Files
ml/hanzo-nn/tests/bf16_add_bisect.rs
zeekayandHanzo Dev 6a363f7ee3 merge: upstream main ddf1b879 and the aarch64 repack branch 35d7ae7c
Three-way per file through the rename: base 39355c6c91, theirs the upstream
merge of main with 35d7ae7c (via 7ef6a694d), ours origin/main.

Lands the barrier thread pool, runtime CPU feature dispatch, aarch64 and x86
repacked quantized kernels (indexed_gemv, gemv_fused_shared_lhs), quantized
embedding on CPU/Metal/CUDA, paged flash-attn kernels, cutile fused MoE, the
in-place op aliasing API, parking_lot storage locks, Metal buffer builders and
debug labels, safetensors 0.8, hf-hub 1.0, tokenizers 0.23.

Ours keeps ROCm, Vulkan, wgpu, the IQ/MXFP4/NVFP4/TQ dtypes, and the i-quant
weight/activation block split in matmul. Our backends ride InplaceOpN and the
quantized embedding; Metal get_rows gains the IQ kernels. ug leaves, as upstream.

Co-authored-by: Hanzo Dev <[email protected]>
2026-09-15 15:52:28 -07:00

18 lines
586 B
Rust

use hanzo_ml::{DType, Device, Result, Tensor};
#[test]
fn bf16_add_sizes() -> Result<()> {
let dev = Device::Cpu;
for n in [1024usize, 32 * 1024, 100 * 1024] {
let x = Tensor::rand(-2f32, 2f32, (n,), &dev)?;
let xb = x.to_dtype(DType::BF16)?;
let y = Tensor::rand(-2f32, 2f32, (n,), &dev)?;
let yb = y.to_dtype(DType::BF16)?;
let want = (&x + &y)?;
let got = (&xb + &yb)?.to_dtype(DType::F32)?;
let d = (&want - &got)?.abs()?.max_all()?.to_scalar::<f32>()?;
println!("n={n}: max diff {d:.4}");
}
Ok(())
}