Gahow Wang
fbd07a578c
New xtrain-tensor crate: DType (F32), shape/stride helpers, Arc-counted host/device Storage with CPU↔CUDA copy, and a contiguous Tensor with creation, host↔device transfer, and a scale() op driving the elementwise kernel. GPU integration tests (host↔device roundtrip + scale correctness) gated behind not(no_cuda); a thin build.rs emits the no_cuda cfg so the kernel call sites compile out locally. Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
27 lines
857 B
Rust
27 lines
857 B
Rust
use std::env;
|
|
use std::path::Path;
|
|
use std::process::Command;
|
|
|
|
// xtrain-tensor calls GPU kernels (via xtrain-cuda's FFI), so it gates those
|
|
// call sites behind `not(no_cuda)` — the same convention xtrain-cuda uses. This
|
|
// build script only detects nvcc and emits that cfg; it compiles no CUDA itself
|
|
// (the kernels are built by xtrain-cuda's build.rs).
|
|
fn main() {
|
|
println!("cargo:rustc-check-cfg=cfg(no_cuda)");
|
|
|
|
let cuda_path = env::var("CUDA_HOME")
|
|
.or_else(|_| env::var("CUDA_PATH"))
|
|
.unwrap_or_else(|_| "/usr/local/cuda".to_string());
|
|
|
|
if !nvcc_available(&cuda_path) {
|
|
println!("cargo:rustc-cfg=no_cuda");
|
|
}
|
|
}
|
|
|
|
fn nvcc_available(cuda_path: &str) -> bool {
|
|
if Command::new("nvcc").arg("--version").output().is_ok() {
|
|
return true;
|
|
}
|
|
Path::new(&format!("{cuda_path}/bin/nvcc")).exists()
|
|
}
|