Update comment in second implementation

Author: AlexanderKalistratov

dpnp/backend/extensions/statistics/sliding_window1d.hpp

@@ -651,11 +651,13 @@ void submit_sliding_window1d_small_kernel(const PaddedSpan<const T, SizeT> &a,
             auto y_start = glid;
             auto y_stop =
                 std::min(y_start + WorkPI * results.size_x(), out.size());
-            int32_t i = 0;
+            uint32_t i = 0;
             for (uint32_t y = y_start; y < y_stop; y += results.size_x()) {
                 out_ptr[y] = results[i++];
             }
-            // due to excessive optimizations this code results in memory
+            // while the code itself seems to be valid, inside correlate
+            // kernel it results in memory corruption. Further investigation
+            // is needed. SAT-7693
             // corruption results.store(&out_ptr[glid],
             //               [out_end](auto &&ptr) { return ptr < out_end; });
         });