add vgetmant*, vscale* vrsqrt*

3 files changed, 171 insertions, 11 deletions

diff --git a/gen/gen_avx512.cpp b/gen/gen_avx512.cpp
index 42b2709..715ed65 100644
--- a/gen/gen_avx512.cpp
+++ b/gen/gen_avx512.cpp
@@ -174,6 +174,8 @@ void putX_XM()
 
 		{ 0x89, "vpexpandd", T_66 | T_0F38 | T_MUST_EVEX | T_YMM | T_EW0 | T_N4 },
 		{ 0x89, "vpexpandq", T_66 | T_0F38 | T_MUST_EVEX | T_YMM | T_EW1 | T_N8 },
+		{ 0x42, "vgetexppd", T_66 | T_0F38 | T_MUST_EVEX | T_YMM | T_EW1 | T_B64 | T_SAE_Z },
+		{ 0x42, "vgetexpps", T_66 | T_0F38 | T_MUST_EVEX | T_YMM | T_EW0 | T_B32 | T_SAE_Z },
 	};
 	for (size_t i = 0; i < NUM_OF_ARRAY(tbl); i++) {
 		const Tbl *p = &tbl[i];
@@ -280,6 +282,30 @@ void putX_X_XM_IMM()
 
 		{ 0x25, "vpternlogd", T_66 | T_0F3A | T_YMM | T_MUST_EVEX | T_EW0 | T_B32, true },
 		{ 0x25, "vpternlogq", T_66 | T_0F3A | T_YMM | T_MUST_EVEX | T_EW1 | T_B64, true },
+
+		{ 0x43, "vgetexpsd", T_66 | T_0F38 | T_MUST_EVEX | T_EW1 | T_SAE_X | T_N8, false },
+		{ 0x43, "vgetexpss", T_66 | T_0F38 | T_MUST_EVEX | T_EW0 | T_SAE_X | T_N4, false },
+		{ 0x27, "vgetmantsd", T_66 | T_0F3A | T_MUST_EVEX | T_EW1 | T_SAE_X | T_N8, true },
+		{ 0x27, "vgetmantss", T_66 | T_0F3A | T_MUST_EVEX | T_EW0 | T_SAE_X | T_N4, true },
+
+		{ 0x54, "vfixupimmpd", T_66 | T_0F3A | T_MUST_EVEX | T_YMM | T_EW1 | T_B64 | T_SAE_Z, true },
+		{ 0x54, "vfixupimmps", T_66 | T_0F3A | T_MUST_EVEX | T_YMM | T_EW0 | T_B32 | T_SAE_Z, true },
+		{ 0x55, "vfixupimmsd", T_66 | T_0F3A | T_MUST_EVEX | T_EW1 | T_SAE_Z | T_N8, true },
+		{ 0x55, "vfixupimmss", T_66 | T_0F3A | T_MUST_EVEX | T_EW0 | T_SAE_Z | T_N4, true },
+
+		{ 0x4D, "vrcp14sd", T_66 | T_0F38 | T_MUST_EVEX | T_EW1 | T_N8, false },
+		{ 0x4D, "vrcp14ss", T_66 | T_0F38 | T_MUST_EVEX | T_EW0 | T_N4, false },
+
+		{ 0x4F, "vrsqrt14sd", T_66 | T_0F38 | T_YMM | T_MUST_EVEX | T_EW1 | T_N8, false },
+		{ 0x4F, "vrsqrt14ss", T_66 | T_0F38 | T_YMM | T_MUST_EVEX | T_EW0 | T_N4, false },
+
+		{ 0x0B, "vrndscalesd", T_66 | T_0F3A | T_MUST_EVEX | T_EW1 | T_N8, true },
+		{ 0x0A, "vrndscaless", T_66 | T_0F3A | T_MUST_EVEX | T_EW0 | T_N4, true },
+
+		{ 0x2C, "vscalefpd", T_66 | T_0F38 | T_YMM | T_MUST_EVEX | T_EW1 | T_B64 | T_ER_Z, false },
+		{ 0x2C, "vscalefps", T_66 | T_0F38 | T_YMM | T_MUST_EVEX | T_EW0 | T_B32 | T_ER_Z, false },
+		{ 0x2D, "vscalefsd", T_66 | T_0F38 | T_MUST_EVEX | T_EW1 | T_ER_X | T_N8, false },
+		{ 0x2D, "vscalefss", T_66 | T_0F38 | T_MUST_EVEX | T_EW0 | T_ER_X | T_N4, false },
 	};
 	for (size_t i = 0; i < NUM_OF_ARRAY(tbl); i++) {
 		const Tbl *p = &tbl[i];
@@ -512,6 +538,33 @@ void putMov()
 	}
 }
 
+void putX_XM_IMM()
+{
+	const struct Tbl {
+		uint8 code;
+		const char *name;
+		int type;
+		bool hasIMM;
+	} tbl[] = {
+		{ 0x26, "vgetmantpd", T_66 | T_0F3A | T_YMM | T_MUST_EVEX | T_EW1 | T_B64 | T_SAE_Z, true },
+		{ 0x26, "vgetmantps", T_66 | T_0F3A | T_YMM | T_MUST_EVEX | T_EW0 | T_B32 | T_SAE_Z, true },
+		{ 0x4C, "vrcp14pd", T_66 | T_0F38 | T_YMM | T_MUST_EVEX | T_EW1 | T_B64, false },
+		{ 0x4C, "vrcp14ps", T_66 | T_0F38 | T_YMM | T_MUST_EVEX | T_EW0 | T_B32, false },
+
+		{ 0x4E, "vrsqrt14pd", T_66 | T_0F38 | T_YMM | T_MUST_EVEX | T_EW1 | T_B64, false },
+		{ 0x4E, "vrsqrt14ps", T_66 | T_0F38 | T_YMM | T_MUST_EVEX | T_EW0 | T_B32, false },
+
+		{ 0x09, "vrndscalepd", T_66 | T_0F3A | T_YMM | T_MUST_EVEX | T_EW1 | T_B64, true },
+		{ 0x08, "vrndscaleps", T_66 | T_0F3A | T_YMM | T_MUST_EVEX | T_EW0 | T_B32, true },
+
+	};
+	for (size_t i = 0; i < NUM_OF_ARRAY(tbl); i++) {
+		const Tbl *p = &tbl[i];
+		std::string type = type2String(p->type);
+		printf("void %s(const Xmm& x, const Operand& op%s) { opAVX_X_XM_IMM(x, op, %s, 0x%02X%s); }\n"
+			, p->name, p->hasIMM ? ", uint8 imm" : "", type.c_str(), p->code, p->hasIMM ? ", imm" : "");
+	}
+}
 int main()
 {
 	puts("#ifndef XBYAK_DISABLE_AVX512");
@@ -530,5 +583,6 @@ int main()
 	putGather();
 	putShuff();
 	putMov();
+	putX_XM_IMM();
 	puts("#endif");
 }
diff --git a/test/make_512.cpp b/test/make_512.cpp
index db2d9d7..589ed6a 100644
--- a/test/make_512.cpp
+++ b/test/make_512.cpp
@@ -1742,16 +1742,6 @@ public:
 		put("vshufi64x2", YMM_KZ, _YMM, _YMM | _MEM | M_1to4, IMM8);
 		put("vshufi64x2", ZMM_KZ, _ZMM, _ZMM | _MEM | M_1to8, IMM8);
 	}
-	void putMisc2()
-	{
-		put("vpternlogd", XMM_KZ, _XMM, _XMM | _MEM | M_1to4, IMM8);
-		put("vpternlogd", YMM_KZ, _YMM, _YMM | _MEM | M_1to8, IMM8);
-		put("vpternlogd", ZMM_KZ, _ZMM, _ZMM | _MEM | M_1to16, IMM8);
-
-		put("vpternlogq", XMM_KZ, _XMM, _XMM | _MEM | M_1to2, IMM8);
-		put("vpternlogq", YMM_KZ, _YMM, _YMM | _MEM | M_1to4, IMM8);
-		put("vpternlogq", ZMM_KZ, _ZMM, _ZMM | _MEM | M_1to8, IMM8);
-	}
 	void putMov()
 	{
 		put("vpmovm2b", _XMM | _YMM | _ZMM, K);
@@ -1803,10 +1793,98 @@ public:
 		put("vpmovuswb", XMM_KZ | _MEM, _XMM | _YMM);
 		put("vpmovuswb", YMM_KZ | _MEM, _ZMM);
 	}
+	void putMisc2()
+	{
+		put("vpternlogd", XMM_KZ, _XMM, _XMM | _MEM | M_1to4, IMM8);
+		put("vpternlogd", YMM_KZ, _YMM, _YMM | _MEM | M_1to8, IMM8);
+		put("vpternlogd", ZMM_KZ, _ZMM, _ZMM | _MEM | M_1to16, IMM8);
+
+		put("vpternlogq", XMM_KZ, _XMM, _XMM | _MEM | M_1to2, IMM8);
+		put("vpternlogq", YMM_KZ, _YMM, _YMM | _MEM | M_1to4, IMM8);
+		put("vpternlogq", ZMM_KZ, _ZMM, _ZMM | _MEM | M_1to8, IMM8);
+
+		put("vgetexppd", XMM_KZ, _XMM | MEM | M_1to2);
+		put("vgetexppd", YMM_KZ, _YMM | MEM | M_1to4);
+		put("vgetexppd", ZMM_KZ, _ZMM | MEM | M_1to8 | ZMM_SAE);
+
+		put("vgetexpps", XMM_KZ, _XMM | MEM | M_1to4);
+		put("vgetexpps", YMM_KZ, _YMM | MEM | M_1to8);
+		put("vgetexpps", ZMM_KZ, _ZMM | MEM | M_1to16 | ZMM_SAE);
+
+		put("vgetexpsd", XMM_KZ, _XMM, _XMM | _MEM | XMM_SAE);
+		put("vgetexpss", XMM_KZ, _XMM, _XMM | _MEM | XMM_SAE);
+
+		put("vgetmantpd", XMM_KZ, _XMM | _MEM | M_1to2, IMM8);
+		put("vgetmantpd", YMM_KZ, _YMM | _MEM | M_1to4, IMM8);
+		put("vgetmantpd", ZMM_KZ, _ZMM | _MEM | M_1to8, IMM8);
+
+		put("vgetmantps", XMM_KZ, _XMM | _MEM | M_1to4, IMM8);
+		put("vgetmantps", YMM_KZ, _YMM | _MEM | M_1to8, IMM8);
+		put("vgetmantps", ZMM_KZ, _ZMM | _MEM | M_1to16, IMM8);
+
+		put("vgetmantsd", XMM_KZ, _XMM, _XMM | _MEM | XMM_SAE, IMM8);
+		put("vgetmantss", XMM_KZ, _XMM, _XMM | _MEM | XMM_SAE, IMM8);
+
+		put("vfixupimmpd", XMM_KZ, _XMM, _XMM | _MEM | M_1to2, IMM8);
+		put("vfixupimmpd", YMM_KZ, _YMM, _YMM | _MEM | M_1to4, IMM8);
+		put("vfixupimmpd", ZMM_KZ, _ZMM, _ZMM | _MEM | M_1to8, IMM8);
+
+		put("vfixupimmps", XMM_KZ, _XMM, _XMM | _MEM | M_1to4, IMM8);
+		put("vfixupimmps", YMM_KZ, _YMM, _YMM | _MEM | M_1to8, IMM8);
+		put("vfixupimmps", ZMM_KZ, _ZMM, _ZMM | _MEM | M_1to16, IMM8);
+
+		put("vfixupimmsd", XMM_KZ, _XMM, _XMM | _MEM, IMM8);
+		put("vfixupimmss", XMM_KZ, _XMM, _XMM | _MEM, IMM8);
+
+		put("vrcp14pd", XMM_KZ, _XMM | _MEM | M_1to2);
+		put("vrcp14pd", YMM_KZ, _YMM | _MEM | M_1to4);
+		put("vrcp14pd", ZMM_KZ, _ZMM | _MEM | M_1to8);
+
+		put("vrcp14ps", XMM_KZ, _XMM | _MEM | M_1to4);
+		put("vrcp14ps", YMM_KZ, _YMM | _MEM | M_1to8);
+		put("vrcp14ps", ZMM_KZ, _ZMM | _MEM | M_1to16);
+
+		put("vrcp14sd", XMM_KZ, _XMM, _XMM | _MEM);
+
+		put("vrcp14ss", XMM_KZ, _XMM, _XMM | _MEM);
+
+		put("vrsqrt14pd", XMM_KZ, _XMM | _MEM | M_1to2);
+		put("vrsqrt14pd", YMM_KZ, _YMM | _MEM | M_1to4);
+		put("vrsqrt14pd", ZMM_KZ, _ZMM | _MEM | M_1to8);
+
+		put("vrsqrt14ps", XMM_KZ, _XMM | _MEM | M_1to4);
+		put("vrsqrt14ps", YMM_KZ, _YMM | _MEM | M_1to8);
+		put("vrsqrt14ps", ZMM_KZ, _ZMM | _MEM | M_1to16);
+
+		put("vrsqrt14sd", XMM_KZ, _XMM, _XMM | _MEM);
+
+		put("vrsqrt14ss", XMM_KZ, _XMM, _XMM | _MEM);
+
+		put("vrndscalepd", XMM_KZ, _XMM | _MEM | M_1to2, IMM8);
+		put("vrndscalepd", YMM_KZ, _YMM | _MEM | M_1to4, IMM8);
+		put("vrndscalepd", ZMM_KZ, _ZMM | _MEM | M_1to8, IMM8);
+
+		put("vrndscaleps", XMM_KZ, _XMM | _MEM | M_1to4, IMM8);
+		put("vrndscaleps", YMM_KZ, _YMM | _MEM | M_1to8, IMM8);
+		put("vrndscaleps", ZMM_KZ, _ZMM | _MEM | M_1to16, IMM8);
+
+		put("vrndscalesd", XMM_KZ, _XMM, _XMM | _MEM, IMM8);
+
+		put("vrndscaless", XMM_KZ, _XMM, _XMM | _MEM, IMM8);
+
+		put("vscalefpd", XMM_KZ, _XMM, _XMM | _MEM | M_1to2);
+		put("vscalefpd", YMM_KZ, _YMM, _YMM | _MEM | M_1to4);
+		put("vscalefpd", ZMM_KZ, _ZMM, _ZMM | _MEM | M_1to8 | ZMM_ER);
+
+		put("vscalefps", XMM_KZ, _XMM, _XMM | _MEM | M_1to4);
+		put("vscalefps", YMM_KZ, _YMM, _YMM | _MEM | M_1to8);
+		put("vscalefps", ZMM_KZ, _ZMM, _ZMM | _MEM | M_1to16 | ZMM_ER);
+	}
 	void putMin()
 	{
 #ifdef XBYAK64
-		putMov();
+		put("vscalefsd", XMM_KZ, _XMM, _XMM | _MEM | XMM_ER);
+		put("vscalefss", XMM_KZ, _XMM, _XMM | _MEM | XMM_ER);
 #endif
 	}
 	void putAVX512()
diff --git a/xbyak/xbyak_avx512.h b/xbyak/xbyak_avx512.h
index 1d74a3a..9db2092 100644
--- a/xbyak/xbyak_avx512.h
+++ b/xbyak/xbyak_avx512.h
@@ -111,6 +111,8 @@ void vexpandpd(const Xmm& x, const Operand& op) { opAVX_X_XM_IMM(x, op, T_66 | T
 void vexpandps(const Xmm& x, const Operand& op) { opAVX_X_XM_IMM(x, op, T_66 | T_0F38 | T_EW0 | T_YMM | T_MUST_EVEX | T_N4, 0x88); }
 void vpexpandd(const Xmm& x, const Operand& op) { opAVX_X_XM_IMM(x, op, T_66 | T_0F38 | T_EW0 | T_YMM | T_MUST_EVEX | T_N4, 0x89); }
 void vpexpandq(const Xmm& x, const Operand& op) { opAVX_X_XM_IMM(x, op, T_66 | T_0F38 | T_EW1 | T_YMM | T_MUST_EVEX | T_N8, 0x89); }
+void vgetexppd(const Xmm& x, const Operand& op) { opAVX_X_XM_IMM(x, op, T_66 | T_0F38 | T_EW1 | T_YMM | T_SAE_Z | T_MUST_EVEX | T_B64, 0x42); }
+void vgetexpps(const Xmm& x, const Operand& op) { opAVX_X_XM_IMM(x, op, T_66 | T_0F38 | T_EW0 | T_YMM | T_SAE_Z | T_MUST_EVEX | T_B32, 0x42); }
 void vpabsq(const Xmm& x, const Operand& op) { opAVX_X_XM_IMM(x, op, T_66 | T_0F38 | T_MUST_EVEX | T_EW1 | T_B64 | T_YMM, 0x1F); }
 void vmovdqa32(const Address& addr, const Xmm& x) { opAVX_X_XM_IMM(x, addr, T_66 | T_0F | T_EW0 | T_YMM | T_ER_X | T_ER_Y | T_ER_Z | T_MUST_EVEX, 0x7F); }
 void vmovdqa64(const Address& addr, const Xmm& x) { opAVX_X_XM_IMM(x, addr, T_66 | T_0F | T_EW1 | T_YMM | T_ER_X | T_ER_Y | T_ER_Z | T_MUST_EVEX, 0x7F); }
@@ -163,6 +165,24 @@ void vpermi2ps(const Xmm& x1, const Xmm& x2, const Operand& op) { opAVX_X_X_XM(x
 void vpermi2pd(const Xmm& x1, const Xmm& x2, const Operand& op) { opAVX_X_X_XM(x1, x2, op, T_66 | T_0F38 | T_EW1 | T_YMM | T_MUST_EVEX | T_B64, 0x77); }
 void vpternlogd(const Xmm& x1, const Xmm& x2, const Operand& op, uint8 imm) { opAVX_X_X_XM(x1, x2, op, T_66 | T_0F3A | T_EW0 | T_YMM | T_MUST_EVEX | T_B32, 0x25, imm); }
 void vpternlogq(const Xmm& x1, const Xmm& x2, const Operand& op, uint8 imm) { opAVX_X_X_XM(x1, x2, op, T_66 | T_0F3A | T_EW1 | T_YMM | T_MUST_EVEX | T_B64, 0x25, imm); }
+void vgetexpsd(const Xmm& x1, const Xmm& x2, const Operand& op) { opAVX_X_X_XM(x1, x2, op, T_66 | T_0F38 | T_EW1 | T_SAE_X | T_MUST_EVEX | T_N8, 0x43); }
+void vgetexpss(const Xmm& x1, const Xmm& x2, const Operand& op) { opAVX_X_X_XM(x1, x2, op, T_66 | T_0F38 | T_EW0 | T_SAE_X | T_MUST_EVEX | T_N4, 0x43); }
+void vgetmantsd(const Xmm& x1, const Xmm& x2, const Operand& op, uint8 imm) { opAVX_X_X_XM(x1, x2, op, T_66 | T_0F3A | T_EW1 | T_SAE_X | T_MUST_EVEX | T_N8, 0x27, imm); }
+void vgetmantss(const Xmm& x1, const Xmm& x2, const Operand& op, uint8 imm) { opAVX_X_X_XM(x1, x2, op, T_66 | T_0F3A | T_EW0 | T_SAE_X | T_MUST_EVEX | T_N4, 0x27, imm); }
+void vfixupimmpd(const Xmm& x1, const Xmm& x2, const Operand& op, uint8 imm) { opAVX_X_X_XM(x1, x2, op, T_66 | T_0F3A | T_EW1 | T_YMM | T_SAE_Z | T_MUST_EVEX | T_B64, 0x54, imm); }
+void vfixupimmps(const Xmm& x1, const Xmm& x2, const Operand& op, uint8 imm) { opAVX_X_X_XM(x1, x2, op, T_66 | T_0F3A | T_EW0 | T_YMM | T_SAE_Z | T_MUST_EVEX | T_B32, 0x54, imm); }
+void vfixupimmsd(const Xmm& x1, const Xmm& x2, const Operand& op, uint8 imm) { opAVX_X_X_XM(x1, x2, op, T_66 | T_0F3A | T_EW1 | T_SAE_Z | T_MUST_EVEX | T_N8, 0x55, imm); }
+void vfixupimmss(const Xmm& x1, const Xmm& x2, const Operand& op, uint8 imm) { opAVX_X_X_XM(x1, x2, op, T_66 | T_0F3A | T_EW0 | T_SAE_Z | T_MUST_EVEX | T_N4, 0x55, imm); }
+void vrcp14sd(const Xmm& x1, const Xmm& x2, const Operand& op) { opAVX_X_X_XM(x1, x2, op, T_66 | T_0F38 | T_EW1 | T_MUST_EVEX | T_N8, 0x4D); }
+void vrcp14ss(const Xmm& x1, const Xmm& x2, const Operand& op) { opAVX_X_X_XM(x1, x2, op, T_66 | T_0F38 | T_EW0 | T_MUST_EVEX | T_N4, 0x4D); }
+void vrsqrt14sd(const Xmm& x1, const Xmm& x2, const Operand& op) { opAVX_X_X_XM(x1, x2, op, T_66 | T_0F38 | T_EW1 | T_YMM | T_MUST_EVEX | T_N8, 0x4F); }
+void vrsqrt14ss(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_N4, 0x4F); }
+void vrndscalesd(const Xmm& x1, const Xmm& x2, const Operand& op, uint8 imm) { opAVX_X_X_XM(x1, x2, op, T_66 | T_0F3A | T_EW1 | T_MUST_EVEX | T_N8, 0x0B, imm); }
+void vrndscaless(const Xmm& x1, const Xmm& x2, const Operand& op, uint8 imm) { opAVX_X_X_XM(x1, x2, op, T_66 | T_0F3A | T_EW0 | T_MUST_EVEX | T_N4, 0x0A, imm); }
+void vscalefpd(const Xmm& x1, const Xmm& x2, const Operand& op) { opAVX_X_X_XM(x1, x2, op, T_66 | T_0F38 | T_EW1 | T_YMM | T_ER_Z | T_MUST_EVEX | T_B64, 0x2C); }
+void vscalefps(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_ER_Z | T_MUST_EVEX | T_B32, 0x2C); }
+void vscalefsd(const Xmm& x1, const Xmm& x2, const Operand& op) { opAVX_X_X_XM(x1, x2, op, T_66 | T_0F38 | T_EW1 | T_ER_X | T_MUST_EVEX | T_N8, 0x2D); }
+void vscalefss(const Xmm& x1, const Xmm& x2, const Operand& op) { opAVX_X_X_XM(x1, x2, op, T_66 | T_0F38 | T_EW0 | T_ER_X | T_MUST_EVEX | T_N4, 0x2D); }
 void vpsraq(const Xmm& x, const Operand& op, uint8 imm) { opAVX_X_X_XM(Xmm(x.getKind(), 4), x, op, T_66 | T_0F | T_EW1 | T_YMM | T_MUST_EVEX | T_B64, 0x72, imm); }
 void vextractf32x4(const Operand& op, const Ymm& r, uint8 imm) { opAVX_X_X_XMcvt(r, true, cvtIdx0(r), op, op.isXMM(), Operand::YMM, T_66 | T_0F3A | T_EW0 | T_YMM | T_MUST_EVEX | T_N16, 0x19, imm); }
 void vextractf64x2(const Operand& op, const Ymm& r, uint8 imm) { opAVX_X_X_XMcvt(r, true, cvtIdx0(r), op, op.isXMM(), Operand::YMM, T_66 | T_0F3A | T_EW1 | T_YMM | T_MUST_EVEX | T_N16, 0x19, imm); }
@@ -254,4 +274,12 @@ void vpmovusdw(const Operand& op, const Xmm& x) { opVmov(op, x, T_F3 | T_0F38 |
 void vpmovwb(const Operand& op, const Xmm& x) { opVmov(op, x, T_F3 | T_0F38 | T_EW0 | T_YMM | T_MUST_EVEX | T_N8 | T_N_VL, 0x30, true); }
 void vpmovswb(const Operand& op, const Xmm& x) { opVmov(op, x, T_F3 | T_0F38 | T_EW0 | T_YMM | T_MUST_EVEX | T_N8 | T_N_VL, 0x20, true); }
 void vpmovuswb(const Operand& op, const Xmm& x) { opVmov(op, x, T_F3 | T_0F38 | T_EW0 | T_YMM | T_MUST_EVEX | T_N8 | T_N_VL, 0x10, true); }
+void vgetmantpd(const Xmm& x, const Operand& op, uint8 imm) { opAVX_X_XM_IMM(x, op, T_66 | T_0F3A | T_EW1 | T_YMM | T_SAE_Z | T_MUST_EVEX | T_B64, 0x26, imm); }
+void vgetmantps(const Xmm& x, const Operand& op, uint8 imm) { opAVX_X_XM_IMM(x, op, T_66 | T_0F3A | T_EW0 | T_YMM | T_SAE_Z | T_MUST_EVEX | T_B32, 0x26, imm); }
+void vrcp14pd(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, 0x4C); }
+void vrcp14ps(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, 0x4C); }
+void vrsqrt14pd(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, 0x4E); }
+void vrsqrt14ps(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, 0x4E); }
+void vrndscalepd(const Xmm& x, const Operand& op, uint8 imm) { opAVX_X_XM_IMM(x, op, T_66 | T_0F3A | T_EW1 | T_YMM | T_MUST_EVEX | T_B64, 0x09, imm); }
+void vrndscaleps(const Xmm& x, const Operand& op, uint8 imm) { opAVX_X_XM_IMM(x, op, T_66 | T_0F3A | T_EW0 | T_YMM | T_MUST_EVEX | T_B32, 0x08, imm); }
 #endif