diff --git a/lib/algorithms/iterative/ImplicitlyRestartedLanczos.h b/lib/algorithms/iterative/ImplicitlyRestartedLanczos.h
index b20afcb9..7541a544 100644
--- a/lib/algorithms/iterative/ImplicitlyRestartedLanczos.h
+++ b/lib/algorithms/iterative/ImplicitlyRestartedLanczos.h
@@ -267,7 +267,7 @@ until convergence
       Qt = Eigen::MatrixXd::Identity(Nm,Nm);
       for(int ip=k2; ip<Nm; ++ip){ 
 	// Eigen replacement for qr_decomp ???
-        clog << "ckpt B1: shift[" << ip << "] = " << eval2[ip] << endl;
+        clog << "ckpt B1: shift[" << ip << "] = " << eval2[ip] << std::endl;
 	qr_decomp(eval,lme,Nm,Nm,Qt,eval2[ip],k1,Nm);
       }
       
diff --git a/lib/qcd/action/fermion/CayleyFermion5Dvec.cc b/lib/qcd/action/fermion/CayleyFermion5Dvec.cc
index 653e6ab3..d5c05299 100644
--- a/lib/qcd/action/fermion/CayleyFermion5Dvec.cc
+++ b/lib/qcd/action/fermion/CayleyFermion5Dvec.cc
@@ -469,7 +469,8 @@ void CayleyFermion5D<Impl>::MooeeInternalAsm(const FermionField &psi, FermionFie
 	}
 	a0 = a0+incr;
 	a1 = a1+incr;
-	a2 = a2+sizeof(Simd::scalar_type);
+	//a2 = a2+sizeof(Simd::scalar_type);
+	a2 = a2+sizeof(typename Simd::scalar_type); //ypj [debug]
       }}
     {
       int lexa = s1+LLs*site;
@@ -701,7 +702,8 @@ void CayleyFermion5D<Impl>::MooeeInternalZAsm(const FermionField &psi, FermionFi
 	}
 	a0 = a0+incr;
 	a1 = a1+incr;
-	a2 = a2+sizeof(Simd::scalar_type);
+	//a2 = a2+sizeof(Simd::scalar_type);
+	a2 = a2+sizeof(typename Simd::scalar_type); // ypj [debug]
       }}
     {
       int lexa = s1+LLs*site;
diff --git a/lib/qcd/action/fermion/DomainWallEOFAFermionvec.cc b/lib/qcd/action/fermion/DomainWallEOFAFermionvec.cc
index 81ce448c..dfd7fd83 100644
--- a/lib/qcd/action/fermion/DomainWallEOFAFermionvec.cc
+++ b/lib/qcd/action/fermion/DomainWallEOFAFermionvec.cc
@@ -475,7 +475,8 @@ namespace QCD {
                         }
                         a0 = a0 + incr;
                         a1 = a1 + incr;
-                        a2 = a2 + sizeof(Simd::scalar_type);
+                        //a2 = a2 + sizeof(Simd::scalar_type);
+                        a2 = a2 + sizeof(typename Simd::scalar_type); // ypj [debug]
                     }
                 }
 
diff --git a/lib/qcd/action/fermion/MobiusEOFAFermionvec.cc b/lib/qcd/action/fermion/MobiusEOFAFermionvec.cc
index c4eaf0f3..3a1f4ef5 100644
--- a/lib/qcd/action/fermion/MobiusEOFAFermionvec.cc
+++ b/lib/qcd/action/fermion/MobiusEOFAFermionvec.cc
@@ -853,7 +853,8 @@ namespace QCD {
 
               a0 = a0 + incr;
               a1 = a1 + incr;
-              a2 = a2 + sizeof(Simd::scalar_type);
+              //a2 = a2 + sizeof(Simd::scalar_type);
+              a2 = a2 + sizeof(typename Simd::scalar_type); // ypj [debug]
             }
           }
 
diff --git a/lib/simd/Grid_avx512.h b/lib/simd/Grid_avx512.h
index 85d27421..10ea6dc6 100644
--- a/lib/simd/Grid_avx512.h
+++ b/lib/simd/Grid_avx512.h
@@ -556,7 +556,7 @@ namespace Optimization {
     v3  = _mm256_add_epi32(v1, v2);
     v1  = _mm256_hadd_epi32(v3, v3);
     v2  = _mm256_hadd_epi32(v1, v1);
-    u1  = _mm256_castsi256_si128(v2)        // upper half
+    u1  = _mm256_castsi256_si128(v2);        // upper half ypj[debug] ; was missing
     u2  = _mm256_extracti128_si256(v2, 1);  // lower half
     ret = _mm_add_epi32(u1, u2);
     return _mm_cvtsi128_si32(ret);