ferrotensor-backend

The GPU compute layer of the LLM stack in Ferrotensor: a wgpu context (WgpuContext), UnifiedTensor buffers, and the hand-written WGSL kernels in src/*.wgsl behind ComputeEngine (tiled matmul, RoPE, RMSNorm, SiLU, softmax, AdamW and SGD updates, losses, int8/affine dequantisation, flash attention v2 forward and backward), plus a buffer pool. Runs on Metal and Vulkan.

use ferrotensor_backend::{ComputeEngine, UnifiedTensor, WgpuContext};

pollster::block_on(async {
    let ctx = WgpuContext::new().await;
    let engine = ComputeEngine::new(&ctx);
    let a = UnifiedTensor::new(&ctx, &[1., 2., 3., 4.], &[4]);
    let b = UnifiedTensor::new(&ctx, &[5., 6., 7., 8.], &[4]);
    let out = UnifiedTensor::empty(&ctx, &[4]);
    engine.add(&ctx, &a, &b, &out);
    println!("{:?}", out.to_vec(&ctx).await); // [6.0, 8.0, 10.0, 12.0]
});

Runnable versions are in examples/: init_wgpu, tensor_creation, fused_op, ops_test and basic_tensor (each has its real output in its doc comment), for example cargo run -p ferrotensor-backend --example ops_test.

The separate ferrotensor crate has its own small kernel set; this crate is the backend of ferrotensor-autograd, ferrotensor-graph and ferrotensor-trainer.