diff --git a/ggml/src/iqk/iqk_quantize.cpp b/ggml/src/iqk/iqk_quantize.cpp
index a96e559c..7b87f49c 100644
--- a/ggml/src/iqk/iqk_quantize.cpp
+++ b/ggml/src/iqk/iqk_quantize.cpp
@@ -1259,34 +1259,34 @@ void quantize_row_iq2_ks_impl(const float * x, void * vy, int n_per_row, const f
             float best = 0, d = 0;
             bool is_shifted = false;
             float sumqx, sumq2;
-            for (int i1 = 0; i1 < kMax_i1; ++i1) {
-                for (int i2 = i1; i2 < kBlockSize; ++i2) {
-                    for (int i3 = std::max(i2, kMin_i3); i3 < kBlockSize; ++i3) {
-                        sumqx = (sumx[i1] - sumx[ 0])*val[0] + (sumx[i2] - sumx[i1])*val[1]
-                              + (sumx[i3] - sumx[i2])*val[2] + (sumx[kBlockSize] - sumx[i3])*val[3];
-                        sumq2 = (sumw[i1] - sumw[ 0])*val[0]*val[0] + (sumw[i2] - sumw[i1])*val[1]*val[1]
-                              + (sumw[i3] - sumw[i2])*val[2]*val[2] + (sumw[kBlockSize] - sumw[i3])*val[3]*val[3];
+            for (int i1 = 1; i1 < kMax_i1; ++i1) {
+                for (int i3 = std::max(i1, kMin_i3); i3 < kBlockSize; ++i3) {
+                    float sumqx_1 = (sumx[i1] - sumx[0])*val[0]          + (sumx[kBlockSize] - sumx[i3])*val[3];
+                    float sumq2_1 = (sumw[i1] - sumw[0])*val[0]*val[0]   + (sumw[kBlockSize] - sumw[i3])*val[3]*val[3];
+                    float sumqx_2 = (sumx[i1] - sumx[0])*sval[0]         + (sumx[kBlockSize] - sumx[i3])*sval[3];
+                    float sumq2_2 = (sumw[i1] - sumw[0])*sval[0]*sval[0] + (sumw[kBlockSize] - sumw[i3])*sval[3]*sval[3];
+                    float sumqx_3 = (sumx[i1] - sumx[0])*val[3]          + (sumx[kBlockSize] - sumx[i3])*val[0];
+                    float sumq2_3 = (sumw[i1] - sumw[0])*val[3]*val[3]   + (sumw[kBlockSize] - sumw[i3])*val[0]*val[0];
+                    float sumqx_4 = (sumx[i1] - sumx[0])*sval[3]         + (sumx[kBlockSize] - sumx[i3])*sval[0];
+                    float sumq2_4 = (sumw[i1] - sumw[0])*sval[3]*sval[3] + (sumw[kBlockSize] - sumw[i3])*sval[0]*sval[0];
+                    for (int i2 = i1; i2 <= i3; ++i2) {
+                        sumqx = sumqx_1 + (sumx[i2] - sumx[i1])*val[1] + (sumx[i3] - sumx[i2])*val[2];
+                        sumq2 = sumq2_1 + (sumw[i2] - sumw[i1])*val[1]*val[1] + (sumw[i3] - sumw[i2])*val[2]*val[2];
                         if (sumq2 > 0 && sumqx*sumqx > best*sumq2) {
                             d = sumqx/sumq2; best = d*sumqx; is_shifted = false;
                         }
-                        sumqx = (sumx[i1] - sumx[ 0])*sval[0] + (sumx[i2] - sumx[i1])*sval[1]
-                              + (sumx[i3] - sumx[i2])*sval[2] + (sumx[kBlockSize] - sumx[i3])*sval[3];
-                        sumq2 = (sumw[i1] - sumw[ 0])*sval[0]*sval[0] + (sumw[i2] - sumw[i1])*sval[1]*sval[1]
-                              + (sumw[i3] - sumw[i2])*sval[2]*sval[2] + (sumw[kBlockSize] - sumw[i3])*sval[3]*sval[3];
+                        sumqx = sumqx_2 + (sumx[i2] - sumx[i1])*sval[1] + (sumx[i3] - sumx[i2])*sval[2];
+                        sumq2 = sumq2_2 + (sumw[i2] - sumw[i1])*sval[1]*sval[1] + (sumw[i3] - sumw[i2])*sval[2]*sval[2];
                         if (sumq2 > 0 && sumqx*sumqx > best*sumq2) {
                             d = sumqx/sumq2; best = d*sumqx; is_shifted = true;
                         }
-                        sumqx = (sumx[i1] - sumx[ 0])*val[3] + (sumx[i2        ] - sumx[i1])*val[2]
-                              + (sumx[i3] - sumx[i2])*val[1] + (sumx[kBlockSize] - sumx[i3])*val[0];
-                        sumq2 = (sumw[i1] - sumw[ 0])*val[3]*val[3] + (sumw[i2        ] - sumw[i1])*val[2]*val[2]
-                              + (sumw[i3] - sumw[i2])*val[1]*val[1] + (sumw[kBlockSize] - sumw[i3])*val[0]*val[0];
+                        sumqx = sumqx_3 + (sumx[i2] - sumx[i1])*val[2] + (sumx[i3] - sumx[i2])*val[1];
+                        sumq2 = sumq2_3 + (sumw[i2] - sumw[i1])*val[2]*val[2] + (sumw[i3] - sumw[i2])*val[1]*val[1];
                         if (sumq2 > 0 && sumqx*sumqx > best*sumq2) {
                             d = sumqx/sumq2; best = d*sumqx; is_shifted = false;
                         }
-                        sumqx = (sumx[i1] - sumx[ 0])*sval[3] + (sumx[i2        ] - sumx[i1])*sval[2]
-                              + (sumx[i3] - sumx[i2])*sval[1] + (sumx[kBlockSize] - sumx[i3])*sval[0];
-                        sumq2 = (sumw[i1] - sumw[ 0])*sval[3]*sval[3] + (sumw[i2        ] - sumw[i1])*sval[2]*sval[2]
-                              + (sumw[i3] - sumw[i2])*sval[1]*sval[1] + (sumw[kBlockSize] - sumw[i3])*sval[0]*sval[0];
+                        sumqx = sumqx_4 + (sumx[i2] - sumx[i1])*sval[2] + (sumx[i3] - sumx[i2])*sval[1];
+                        sumq2 = sumq2_4 + (sumw[i2] - sumw[i1])*sval[2]*sval[2] + (sumw[i3] - sumw[i2])*sval[1]*sval[1];
                         if (sumq2 > 0 && sumqx*sumqx > best*sumq2) {
                             d = sumqx/sumq2; best = d*sumqx; is_shifted = true;
                         }