Cross-Vendor f64 Scientific GPU Computing in Rust and WGSL — No CUDA Required

Cross-vendor f64 scientific GPU computing without CUDA. This page documents a pure Rust GPU compute pipeline using WebGPU/WGSL that runs on NVIDIA, AMD, and Intel GPUs — no vendor SDK, no proprietary dependencies, no cloud.

Last Updated: July 31, 2026
License: CC-BY-SA 4.0

✓ VALIDATED ON LIVE HARDWARE — strandGate RTX 3090: 2,130 matmul/sec, 746 pipelines/sec, 98 capabilities LIVE. Dual-vendor proof: RTX 3090 + RX 6950 XT, 100% pass rate. 952 validated WGSL shaders. Current numbers: Evidence Snapshot (measured 2026-08-04-PM).

The Claim

Four public primals — BarraCuda (math), 🐸🍄 ToadStool (orchestration), 🪸🌊 coralReef (compiler), and coral-glowplug (hardware lifecycle) — together replace the NVIDIA CUDA toolchain for scientific computing. Not all of it yet. But a clear, advancing front that already produces paper-parity lattice QCD on a $500 consumer GPU.

This is not a research prototype. These are production-grade primals with 135,000+ combined test functions across 3,598,358 lines of Rust, zero unsafe code, and zero C dependencies in application code.

What’s Already Replaced

1. CUDA Runtime → toadStool + wgpu/Vulkan

Key difference: toadStool discovers hardware at runtime by capability, not by vendor ID. The same code discovers NVIDIA, AMD, Intel, and BrainChip NPU. There is no concept of “CUDA device 0” — there is “the device that supports f64 and has >8GB VRAM.”

2. cuBLAS / cuFFT / cuDNN → BarraCuda WGSL Shaders

952 WGSL shaders total — every one is f64-canonical (native f64 on pro GPUs, DF64 emulation on consumer GPUs). Key domains:

3. nvcc / ptxas → coralReef Sovereign Compiler

What “46/46 shaders compile” means: 46 representative 🐟⚡ barraCuda shaders — covering every domain (bio, physics, ML, linear algebra) — compile from WGSL to native SASS (SM70 Volta, SM86 Ampere) and native RDNA2 (GFX1030) without ANY vendor toolchain. No nvcc, no ptxas, no ROCm. Pure Rust compiler.

4. NVIDIA Kernel Driver → coral-glowplug + VFIO

coral-glowplug is a systemd daemon that manages GPU lifecycle without any NVIDIA software. It binds GPUs to vfio-pci at boot, provides hot-swap personality management (VfioPersonality, NouveauPersonality, AmdgpuPersonality), health monitoring, and graceful shutdown — all via JSON-RPC 2.0 over Unix socket.

Current Performance vs CUDA/Kokkos

The gap is narrowing: 🔥♨️ hotSpring Kokkos parity tracking shows 27×→12.4×→3.7× improvement over three months. The remaining gap is primarily in DF64 transcendental functions (exp, log, sin, cos) where NVIDIA’s NVVM has hand-optimized silicon paths. 🪸🌊 coralReef’s sovereign transcendentals use Newton-Raphson iteration at slightly higher latency.

What’s Coming Next (Near-Term, Given Velocity)

Based on the 27-day sprint velocity (architecture/EVOLUTION_TIMELINE.md), with BarraCuda gaining ~50 new shaders/month and 🪸🌊 coralReef closing 2–3 NVVM bypass patterns per iteration:

3 Months (June 2026)

6 Months (September 2026)

12 Months (March 2027)

The Proprietary Cost of What We Replace

Annual lab savings (conservative): $0 in software licenses (all were technically free) but ~$50K–200K in HPC allocations avoided for a lab running regular plasma, QCD, or bioinformatics GPU workloads. Plus the unmeasurable value of zero-queue 24/7 access and data that never leaves the lab.

The Sovereign Stack Diagram

Application Layer
  └── Springs (wetSpring, hotSpring, neuralSpring, etc.)
      └── Science experiments: reproduce published papers

Math Layer
  └── BarraCuda (

952 WGSL shaders)
      └── f64-canonical math: what to compute

Orchestration Layer
  └── ToadStool S157 (96 JSON-RPC methods, 21,156 tests)
      └── Hardware discovery: where and how to compute

Compiler Layer
  └── coralReef Phase 10 Iter 53 (2,241 tests)
      └── WGSL → SPIR-V → native SASS/RDNA2

Hardware Layer
  └── coral-glowplug (systemd daemon, JSON-RPC)
      └── VFIO device lifecycle: boot, bind, dispatch, recover

Nothing above depends on:
  ✗ NVIDIA CUDA toolkit
  ✗ nvcc / ptxas / NVVM
  ✗ nvidia.ko kernel module
  ✗ Any C/C++ library
  ✗ Any cloud service
  ✗ Any vendor-specific API

Every layer is pure Rust, AGPL-3.0, and publicly auditable. The entire scientific computing stack — from math primitives to GPU binary compilation to hardware lifecycle management — belongs to the user.

Reproduce It

git clone https://github.com/ecoPrimals/barraCuda && cd barraCuda
cargo test --workspace          # all tests pass
cargo run --release --bin validate  # exit 0 = pass

Hardware tested: NVIDIA RTX 4070, RTX 5090, AMD RDNA2, Intel Arc
Precision: f64 via Vulkan SHADER_F64 extension
Date: July 2026
Author: ecoPrimal (ORCID 0009-0004-2141-0321)

Limitations

• f64 GPU support requires Vulkan shaderFloat64 — some mobile and integrated GPUs lack this
• Shader compilation is pure Rust (coralReef) but currently targets SPIR-V; Metal/DX12 backends are planned
• Performance comparisons to CUDA are indirect — we benchmark scientific output, not raw FLOPS
• No CUDA interop; this is a clean replacement, not a compatibility layer

Repositories:
ecoPrimals/barraCuda · ecoPrimals/toadStool · ecoPrimals/coralReef · syntheticChemistry/hotSpring