build(ml-alpha): bust build.rs cache to force cubin rebuild

Cluster ensure-binary at c3ee5e165 produced a binary whose cubin
lacked the batched kernel symbols (CUDA_ERROR_NOT_FOUND on
cfc_step_batched at trainer init). Local cubins have all symbols —
the cluster's persistent /cargo-target PVC appears to have cached
pre-batched build artifacts that cargo's incremental compilation
deemed up-to-date despite the .cu source changes.

Forcing a fresh build.rs run via a sentinel comment bumps the
build script's fingerprint, which invalidates all build.rs outputs
(cubins) and re-invokes nvcc for every .cu file in KERNELS.

Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
This commit is contained in:
jgrusewski
2026-05-17 10:53:38 +02:00
parent affb0e24cf
commit 8851e98bdc

View File

@@ -17,6 +17,10 @@ const KERNELS: &[&str] = &[
"adamw_step",
];
// Cache bust v3 (2026-05-17): batched cfc/heads kernels + transpose_3d_swap_01
// added in commit 829ddfa62 + c3ee5e165. Force rebuild on cluster nodes whose
// /cargo-target persistent volume still has the pre-batched cubins.
fn main() {
println!("cargo:rerun-if-changed=build.rs");