style: cargo fmt --workspace (whitespace/wrapping only, no semantic change)

Whole-workspace rustfmt pass picked up while iterating on Mamba GPU
backward work. Verified formatting-only via diff sampling; no logic
changed.

Co-Authored-By: Claude Sonnet 5 <[email protected]>

demos/rtx-distllm-demo/src/attention.rs

@@ -100,11 +100,9 @@ impl DistributedAttention {
         let device = Device::cpu();
         let d_model = self.num_heads * self.head_dim;
 
-        let query =
-            Tensor::randn(&[1, seq_len, d_model], &device).map_err(|e| e.to_string())?;
+        let query = Tensor::randn(&[1, seq_len, d_model], &device).map_err(|e| e.to_string())?;
         let key = Tensor::randn(&[1, seq_len, d_model], &device).map_err(|e| e.to_string())?;
-        let value =
-            Tensor::randn(&[1, seq_len, d_model], &device).map_err(|e| e.to_string())?;
+        let value = Tensor::randn(&[1, seq_len, d_model], &device).map_err(|e| e.to_string())?;
 
         let output = query
             .scaled_dot_product_attention(&key, &value, None)
@@ -563,7 +561,9 @@ mod tests {
     #[test]
     fn test_attention_forward() {
         let attn = DistributedAttention::new(4, 4, 16, 1);
-        let (output, elapsed) = attn.forward(32, 0).expect("real attention forward should succeed");
+        let (output, elapsed) = attn
+            .forward(32, 0)
+            .expect("real attention forward should succeed");
         // Output is [batch=1, seq_len, num_heads * head_dim] = [1, 32, 64].
         assert_eq!(output.shape().dims(), &[1, 32, 64]);
         assert!(elapsed.as_nanos() > 0 || elapsed.as_nanos() == 0); // measured, not fabricated