Ready for fvm steady case

include/numerics/matvec.h

@@ -38,16 +38,24 @@ namespace numerics{
 	    const uint64_t m = A.rows();
 	    const uint64_t n = A.cols();
 
-	    utils::Vector<T> y(m, T{0});
+	    utils::Vector<T> y(m, T{0}); // <-- y has length m (rows)
 
+
+        const T* xptr = x.data();
+        const T* Aptr = A.data();	// row-major: A(i,j) == Aptr[i*n + j]
+
+        // Each row i is an independent dot product: y[i] = dot(A[i,*], x)
 	    #pragma omp parallel for schedule(static)
 	    for (uint64_t i = 0; i < m; ++i) {
-	        T acc = T{0};
+            const T* row = Aptr + i * n;     // contiguous row i
+            T acc = T{0};
+            #pragma omp simd reduction(+:acc)
 	        for (uint64_t j = 0; j < n; ++j) {
-	            acc += A(i, j) * x[j];
+	            acc += row[j] * xptr[j];
 	        }
 	        y[i] = acc;
-	    }
+		}
+
 	    return y;
 	}