fix vcmppd/vcmpps for ptr_b
diff --git a/gen/gen_avx512.cpp b/gen/gen_avx512.cpp
index 069ca54..84cd612 100644
--- a/gen/gen_avx512.cpp
+++ b/gen/gen_avx512.cpp
@@ -103,8 +103,8 @@
int type;
bool hasIMM;
} tbl[] = {
- { 0xC2, "vcmppd", T_0F | T_MUST_EVEX | T_EW1 | T_SAE_Z | T_YMM | T_66, true },
- { 0xC2, "vcmpps", T_0F | T_MUST_EVEX | T_EW0 | T_SAE_Z | T_YMM, true },
+ { 0xC2, "vcmppd", T_0F | T_MUST_EVEX | T_EW1 | T_SAE_Z | T_YMM | T_66 | T_B64, true },
+ { 0xC2, "vcmpps", T_0F | T_MUST_EVEX | T_EW0 | T_SAE_Z | T_YMM | T_B32, true },
{ 0xC2, "vcmpsd", T_0F | T_MUST_EVEX | T_EW1 | T_SAE_Z | T_F2 | T_N8, true },
{ 0xC2, "vcmpss", T_0F | T_MUST_EVEX | T_EW0 | T_SAE_Z | T_F3 | T_N4, true },
diff --git a/readme.md b/readme.md
index 421674d..9c5f2a6 100644
--- a/readme.md
+++ b/readme.md
@@ -1,5 +1,5 @@
-# Xbyak 5.78 ; JIT assembler for x86(IA32), x64(AMD64, x86-64) by C++
+# Xbyak 5.79 ; JIT assembler for x86(IA32), x64(AMD64, x86-64) by C++
## Abstract
@@ -392,6 +392,7 @@
http://opensource.org/licenses/BSD-3-Clause
## History
+* 2019/Apr/27 ver 5.79 vcmppd/vcmpps supports ptr_b(thanks to jkopinsky)
* 2019/Apr/15 ver 5.78 rewrite Reg::changeBit() (thanks to MerryMage)
* 2019/Mar/06 ver 5.77 fix number of cores that share LLC cache by densamoilov
* 2019/Jan/17 ver 5.76 add Cpu::getNumCores() by shelleygoel
diff --git a/readme.txt b/readme.txt
index 20c77e9..e75f90f 100644
--- a/readme.txt
+++ b/readme.txt
@@ -1,5 +1,5 @@
- C++用x86(IA-32), x64(AMD64, x86-64) JITアセンブラ Xbyak 5.78
+ C++用x86(IA-32), x64(AMD64, x86-64) JITアセンブラ Xbyak 5.79
-----------------------------------------------------------------------------
◎概要
@@ -373,6 +373,7 @@
-----------------------------------------------------------------------------
◎履歴
+2019/04/27 ver 5.79 vcmppd/vcmppsのptr_b対応忘れ(thanks to jkopinsky)
2019/04/15 ver 5.78 Reg::changeBit()のリファクタリング(thanks to MerryMage)
2019/03/06 ver 5.77 LLCキャッシュを共有数CPU数の修整(by densamoilov)
2019/01/17 ver 5.76 Cpu::getNumCores()追加(by shelleygoel)
diff --git a/test/make_512.cpp b/test/make_512.cpp
index 49d082c..ecdaea4 100644
--- a/test/make_512.cpp
+++ b/test/make_512.cpp
@@ -615,6 +615,13 @@
put(p->name, K, _YMM, _YMM | MEM, IMM8);
put(p->name, K, _ZMM, _ZMM | MEM, IMM8);
}
+ put("vcmppd", K, XMM, M_1to2, IMM8);
+ put("vcmppd", K, YMM, M_1to4, IMM8);
+ put("vcmppd", K, ZMM, M_1to8, IMM8);
+
+ put("vcmpps", K, XMM, M_1to4, IMM8);
+ put("vcmpps", K, YMM, M_1to8, IMM8);
+ put("vcmpps", K, ZMM, M_1to16, IMM8);
}
put("vcmppd", K2, ZMM, ZMM_SAE, IMM);
#ifdef XBYAK64
diff --git a/xbyak/xbyak.h b/xbyak/xbyak.h
index f768927..c28a536 100644
--- a/xbyak/xbyak.h
+++ b/xbyak/xbyak.h
@@ -113,7 +113,7 @@
enum {
DEFAULT_MAX_CODE_SIZE = 4096,
- VERSION = 0x5780 /* 0xABCD = A.BC(D) */
+ VERSION = 0x5790 /* 0xABCD = A.BC(D) */
};
#ifndef MIE_INTEGER_TYPE_DEFINED
diff --git a/xbyak/xbyak_mnemonic.h b/xbyak/xbyak_mnemonic.h
index bb40f8c..2733c61 100644
--- a/xbyak/xbyak_mnemonic.h
+++ b/xbyak/xbyak_mnemonic.h
@@ -1,4 +1,4 @@
-const char *getVersionString() const { return "5.78"; }
+const char *getVersionString() const { return "5.79"; }
void adc(const Operand& op, uint32 imm) { opRM_I(op, imm, 0x10, 2); }
void adc(const Operand& op1, const Operand& op2) { opRM_RM(op1, op2, 0x10); }
void adcx(const Reg32e& reg, const Operand& op) { opGen(reg, op, 0xF6, 0x66, isREG32_REG32orMEM, NONE, 0x38); }
@@ -1676,8 +1676,8 @@
void vbroadcasti32x8(const Zmm& z, const Operand& op) { opAVX_X_XM_IMM(z, op, T_66 | T_0F38 | T_YMM | T_MUST_EVEX | T_EW0 | T_N32, 0x5B); }
void vbroadcasti64x2(const Ymm& y, const Operand& op) { opAVX_X_XM_IMM(y, op, T_66 | T_0F38 | T_YMM | T_MUST_EVEX | T_EW1 | T_N16, 0x5A); }
void vbroadcasti64x4(const Zmm& z, const Operand& op) { opAVX_X_XM_IMM(z, op, T_66 | T_0F38 | T_YMM | T_MUST_EVEX | T_EW1 | T_N32, 0x5B); }
-void vcmppd(const Opmask& k, const Xmm& x, const Operand& op, uint8 imm) { opAVX_K_X_XM(k, x, op, T_66 | T_0F | T_EW1 | T_YMM | T_SAE_Z | T_MUST_EVEX, 0xC2, imm); }
-void vcmpps(const Opmask& k, const Xmm& x, const Operand& op, uint8 imm) { opAVX_K_X_XM(k, x, op, T_0F | T_EW0 | T_YMM | T_SAE_Z | T_MUST_EVEX, 0xC2, imm); }
+void vcmppd(const Opmask& k, const Xmm& x, const Operand& op, uint8 imm) { opAVX_K_X_XM(k, x, op, T_66 | T_0F | T_EW1 | T_YMM | T_SAE_Z | T_MUST_EVEX | T_B64, 0xC2, imm); }
+void vcmpps(const Opmask& k, const Xmm& x, const Operand& op, uint8 imm) { opAVX_K_X_XM(k, x, op, T_0F | T_EW0 | T_YMM | T_SAE_Z | T_MUST_EVEX | T_B32, 0xC2, imm); }
void vcmpsd(const Opmask& k, const Xmm& x, const Operand& op, uint8 imm) { opAVX_K_X_XM(k, x, op, T_N8 | T_F2 | T_0F | T_EW1 | T_SAE_Z | T_MUST_EVEX, 0xC2, imm); }
void vcmpss(const Opmask& k, const Xmm& x, const Operand& op, uint8 imm) { opAVX_K_X_XM(k, x, op, T_N4 | T_F3 | T_0F | T_EW0 | T_SAE_Z | T_MUST_EVEX, 0xC2, imm); }
void vcompressb(const Operand& op, const Xmm& x) { opAVX_X_XM_IMM(x, op, T_N1 | T_66 | T_0F38 | T_EW0 | T_YMM | T_MUST_EVEX, 0x63); }
