diff --git a/src/coreclr/jit/emit.h b/src/coreclr/jit/emit.h
index 3327e4bb2cdf41..05246d737bd6ea 100644
--- a/src/coreclr/jit/emit.h
+++ b/src/coreclr/jit/emit.h
@@ -2088,6 +2088,7 @@ class emitter
 #define PERFSCORE_LATENCY_14C  14.0f
 #define PERFSCORE_LATENCY_15C  15.0f
 #define PERFSCORE_LATENCY_16C  16.0f
+#define PERFSCORE_LATENCY_17C  17.0f
 #define PERFSCORE_LATENCY_18C  18.0f
 #define PERFSCORE_LATENCY_20C  20.0f
 #define PERFSCORE_LATENCY_22C  22.0f
diff --git a/src/coreclr/jit/emitxarch.cpp b/src/coreclr/jit/emitxarch.cpp
index ff65e2d883a25f..c7380c61793dc8 100644
--- a/src/coreclr/jit/emitxarch.cpp
+++ b/src/coreclr/jit/emitxarch.cpp
@@ -20913,7 +20913,7 @@ emitter::insExecutionCharacteristics emitter::getInsExecutionCharacteristics(ins
             {
                 // ins   reg, mem
                 result.insThroughput = PERFSCORE_THROUGHPUT_2X;
-                result.insLatency += opSize == EA_32BYTE ? PERFSCORE_LATENCY_3C : PERFSCORE_LATENCY_2C;
+                result.insLatency += opSize == EA_16BYTE ? PERFSCORE_LATENCY_2C : PERFSCORE_LATENCY_3C;
             }
             else
             {
@@ -20943,11 +20943,12 @@ emitter::insExecutionCharacteristics emitter::getInsExecutionCharacteristics(ins
             break;
         }
 
+        case INS_lddqu:
         case INS_movntdqa:
         {
             assert(memAccessKind == PERFSCORE_MEMORY_READ);
             result.insThroughput = PERFSCORE_THROUGHPUT_2X;
-            result.insLatency += opSize == EA_32BYTE ? PERFSCORE_LATENCY_3C : PERFSCORE_LATENCY_2C;
+            result.insLatency += opSize == EA_16BYTE ? PERFSCORE_LATENCY_2C : PERFSCORE_LATENCY_3C;
             break;
         }
 
@@ -20963,7 +20964,7 @@ emitter::insExecutionCharacteristics emitter::getInsExecutionCharacteristics(ins
             else if (memAccessKind == PERFSCORE_MEMORY_READ)
             {
                 result.insThroughput = PERFSCORE_THROUGHPUT_2X;
-                result.insLatency += opSize == EA_32BYTE ? PERFSCORE_LATENCY_3C : PERFSCORE_LATENCY_2C;
+                result.insLatency += (EA_SIZE_IN_BYTES(opSize) <= 16) ? PERFSCORE_LATENCY_2C : PERFSCORE_LATENCY_3C;
             }
             else
             {
@@ -20973,13 +20974,6 @@ emitter::insExecutionCharacteristics emitter::getInsExecutionCharacteristics(ins
             break;
         }
 
-        case INS_lddqu:
-        {
-            result.insThroughput = PERFSCORE_THROUGHPUT_2X;
-            result.insLatency += opSize == EA_32BYTE ? PERFSCORE_LATENCY_3C : PERFSCORE_LATENCY_2C;
-            break;
-        }
-
         case INS_vpmovdb:
         case INS_vpmovdw:
         case INS_vpmovqb:
@@ -21147,7 +21141,7 @@ emitter::insExecutionCharacteristics emitter::getInsExecutionCharacteristics(ins
             }
             else
             {
-                result.insLatency += opSize == EA_32BYTE ? PERFSCORE_LATENCY_4C : PERFSCORE_LATENCY_2C;
+                result.insLatency += opSize == EA_16BYTE ? PERFSCORE_LATENCY_2C : PERFSCORE_LATENCY_4C;
                 result.insThroughput = PERFSCORE_THROUGHPUT_1C;
             }
             break;
@@ -21231,7 +21225,7 @@ emitter::insExecutionCharacteristics emitter::getInsExecutionCharacteristics(ins
         case INS_pmovzxdq:
         {
             result.insThroughput = PERFSCORE_THROUGHPUT_1C;
-            result.insLatency += opSize == EA_32BYTE ? PERFSCORE_LATENCY_3C : PERFSCORE_LATENCY_1C;
+            result.insLatency += opSize == EA_16BYTE ? PERFSCORE_LATENCY_1C : PERFSCORE_LATENCY_3C;
             break;
         }
 
@@ -21259,7 +21253,7 @@ emitter::insExecutionCharacteristics emitter::getInsExecutionCharacteristics(ins
         case INS_vcvtuqq2ps:
         {
             result.insThroughput = PERFSCORE_THROUGHPUT_1C;
-            result.insLatency += opSize == EA_32BYTE ? PERFSCORE_LATENCY_7C : PERFSCORE_LATENCY_5C;
+            result.insLatency += opSize == EA_16BYTE ? PERFSCORE_LATENCY_5C : PERFSCORE_LATENCY_7C;
             break;
         }
 
@@ -21349,8 +21343,22 @@ emitter::insExecutionCharacteristics emitter::getInsExecutionCharacteristics(ins
         case INS_vpgatherdd:
         case INS_vgatherdps:
         {
-            result.insThroughput = PERFSCORE_THROUGHPUT_4C;
-            result.insLatency += opSize == EA_32BYTE ? PERFSCORE_LATENCY_13C : PERFSCORE_LATENCY_11C;
+            if (opSize == EA_16BYTE)
+            {
+                result.insThroughput = PERFSCORE_THROUGHPUT_2C;
+                result.insLatency += PERFSCORE_LATENCY_11C;
+            }
+            else if (opSize == EA_32BYTE)
+            {
+                result.insThroughput = PERFSCORE_THROUGHPUT_4C;
+                result.insLatency += PERFSCORE_LATENCY_13C;
+            }
+            else
+            {
+                assert(opSize == EA_64BYTE);
+                result.insThroughput = PERFSCORE_THROUGHPUT_8C;
+                result.insLatency += PERFSCORE_LATENCY_17C;
+            }
             break;
         }
 
@@ -21361,8 +21369,22 @@ emitter::insExecutionCharacteristics emitter::getInsExecutionCharacteristics(ins
         case INS_vgatherqps:
         case INS_vgatherqpd:
         {
-            result.insThroughput = PERFSCORE_THROUGHPUT_4C;
-            result.insLatency += opSize == EA_32BYTE ? PERFSCORE_LATENCY_11C : PERFSCORE_LATENCY_9C;
+            if (opSize == EA_16BYTE)
+            {
+                result.insThroughput = PERFSCORE_THROUGHPUT_1C;
+                result.insLatency += PERFSCORE_LATENCY_9C;
+            }
+            else if (opSize == EA_32BYTE)
+            {
+                result.insThroughput = PERFSCORE_THROUGHPUT_2C;
+                result.insLatency += PERFSCORE_LATENCY_11C;
+            }
+            else
+            {
+                assert(opSize == EA_64BYTE);
+                result.insThroughput = PERFSCORE_THROUGHPUT_4C;
+                result.insLatency += PERFSCORE_LATENCY_13C;
+            }
             break;
         }