add vpermi2b

gregory38 · Jul 23, 2016 · 9af3532 · 9af3532
1 parent fbde291
commit 9af3532
Show file tree

Hide file tree

Showing 5 changed files with 18 additions and 3 deletions.
diff --git a/gen/gen_avx512.cpp b/gen/gen_avx512.cpp
@@ -282,6 +282,7 @@ void putX_X_XM_IMM()
 		{ 0x7F, "vpermt2ps", T_66 | T_0F38 | T_YMM | T_MUST_EVEX | T_EW0 | T_B32, false },
 		{ 0x7F, "vpermt2pd", T_66 | T_0F38 | T_YMM | T_MUST_EVEX | T_EW1 | T_B64, false },
 
+		{ 0x75, "vpermi2b", T_66 | T_0F38 | T_YMM | T_MUST_EVEX | T_EW0, false },
 		{ 0x75, "vpermi2w", T_66 | T_0F38 | T_YMM | T_MUST_EVEX | T_EW1, false },
 		{ 0x76, "vpermi2d", T_66 | T_0F38 | T_YMM | T_MUST_EVEX | T_EW0 | T_B32, false },
 		{ 0x76, "vpermi2q", T_66 | T_0F38 | T_YMM | T_MUST_EVEX | T_EW1 | T_B64, false },

diff --git a/readme.md b/readme.md
@@ -143,18 +143,23 @@ vaddpd xmm1, xmm2, [rax+256]{1to2}      --> vaddpd(xmm1, xmm2, ptr_b [rax+256]);
 vaddpd ymm1, ymm2, [rax+256]{1to4}      --> vaddpd(ymm1, ymm2, ptr_b [rax+256]);
 vaddpd zmm1, zmm2, [rax+256]{1to8}      --> vaddpd(zmm1, zmm2, ptr_b [rax+256]);
 vaddps zmm1, zmm2, [rax+rcx*8+8]{1to16} --> vaddps(zmm1, zmm2, ptr_b [rax+rcx*8+8]);
-vmovsd [rax]{k1}, xmm4                  --> vmovsd (ptr [rax] | k1, xmm4);
+vmovsd [rax]{k1}, xmm4                  --> vmovsd(ptr [rax] | k1, xmm4);
 
 vcvtpd2dq xmm16, oword [eax+33]         --> vcvtpd2dq(xmm16, ptr [eax+33]); // default oword(m128)
 vcvtpd2dq xmm21, [eax+32]{1to2}         --> vcvtpd2dq(xmm21, ptr_b [eax+32]);
 vcvtpd2dq xmm0, yword [eax+33]          --> vcvtpd2dq(xmm0, yword [eax+33]); // use yword for m256
 vcvtpd2dq xmm19, [eax+32]{1to4}         --> vcvtpd2dq(xmm19, yword_b [eax+32]); // use yword_b to broadcast
+
+vfpclassps k5{k3}, zword [rax+64], 5    --> vfpclassps(k5|k3, zword [rax+64], 5); // specify m512
+vfpclasspd k5{k3}, [rax+64]{1to2}, 5    --> vfpclasspd(k5|k3, xword_b [rax+64], 5); // broadcast 64-bit to 128-bit
+vfpclassps k5{k3}, [rax+64]{1to4}, 5    --> vfpclassps(k5|k3, xword_b [rax+64], 5); // broadcast 32-bit to 128-bit
 ```
 Remark
 * k1, ..., k7 are new opmask registers.
 * use `| T_z`, `| T_sae`, `| T_rn_sae`, `| T_rd_sae`, `| T_ru_sae`, `| T_rz_sae` instead of `,{z}`, `,{sae}`, `,{rn-sae}`, `,{rd-sae}`, `,{ru-sae}`, `,{rz-sae}` respectively.
 * `k4 | k3` is different from `k3 | k4`.
-* use `ptr_b`, `yword_b' for broadcast `{1toX}`. X is automatically determined.
+* use `ptr_b` for broadcast `{1toX}`. X is automatically determined.
+* specify xword/yword/zword(_b) for m128/m256/m512.
 
 ### Label
 

diff --git a/readme.txt b/readme.txt
@@ -131,25 +131,31 @@ vaddpd zmm2, zmm5, zmm30                --> vaddpd(zmm2, zmm5, zmm30);
 vaddpd zmm2{k5}, zmm4, zmm2             --> vaddpd(zmm2 | k5, zmm4, zmm2);
 vaddpd zmm2{k5}{z}, zmm4, zmm2          --> vaddpd(zmm2 | k5 | T_z, zmm4, zmm2);
 vaddpd zmm2{k5}{z}, zmm4, zmm2,{rd-sae} --> vaddpd(zmm2 | k5 | T_z, zmm4, zmm2 | T_rd_sae);
+                                            vaddpd(zmm2 | k5 | T_z | T_rd_sae, zmm4, zmm2); // the position of `|` is arbitrary.
 vcmppd k4{k3}, zmm1, zmm2, {sae}, 5     --> vcmppd(k4 | k3, zmm1, zmm2 | T_sae, 5);
 
 vaddpd xmm1, xmm2, [rax+256]{1to2}      --> vaddpd(xmm1, xmm2, ptr_b [rax+256]);
 vaddpd ymm1, ymm2, [rax+256]{1to4}      --> vaddpd(ymm1, ymm2, ptr_b [rax+256]);
 vaddpd zmm1, zmm2, [rax+256]{1to8}      --> vaddpd(zmm1, zmm2, ptr_b [rax+256]);
 vaddps zmm1, zmm2, [rax+rcx*8+8]{1to16} --> vaddps(zmm1, zmm2, ptr_b [rax+rcx*8+8]);
-vmovsd [rax]{k1}, xmm4                  --> vmovsd (ptr [rax] | k1, xmm4);
+vmovsd [rax]{k1}, xmm4                  --> vmovsd(ptr [rax] | k1, xmm4);
 
 vcvtpd2dq xmm16, oword [eax+33]         --> vcvtpd2dq(xmm16, ptr [eax+33]); // default oword(m128)
 vcvtpd2dq xmm21, [eax+32]{1to2}         --> vcvtpd2dq(xmm21, ptr_b [eax+32]);
 vcvtpd2dq xmm0, yword [eax+33]          --> vcvtpd2dq(xmm0, yword [eax+33]); // use yword for m256
 vcvtpd2dq xmm19, [eax+32]{1to4}         --> vcvtpd2dq(xmm19, yword_b [eax+32]); // use yword_b to broadcast
 
+vfpclassps k5{k3}, zword [rax+64], 5    --> vfpclassps(k5|k3, zword [rax+64], 5); // specify m512
+vfpclasspd k5{k3}, [rax+64]{1to2}, 5    --> vfpclasspd(k5|k3, xword_b [rax+64], 5); // broadcast 64-bit to 128-bit
+vfpclassps k5{k3}, [rax+64]{1to4}, 5    --> vfpclassps(k5|k3, xword_b [rax+64], 5); // broadcast 32-bit to 128-bit
+
 
 注意
 * k1, ..., k7 は新しいopmaskレジスタです。
 * z, sae, rn-sae, rd-sae, ru-sae, rz-saeの代わりにT_z, T_sae, T_rn_sae, T_rd_sae, T_ru_sae, T_rz_saeを使ってください。
 * `k4 | k3`と`k3 | k4`は意味が異なります。
 * {1toX}の代わりにptr_bを使ってください。Xは自動的に決まります。
+* 一部の命令はメモリサイズを指定するためにxword/yword/zword(_b)を使ってください。
 
 ・ラベル
 

diff --git a/test/make_512.cpp b/test/make_512.cpp
@@ -1755,6 +1755,7 @@ class Test {
 			{ "vpermt2ps", M_1to4 },
 			{ "vpermt2pd", M_1to2 },
 
+			{ "vpermi2b", 0 },
 			{ "vpermi2w", 0 },
 			{ "vpermi2d", M_1to4 },
 			{ "vpermi2q", M_1to2 },
@@ -2085,6 +2086,7 @@ class Test {
 	void putMin()
 	{
 #ifdef XBYAK64
+		putClass();
 #endif
 	}
 	void putAVX512()

diff --git a/xbyak/xbyak_avx512.h b/xbyak/xbyak_avx512.h
@@ -164,6 +164,7 @@ void vpermt2d(const Xmm& x1, const Xmm& x2, const Operand& op) { opAVX_X_X_XM(x1
 void vpermt2q(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, 0x7E); }
 void vpermt2ps(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, 0x7F); }
 void vpermt2pd(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, 0x7F); }
+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 vpermi2w(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, 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); }
 void vpermi2q(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, 0x76); }