diff --git a/gen/gen_avx512.cpp b/gen/gen_avx512.cpp
index 1a3b469..4969921 100644
--- a/gen/gen_avx512.cpp
+++ b/gen/gen_avx512.cpp
@@ -363,8 +363,11 @@
 		{ 0x73, "vpshrdvd", T_66 | T_0F38 | T_YMM | T_MUST_EVEX | T_EW0 | T_SAE_Z, false },
 		{ 0x73, "vpshrdvq", T_66 | T_0F38 | T_YMM | T_MUST_EVEX | T_EW1 | T_SAE_Z, false },
 
-		{ 0x50, "vpdpbusd", T_66 | T_0F38 | T_YMM | T_MUST_EVEX | T_EW0 | T_SAE_Z, false },
-		{ 0x51, "vpdpbusds", T_66 | T_0F38 | T_YMM | T_MUST_EVEX | T_EW0 | T_SAE_Z, false },
+		{ 0x50, "vpdpbusd", T_66 | T_0F38 | T_YMM | T_MUST_EVEX | T_EW0 | T_SAE_Z | T_B32, false },
+		{ 0x51, "vpdpbusds", T_66 | T_0F38 | T_YMM | T_MUST_EVEX | T_EW0 | T_SAE_Z | T_B32, false },
+
+		{ 0x52, "vpdpwssd", T_66 | T_0F38 | T_YMM | T_MUST_EVEX | T_EW0 | T_SAE_Z | T_B32, false },
+		{ 0x53, "vpdpwssds", T_66 | T_0F38 | T_YMM | T_MUST_EVEX | T_EW0 | T_SAE_Z | T_B32, false },
 	};
 	for (size_t i = 0; i < NUM_OF_ARRAY(tbl); i++) {
 		const Tbl *p = &tbl[i];
diff --git a/test/misc.cpp b/test/misc.cpp
index 461e995..a786690 100644
--- a/test/misc.cpp
+++ b/test/misc.cpp
@@ -408,9 +408,33 @@
 			vpdpbusd(ymm5|k3|T_z, ymm20, ptr [rax + 0x40]);
 			vpdpbusd(zmm5|k3|T_z, zmm20, ptr [rax + 0x40]);
 
+			vpdpbusd(xmm5|k3|T_z, xmm20, ptr_b [rax + 0x40]);
+			vpdpbusd(ymm5|k3|T_z, ymm20, ptr_b [rax + 0x40]);
+			vpdpbusd(zmm5|k3|T_z, zmm20, ptr_b [rax + 0x40]);
+
 			vpdpbusds(xmm5|k3|T_z, xmm20, ptr [rax + 0x40]);
 			vpdpbusds(ymm5|k3|T_z, ymm20, ptr [rax + 0x40]);
 			vpdpbusds(zmm5|k3|T_z, zmm20, ptr [rax + 0x40]);
+
+			vpdpbusds(xmm5|k3|T_z, xmm20, ptr_b [rax + 0x40]);
+			vpdpbusds(ymm5|k3|T_z, ymm20, ptr_b [rax + 0x40]);
+			vpdpbusds(zmm5|k3|T_z, zmm20, ptr_b [rax + 0x40]);
+
+			vpdpwssd(xmm5|k3|T_z, xmm20, ptr [rax + 0x40]);
+			vpdpwssd(ymm5|k3|T_z, ymm20, ptr [rax + 0x40]);
+			vpdpwssd(zmm5|k3|T_z, zmm20, ptr [rax + 0x40]);
+
+			vpdpwssd(xmm5|k3|T_z, xmm20, ptr_b [rax + 0x40]);
+			vpdpwssd(ymm5|k3|T_z, ymm20, ptr_b [rax + 0x40]);
+			vpdpwssd(zmm5|k3|T_z, zmm20, ptr_b [rax + 0x40]);
+
+			vpdpwssds(xmm5|k3|T_z, xmm20, ptr [rax + 0x40]);
+			vpdpwssds(ymm5|k3|T_z, ymm20, ptr [rax + 0x40]);
+			vpdpwssds(zmm5|k3|T_z, zmm20, ptr [rax + 0x40]);
+
+			vpdpwssds(xmm5|k3|T_z, xmm20, ptr_b [rax + 0x40]);
+			vpdpwssds(ymm5|k3|T_z, ymm20, ptr_b [rax + 0x40]);
+			vpdpwssds(zmm5|k3|T_z, zmm20, ptr_b [rax + 0x40]);
 		}
 	} c;
 	const uint8_t tbl[] = {
@@ -418,9 +442,33 @@
 		0x62, 0xf2, 0x5d, 0xa3, 0x50, 0x68, 0x02,
 		0x62, 0xf2, 0x5d, 0xc3, 0x50, 0x68, 0x01,
 
+		0x62, 0xf2, 0x5d, 0x93, 0x50, 0x68, 0x10,
+		0x62, 0xf2, 0x5d, 0xb3, 0x50, 0x68, 0x10,
+		0x62, 0xf2, 0x5d, 0xd3, 0x50, 0x68, 0x10,
+
 		0x62, 0xf2, 0x5d, 0x83, 0x51, 0x68, 0x04,
 		0x62, 0xf2, 0x5d, 0xa3, 0x51, 0x68, 0x02,
 		0x62, 0xf2, 0x5d, 0xc3, 0x51, 0x68, 0x01,
+
+		0x62, 0xf2, 0x5d, 0x93, 0x51, 0x68, 0x10,
+		0x62, 0xf2, 0x5d, 0xb3, 0x51, 0x68, 0x10,
+		0x62, 0xf2, 0x5d, 0xd3, 0x51, 0x68, 0x10,
+
+		0x62, 0xf2, 0x5d, 0x83, 0x52, 0x68, 0x04,
+		0x62, 0xf2, 0x5d, 0xa3, 0x52, 0x68, 0x02,
+		0x62, 0xf2, 0x5d, 0xc3, 0x52, 0x68, 0x01,
+
+		0x62, 0xf2, 0x5d, 0x93, 0x52, 0x68, 0x10,
+		0x62, 0xf2, 0x5d, 0xb3, 0x52, 0x68, 0x10,
+		0x62, 0xf2, 0x5d, 0xd3, 0x52, 0x68, 0x10,
+
+		0x62, 0xf2, 0x5d, 0x83, 0x53, 0x68, 0x04,
+		0x62, 0xf2, 0x5d, 0xa3, 0x53, 0x68, 0x02,
+		0x62, 0xf2, 0x5d, 0xc3, 0x53, 0x68, 0x01,
+
+		0x62, 0xf2, 0x5d, 0x93, 0x53, 0x68, 0x10,
+		0x62, 0xf2, 0x5d, 0xb3, 0x53, 0x68, 0x10,
+		0x62, 0xf2, 0x5d, 0xd3, 0x53, 0x68, 0x10,
 	};
 	const size_t n = sizeof(tbl) / sizeof(tbl[0]);
 	CYBOZU_TEST_EQUAL(c.getSize(), n);
diff --git a/xbyak/xbyak_mnemonic.h b/xbyak/xbyak_mnemonic.h
index f1f50e0..861ba2b 100644
--- a/xbyak/xbyak_mnemonic.h
+++ b/xbyak/xbyak_mnemonic.h
@@ -1787,8 +1787,10 @@
 void vpcompressq(const Operand& op, const Xmm& x) { opAVX_X_XM_IMM(x, op, T_N8 | T_66 | T_0F38 | T_EW1 | T_YMM | T_MUST_EVEX, 0x8B); }
 void vpconflictd(const Xmm& x, const Operand& op) { opAVX_X_XM_IMM(x, op, T_66 | T_0F38 | T_EW0 | T_YMM | T_MUST_EVEX | T_B32, 0xC4); }
 void vpconflictq(const Xmm& x, const Operand& op) { opAVX_X_XM_IMM(x, op, T_66 | T_0F38 | T_EW1 | T_YMM | T_MUST_EVEX | T_B64, 0xC4); }
-void vpdpbusd(const Xmm& x1, const Xmm& x2, const Operand& op) { opAVX_X_X_XM(x1, x2, op, T_66 | T_0F38 | T_EW0 | T_YMM | T_SAE_Z | T_MUST_EVEX, 0x50); }
-void vpdpbusds(const Xmm& x1, const Xmm& x2, const Operand& op) { opAVX_X_X_XM(x1, x2, op, T_66 | T_0F38 | T_EW0 | T_YMM | T_SAE_Z | T_MUST_EVEX, 0x51); }
+void vpdpbusd(const Xmm& x1, const Xmm& x2, const Operand& op) { opAVX_X_X_XM(x1, x2, op, T_66 | T_0F38 | T_EW0 | T_YMM | T_SAE_Z | T_MUST_EVEX | T_B32, 0x50); }
+void vpdpbusds(const Xmm& x1, const Xmm& x2, const Operand& op) { opAVX_X_X_XM(x1, x2, op, T_66 | T_0F38 | T_EW0 | T_YMM | T_SAE_Z | T_MUST_EVEX | T_B32, 0x51); }
+void vpdpwssd(const Xmm& x1, const Xmm& x2, const Operand& op) { opAVX_X_X_XM(x1, x2, op, T_66 | T_0F38 | T_EW0 | T_YMM | T_SAE_Z | T_MUST_EVEX | T_B32, 0x52); }
+void vpdpwssds(const Xmm& x1, const Xmm& x2, const Operand& op) { opAVX_X_X_XM(x1, x2, op, T_66 | T_0F38 | T_EW0 | T_YMM | T_SAE_Z | T_MUST_EVEX | T_B32, 0x53); }
 void vpermb(const Xmm& x1, const Xmm& x2, const Operand& op) { opAVX_X_X_XM(x1, x2, op, T_66 | T_0F38 | T_EW0 | T_YMM | T_MUST_EVEX, 0x8D); }
 void vpermi2b(const Xmm& x1, const Xmm& x2, const Operand& op) { opAVX_X_X_XM(x1, x2, op, T_66 | T_0F38 | T_EW0 | T_YMM | T_MUST_EVEX, 0x75); }
 void vpermi2d(const Xmm& x1, const Xmm& x2, const Operand& op) { opAVX_X_X_XM(x1, x2, op, T_66 | T_0F38 | T_EW0 | T_YMM | T_MUST_EVEX | T_B32, 0x76); }