phase 5: naive multi-head attention

- Batched GEMM via cublasGemmStridedBatchedEx
- Causal mask CUDA kernel (F32 + BF16)
- Element-wise scale CUDA kernel (F32 + BF16)
- attention() composing: batched_matmul + scale + causal_mask + softmax
- Fixed to_device/contiguous infinite recursion (GPU contiguous via CPU round-trip)
- 5 attention tests passing (max_err < 3e-7 F32)
- Total: 61 tests passing across all crates

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>

crates/xserv-kernels/build.rs

@@ -22,6 +22,7 @@ fn main() {
         .file("../../csrc/reduce/softmax.cu")
         .file("../../csrc/embedding/embedding.cu")
         .file("../../csrc/embedding/rope.cu")
+        .file("../../csrc/attention/causal_mask.cu")
         .compile("xserv_kernels");
 
     println!("cargo:rerun-if-changed=../../csrc/");