Fix: Increase N for statistically significant performance benchmarking

Author: 0x822a5b87

challenges/medium/17_dot_product/challenge.html

@@ -27,5 +27,5 @@ <h2>Constraints</h2>
     <li><code>A</code> and <code>B</code> have identical lengths</li>
     <li>1 ≤ <code>N</code> ≤ 100,000,000</li>
 
-  <li>Performance is measured with <code>N</code> = 5</li>
+  <li>Performance is measured with <code>N</code> = 4194304</li>
 </ul>

challenges/medium/17_dot_product/challenge.py

@@ -105,7 +105,7 @@ def generate_functional_test(self) -> List[Dict[str, Any]]:
 
     def generate_performance_test(self) -> Dict[str, Any]:
         dtype = torch.float32
-        N = 5
+        N = 4194304
         A = torch.empty(N, device="cuda", dtype=dtype).uniform_(-1.0, 1.0)
         B = torch.empty(N, device="cuda", dtype=dtype).uniform_(-1.0, 1.0)
         result = torch.zeros(1, device="cuda", dtype=dtype)