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/src/lib.rs

@@ -1,4 +1,5 @@
 pub mod activation;
+pub mod attention;
 pub mod embedding;
 pub mod gemm;
 pub mod layernorm;
@@ -6,9 +7,10 @@ pub mod rmsnorm;
 pub mod rope;
 pub mod softmax;
 
-pub use activation::{gelu, silu};
+pub use activation::{gelu, scale, silu};
+pub use attention::attention;
 pub use embedding::embedding;
-pub use gemm::{matmul, GemmBackend};
+pub use gemm::{batched_matmul, matmul, GemmBackend};
 pub use layernorm::layernorm;
 pub use rmsnorm::rmsnorm;
 pub use rope::{rope_inplace, RopeCache};