Expand description
platform category (reorg).
Re-exports§
pub use precision_timer::format_duration_human;pub use precision_timer::BenchmarkResult;pub use precision_timer::DurationPrecisionExt;pub use precision_timer::PrecisionTimer;
Modules§
- accelerator_
inventory - compute_
bridge - Compute bridge — the engine’s one shared “which hardware runs this kernel” surface (HARDWARE_BACKEND_AUTOSELECT_PLAN.md; CLAUDE.md §13).
- device_
benchmark - STELLAR §A AH-track H1(a) — cross-circuit capability benchmark (decisions D26/D30).
- git_
bridge - Minimal Merkle-DAG for Q42 volume history (Phase 1 — §4.5 Option B).
- gpu
- GPU integration for calculus modality.
- hardware_
passport - STELLAR §A AH-track H1(a) cache — the hardware passport (decision D26).
- hetero_
dispatch - Vendor-neutral heterogeneous compute dispatch + storage / precision policy.
- host
- Host-side I/O management for calculus modality.
- kml_
bridge - KML import/export bridge.
- local_
scheduler - Local Scheduler — Thread-per-Core Production Queue
- npu_ffi
- GPU sieve and native host domain bridges
Uses
wgputo execute cross-platform GPU compute shaders (Vulkan/DX12/Metal/WebGPU) that filter the 5th Vector of the NQuin parallel arrays. - platform_
scheduler - Platform-aware thread QoS and core-affinity binding.
- precision_
timer - Precision Timer and Sub-Millisecond Benchmark Telemetry.
- tee_ffi