commit f402faadcf6bc44698206ac298040106b1453779
author MITSUNARI Shigeo <herumi@nifty.com> Mon May 27 16:57:59 2019 +0900
committer MITSUNARI Shigeo <herumi@nifty.com> Tue May 28 06:44:08 2019 +0900
tree 98300ef0d02594f7a8bb995e93a62eacf14db847
parent 4cfd520878a1d14acde5441629fe3f705d5470ca

add vp2intersectd/vp2intersectq

gen/gen_avx512.cpp

diff --git a/gen/gen_avx512.cpp b/gen/gen_avx512.cpp
index 3f0508c..15fc62e 100644
--- a/gen/gen_avx512.cpp
+++ b/gen/gen_avx512.cpp
@@ -716,6 +716,8 @@
 	puts("void vpshufbitqmb(const Opmask& k, const Xmm& x, const Operand& op) { opVex(k, &x, op, T_66 | T_0F38 | T_EW0 | T_YMM | T_MUST_EVEX, 0x8F); }");
 	puts("void vcvtneps2bf16(const Xmm& x, const Operand& op) { int xBit = x.getBit(); int opBit = op.getBit(); if (xBit == 256 && opBit == 0) opBit = 512; if (!(xBit == 128 && (opBit == 128 || opBit == 256)) && !(xBit == 256 && opBit == 512)) throw Error(ERR_BAD_COMBINATION); Xmm t = x; t.setBit(opBit); opAVX_X_XM_IMM(t, op, T_F3 | T_0F38 | T_EW0 | T_YMM | T_SAE_Z | T_MUST_EVEX | T_B32, 0x72); }");
 
+	puts("void vp2intersectd(const Opmask& k, const Xmm& x, const Operand& op) { if (k.getOpmaskIdx() != 0) throw Error(ERR_OPMASK_IS_ALREADY_SET); opAVX_K_X_XM(k, x, op, T_F2 | T_0F38 | T_YMM | T_EVEX | T_EW0 | T_B32, 0x68); }");
+	puts("void vp2intersectq(const Opmask& k, const Xmm& x, const Operand& op) { if (k.getOpmaskIdx() != 0) throw Error(ERR_OPMASK_IS_ALREADY_SET); opAVX_K_X_XM(k, x, op, T_F2 | T_0F38 | T_YMM | T_EVEX | T_EW1 | T_B64, 0x68); }");
 }
 
 void putV4FMA()

readme.md

diff --git a/readme.md b/readme.md
index 7449d22..a39d30e 100644
--- a/readme.md
+++ b/readme.md
@@ -392,6 +392,7 @@
 http://opensource.org/licenses/BSD-3-Clause
 
 ## History
+* 2019/May/27 support vp2intersectd, vp2intersectq (not tested)
 * 2019/May/26 ver 5.80 support vcvtne2ps2bf16, vcvtneps2bf16, vdpbf16ps
 * 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)

readme.txt

diff --git a/readme.txt b/readme.txt
index 3eaffdb..2c77042 100644
--- a/readme.txt
+++ b/readme.txt
@@ -373,6 +373,7 @@
 -----------------------------------------------------------------------------

 ◎履歴

 

+2019/05/27 support vp2intersectd, vp2intersectq (not tested)

 2019/05/26 ver 5.80 support vcvtne2ps2bf16, vcvtneps2bf16, vdpbf16ps

 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)

xbyak/xbyak_mnemonic.h

diff --git a/xbyak/xbyak_mnemonic.h b/xbyak/xbyak_mnemonic.h
index 732b097..f947286 100644
--- a/xbyak/xbyak_mnemonic.h
+++ b/xbyak/xbyak_mnemonic.h
@@ -1772,6 +1772,8 @@
 void vmovdqu64(const Xmm& x, const Operand& op) { opAVX_X_XM_IMM(x, op, T_F3 | T_0F | T_EW1 | T_YMM | T_ER_X | T_ER_Y | T_ER_Z | T_MUST_EVEX, 0x6F); }
 void vmovdqu8(const Address& addr, const Xmm& x) { opAVX_X_XM_IMM(x, addr, T_F2 | T_0F | T_EW0 | T_YMM | T_ER_X | T_ER_Y | T_ER_Z | T_MUST_EVEX | T_M_K, 0x7F); }
 void vmovdqu8(const Xmm& x, const Operand& op) { opAVX_X_XM_IMM(x, op, T_F2 | T_0F | T_EW0 | T_YMM | T_ER_X | T_ER_Y | T_ER_Z | T_MUST_EVEX, 0x6F); }
+void vp2intersectd(const Opmask& k, const Xmm& x, const Operand& op) { if (k.getOpmaskIdx() != 0) throw Error(ERR_OPMASK_IS_ALREADY_SET); opAVX_K_X_XM(k, x, op, T_F2 | T_0F38 | T_YMM | T_EVEX | T_EW0 | T_B32, 0x68); }
+void vp2intersectq(const Opmask& k, const Xmm& x, const Operand& op) { if (k.getOpmaskIdx() != 0) throw Error(ERR_OPMASK_IS_ALREADY_SET); opAVX_K_X_XM(k, x, op, T_F2 | T_0F38 | T_YMM | T_EVEX | T_EW1 | T_B64, 0x68); }
 void vp4dpwssd(const Zmm& z1, const Zmm& z2, const Address& addr) { opAVX_X_X_XM(z1, z2, addr, T_0F38 | T_F2 | T_EW0 | T_YMM | T_MUST_EVEX | T_N16, 0x52); }
 void vp4dpwssds(const Zmm& z1, const Zmm& z2, const Address& addr) { opAVX_X_X_XM(z1, z2, addr, T_0F38 | T_F2 | T_EW0 | T_YMM | T_MUST_EVEX | T_N16, 0x53); }
 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); }