rscrypto 0.3.1

# Platforms

`rscrypto` is built around a three-tier dispatch model so the same source compiles to a portable Rust path on every target and a hardware-accelerated path on every supported CPU.

## Dispatch Model

1. **Compile-time** — `#[cfg(target_feature = "...")]` selects the strongest backend permitted by `RUSTFLAGS` / `target-feature`.
2. **Runtime detection** (`std` only) — cached `platform::caps()` probes CPU features once at startup via `is_x86_feature_detected!` and the aarch64 / s390x / ppc64le / riscv64 equivalents, then dispatches to the strongest available kernel.
3. **Portable Rust fallback** — always present. The portable implementation is the source of truth; SIMD and ASM kernels are differentially tested against it.

In `no_std` builds, only the compile-time tier runs.

The `portable-only` feature hard-disables tier 2 even with `std` enabled. See [`features.md`](features.md#portable-only).

## Acceleration Matrix

Availability of any specific backend depends on what the target CPU advertises and what `target-feature` permits. The portable Rust fallback is present on every target listed below.

| Target family | Backends used (when CPU advertises them) |
|---|---|
| x86_64 | SSE4.2 CRC32; SSSE3 / PCLMULQDQ; AVX2; AES-NI; SHA-NI; AVX-512F / VL / BW / DQ; AVX-512IFMA; VPCLMULQDQ; VAES |
| aarch64 / Apple Silicon | NEON; AES; PMULL; CRC; SHA2; SHA3 / EOR3; SHA512; SVE2-PMULL where available |
| s390x (IBM Z) | z/Vector; vector enhancements; CPACF / MSA; VGFM |
| ppc64le (POWER) | AltiVec; VSX; POWER8 vector / crypto; POWER9 / POWER10 vector; VPMSUMD |
| riscv64 | V / RVV; Zbc; Zvbc; Zbkc; Zkne / Zknd; Zvkned; Zkt / Zvkt |
| wasm32 | SIMD128 where enabled |

## `no_std` Targets

The following `no_std` targets are built in CI:

- `thumbv6m-none-eabi`
- `riscv32imac-unknown-none-elf`
- `aarch64-unknown-none`
- `x86_64-unknown-none`
- `wasm32-unknown-unknown`
- `wasm32-wasip1`

Other `no_std` targets in the same families (e.g. `thumbv7em-*`, larger RISC-V profiles) generally work; open an issue if a target you care about is missing.

## Per-Platform Benchmark Scorecard

Current geomean speedups by platform live in [`benchmark_results/OVERVIEW.md`](../benchmark_results/OVERVIEW.md#coverage-matrix). The current public set includes the 2026-05-27 nine-runner Linux CI matrix and a 2026-06-01 local Apple Silicon MBP M1 full run.