dense Vec<f32> buffers produced by the two audited accessors in kernels.

Re-exports

pub use backend::CpuBackend;

pub use backend::has_onednn;

pub use provider::CpuExecutionProvider;

pub use kernels::slice::SliceAxisPlan;

pub use kernels::slice::slice_axes_steps;

pub use kernels::slice::slice_plan;

Modules

backend

CPU GEMM backend selection (docs/ORT2.md §25.2 “CPU Backend Strategy”).

dtype

Reusable dtype-generic machinery for the arithmetic CPU kernels (docs/ORT2.md §4.4; project preference “不同的dtype，是不是可以用 template”).

kernels

CPU kernels for the Phase-1 BERT-on-CPU correctness milestone (docs/ORT2.md §4.4). One [Kernel] per ONNX op, keyed purely by op type — there are no model-specific shapes or names anywhere in this crate; BERT is only the validation target.

provider

The CpuExecutionProvider: a host execution provider backed by pure-Rust reference kernels (docs/ORT2.md §4.4).

strided

Strided-view access helpers shared by the CPU kernels (docs/ORT2.md §5).

Macros

dispatch_arith

Map a runtime DataType to a monomorphized body over the matching Rust element type, across the full ONNX numeric set (floats + signed/unsigned integers). Binds $T via a local type alias; unsupported dtypes yield a RULE #1 error. The body must evaluate to Result<()>.

dispatch_float

Like [dispatch_arith] but restricted to the floating-point dtypes ONNX defines transcendental / accumulate ops over (f32, f16, bf16, f64).