True16 for v_pack_b32_f16 in MC

Author: broxigarchen

llvm/lib/Target/AMDGPU/VOP3Instructions.td

@@ -642,7 +642,7 @@ defm V_MAD_I32_I16 : VOP3Inst <"v_mad_i32_i16", VOP3_Profile<VOP_I32_I16_I16_I32
 defm V_CVT_PKNORM_I16_F16 : VOP3Inst <"v_cvt_pknorm_i16_f16", VOP3_Profile<VOP_B32_F16_F16, VOP3_OPSEL>>;
 defm V_CVT_PKNORM_U16_F16 : VOP3Inst <"v_cvt_pknorm_u16_f16", VOP3_Profile<VOP_B32_F16_F16, VOP3_OPSEL>>;
 
-defm V_PACK_B32_F16 : VOP3Inst <"v_pack_b32_f16", VOP3_Profile<VOP_B32_F16_F16, VOP3_OPSEL>>;
+defm V_PACK_B32_F16 : VOP3Inst_t16 <"v_pack_b32_f16", VOP_B32_F16_F16>;
 
 let isReMaterializable = 1 in {
 defm V_SUB_I32 : VOP3Inst <"v_sub_i32", VOP3_Profile<VOP_I32_I32_I32_ARITH>>;
@@ -1751,7 +1751,7 @@ defm V_MIN_U16             : VOP3Only_Realtriple_t16_and_fake16_gfx11_gfx12<0x30
 defm V_MIN_I16             : VOP3Only_Realtriple_t16_and_fake16_gfx11_gfx12<0x30c, "v_min_i16">;
 defm V_ADD_NC_I16          : VOP3_Realtriple_t16_and_fake16_gfx11_gfx12<0x30d, "v_add_nc_i16", "V_ADD_I16">;
 defm V_SUB_NC_I16          : VOP3_Realtriple_t16_and_fake16_gfx11_gfx12<0x30e, "v_sub_nc_i16", "V_SUB_I16">;
-defm V_PACK_B32_F16        : VOP3_Realtriple_gfx11_gfx12<0x311>;
+defm V_PACK_B32_F16        : VOP3_Realtriple_t16_and_fake16_gfx11_gfx12<0x311, "v_pack_b32_f16">;
 defm V_CVT_PK_NORM_I16_F16 : VOP3_Realtriple_with_name_gfx11_gfx12<0x312, "V_CVT_PKNORM_I16_F16" , "v_cvt_pk_norm_i16_f16" >;
 defm V_CVT_PK_NORM_U16_F16 : VOP3_Realtriple_with_name_gfx11_gfx12<0x313, "V_CVT_PKNORM_U16_F16" , "v_cvt_pk_norm_u16_f16" >;
 defm V_SUB_NC_I32          : VOP3_Realtriple_with_name_gfx11_gfx12<0x325, "V_SUB_I32", "v_sub_nc_i32">;

llvm/test/MC/AMDGPU/gfx11_asm_vop3.s

@@ -5207,11 +5207,11 @@ v_or_b16 v5, src_scc, vcc_lo
 v_or_b16 v255, 0xfe0b, vcc_hi
 // GFX11: v_or_b16 v255, 0xfe0b, vcc_hi           ; encoding: [0xff,0x00,0x63,0xd7,0xff,0xd6,0x00,0x00,0x0b,0xfe,0x00,0x00]
 
-v_pack_b32_f16 v5, v1, v2
-// GFX11: v_pack_b32_f16 v5, v1, v2               ; encoding: [0x05,0x00,0x11,0xd7,0x01,0x05,0x02,0x00]
+v_pack_b32_f16 v5, v1.l, v2.l
+// GFX11: v_pack_b32_f16 v5, v1.l, v2.l           ; encoding: [0x05,0x00,0x11,0xd7,0x01,0x05,0x02,0x00]
 
-v_pack_b32_f16 v5, v255, v255
-// GFX11: v_pack_b32_f16 v5, v255, v255           ; encoding: [0x05,0x00,0x11,0xd7,0xff,0xff,0x03,0x00]
+v_pack_b32_f16 v5, v255.l, v255.l
+// GFX11: v_pack_b32_f16 v5, v255.l, v255.l       ; encoding: [0x05,0x00,0x11,0xd7,0xff,0xff,0x03,0x00]
 
 v_pack_b32_f16 v5, s1, s2
 // GFX11: v_pack_b32_f16 v5, s1, s2               ; encoding: [0x05,0x00,0x11,0xd7,0x01,0x04,0x00,0x00]
@@ -5243,7 +5243,7 @@ v_pack_b32_f16 v5, null, exec_lo
 v_pack_b32_f16 v5, -1, exec_hi
 // GFX11: v_pack_b32_f16 v5, -1, exec_hi          ; encoding: [0x05,0x00,0x11,0xd7,0xc1,0xfe,0x00,0x00]
 
-v_pack_b32_f16 v5, 0.5, -m0 op_sel:[0,0,0]
+v_pack_b32_f16 v5, 0.5, -m0
 // GFX11: v_pack_b32_f16 v5, 0.5, -m0             ; encoding: [0x05,0x00,0x11,0xd7,0xf0,0xfa,0x00,0x40]
 
 v_pack_b32_f16 v5, -src_scc, |vcc_lo| op_sel:[1,0,0]
@@ -5252,6 +5252,18 @@ v_pack_b32_f16 v5, -src_scc, |vcc_lo| op_sel:[1,0,0]
 v_pack_b32_f16 v255, -|0xfe0b|, -|vcc_hi| op_sel:[0,1,0]
 // GFX11: v_pack_b32_f16 v255, -|0xfe0b|, -|vcc_hi| op_sel:[0,1,0] ; encoding: [0xff,0x13,0x11,0xd7,0xff,0xd6,0x00,0x60,0x0b,0xfe,0x00,0x00]
 
+v_pack_b32_f16 v5, v1.h, v2.l
+// GFX11: v_pack_b32_f16 v5, v1.h, v2.l op_sel:[1,0,0] ; encoding: [0x05,0x08,0x11,0xd7,0x01,0x05,0x02,0x00]
+
+v_pack_b32_f16 v5, v255.l, v255.h
+// GFX11: v_pack_b32_f16 v5, v255.l, v255.h op_sel:[0,1,0] ; encoding: [0x05,0x10,0x11,0xd7,0xff,0xff,0x03,0x00]
+
+v_pack_b32_f16 v5, -src_scc, |vcc_lo|
+// GFX11: v_pack_b32_f16 v5, -src_scc, |vcc_lo|   ; encoding: [0x05,0x02,0x11,0xd7,0xfd,0xd4,0x00,0x20]
+
+v_pack_b32_f16 v255, -|0xfe0b|, -|vcc_hi|
+// GFX11: v_pack_b32_f16 v255, -|0xfe0b|, -|vcc_hi| ; encoding: [0xff,0x03,0x11,0xd7,0xff,0xd6,0x00,0x60,0x0b,0xfe,0x00,0x00]
+
 v_perm_b32 v5, v1, v2, s3
 // GFX11: v_perm_b32 v5, v1, v2, s3               ; encoding: [0x05,0x00,0x44,0xd6,0x01,0x05,0x0e,0x00]
 

llvm/test/MC/AMDGPU/gfx11_asm_vop3_dpp16.s

@@ -3674,47 +3674,53 @@ v_or_b16_e64_dpp v5, v1, v2 row_xmask:0 row_mask:0x1 bank_mask:0x3 bound_ctrl:1
 v_or_b16_e64_dpp v255, v255, v255 row_xmask:15 row_mask:0x3 bank_mask:0x0 bound_ctrl:0 fi:1
 // GFX11: v_or_b16_e64_dpp v255, v255, v255 row_xmask:15 row_mask:0x3 bank_mask:0x0 fi:1 ; encoding: [0xff,0x00,0x63,0xd7,0xfa,0xfe,0x03,0x00,0xff,0x6f,0x05,0x30]
 
-v_pack_b32_f16_e64_dpp v5, v1, v2 quad_perm:[3,2,1,0]
-// GFX11: v_pack_b32_f16_e64_dpp v5, v1, v2 quad_perm:[3,2,1,0] row_mask:0xf bank_mask:0xf ; encoding: [0x05,0x00,0x11,0xd7,0xfa,0x04,0x02,0x00,0x01,0x1b,0x00,0xff]
+v_pack_b32_f16_e64_dpp v5, v1.l, v2.l quad_perm:[3,2,1,0]
+// GFX11: v_pack_b32_f16_e64_dpp v5, v1.l, v2.l quad_perm:[3,2,1,0] row_mask:0xf bank_mask:0xf ; encoding: [0x05,0x00,0x11,0xd7,0xfa,0x04,0x02,0x00,0x01,0x1b,0x00,0xff]
 
-v_pack_b32_f16_e64_dpp v5, v1, v2 quad_perm:[0,1,2,3]
-// GFX11: v_pack_b32_f16_e64_dpp v5, v1, v2 quad_perm:[0,1,2,3] row_mask:0xf bank_mask:0xf ; encoding: [0x05,0x00,0x11,0xd7,0xfa,0x04,0x02,0x00,0x01,0xe4,0x00,0xff]
+v_pack_b32_f16_e64_dpp v5, v1.l, v2.l quad_perm:[0,1,2,3]
+// GFX11: v_pack_b32_f16_e64_dpp v5, v1.l, v2.l quad_perm:[0,1,2,3] row_mask:0xf bank_mask:0xf ; encoding: [0x05,0x00,0x11,0xd7,0xfa,0x04,0x02,0x00,0x01,0xe4,0x00,0xff]
 
-v_pack_b32_f16_e64_dpp v5, v1, v2 row_mirror
-// GFX11: v_pack_b32_f16_e64_dpp v5, v1, v2 row_mirror row_mask:0xf bank_mask:0xf ; encoding: [0x05,0x00,0x11,0xd7,0xfa,0x04,0x02,0x00,0x01,0x40,0x01,0xff]
+v_pack_b32_f16_e64_dpp v5, v1.l, v2.l row_mirror
+// GFX11: v_pack_b32_f16_e64_dpp v5, v1.l, v2.l row_mirror row_mask:0xf bank_mask:0xf ; encoding: [0x05,0x00,0x11,0xd7,0xfa,0x04,0x02,0x00,0x01,0x40,0x01,0xff]
 
-v_pack_b32_f16_e64_dpp v5, v1, v2 row_half_mirror
-// GFX11: v_pack_b32_f16_e64_dpp v5, v1, v2 row_half_mirror row_mask:0xf bank_mask:0xf ; encoding: [0x05,0x00,0x11,0xd7,0xfa,0x04,0x02,0x00,0x01,0x41,0x01,0xff]
+v_pack_b32_f16_e64_dpp v5, v1.l, v2.l row_half_mirror
+// GFX11: v_pack_b32_f16_e64_dpp v5, v1.l, v2.l row_half_mirror row_mask:0xf bank_mask:0xf ; encoding: [0x05,0x00,0x11,0xd7,0xfa,0x04,0x02,0x00,0x01,0x41,0x01,0xff]
 
-v_pack_b32_f16_e64_dpp v5, v1, v2 row_shl:1
-// GFX11: v_pack_b32_f16_e64_dpp v5, v1, v2 row_shl:1 row_mask:0xf bank_mask:0xf ; encoding: [0x05,0x00,0x11,0xd7,0xfa,0x04,0x02,0x00,0x01,0x01,0x01,0xff]
+v_pack_b32_f16_e64_dpp v5, v1.l, v2.l row_shl:1
+// GFX11: v_pack_b32_f16_e64_dpp v5, v1.l, v2.l row_shl:1 row_mask:0xf bank_mask:0xf ; encoding: [0x05,0x00,0x11,0xd7,0xfa,0x04,0x02,0x00,0x01,0x01,0x01,0xff]
 
-v_pack_b32_f16_e64_dpp v5, v1, v2 row_shl:15
-// GFX11: v_pack_b32_f16_e64_dpp v5, v1, v2 row_shl:15 row_mask:0xf bank_mask:0xf ; encoding: [0x05,0x00,0x11,0xd7,0xfa,0x04,0x02,0x00,0x01,0x0f,0x01,0xff]
+v_pack_b32_f16_e64_dpp v5, v1.l, v2.l row_shl:15
+// GFX11: v_pack_b32_f16_e64_dpp v5, v1.l, v2.l row_shl:15 row_mask:0xf bank_mask:0xf ; encoding: [0x05,0x00,0x11,0xd7,0xfa,0x04,0x02,0x00,0x01,0x0f,0x01,0xff]
 
-v_pack_b32_f16_e64_dpp v5, v1, v2 row_shr:1
-// GFX11: v_pack_b32_f16_e64_dpp v5, v1, v2 row_shr:1 row_mask:0xf bank_mask:0xf ; encoding: [0x05,0x00,0x11,0xd7,0xfa,0x04,0x02,0x00,0x01,0x11,0x01,0xff]
+v_pack_b32_f16_e64_dpp v5, v1.l, v2.l row_shr:1
+// GFX11: v_pack_b32_f16_e64_dpp v5, v1.l, v2.l row_shr:1 row_mask:0xf bank_mask:0xf ; encoding: [0x05,0x00,0x11,0xd7,0xfa,0x04,0x02,0x00,0x01,0x11,0x01,0xff]
 
-v_pack_b32_f16_e64_dpp v5, v1, v2 row_shr:15
-// GFX11: v_pack_b32_f16_e64_dpp v5, v1, v2 row_shr:15 row_mask:0xf bank_mask:0xf ; encoding: [0x05,0x00,0x11,0xd7,0xfa,0x04,0x02,0x00,0x01,0x1f,0x01,0xff]
+v_pack_b32_f16_e64_dpp v5, v1.l, v2.l row_shr:15
+// GFX11: v_pack_b32_f16_e64_dpp v5, v1.l, v2.l row_shr:15 row_mask:0xf bank_mask:0xf ; encoding: [0x05,0x00,0x11,0xd7,0xfa,0x04,0x02,0x00,0x01,0x1f,0x01,0xff]
 
-v_pack_b32_f16_e64_dpp v5, v1, v2 row_ror:1
-// GFX11: v_pack_b32_f16_e64_dpp v5, v1, v2 row_ror:1 row_mask:0xf bank_mask:0xf ; encoding: [0x05,0x00,0x11,0xd7,0xfa,0x04,0x02,0x00,0x01,0x21,0x01,0xff]
+v_pack_b32_f16_e64_dpp v5, v1.l, v2.l row_ror:1
+// GFX11: v_pack_b32_f16_e64_dpp v5, v1.l, v2.l row_ror:1 row_mask:0xf bank_mask:0xf ; encoding: [0x05,0x00,0x11,0xd7,0xfa,0x04,0x02,0x00,0x01,0x21,0x01,0xff]
 
-v_pack_b32_f16_e64_dpp v5, v1, v2 row_ror:15
-// GFX11: v_pack_b32_f16_e64_dpp v5, v1, v2 row_ror:15 row_mask:0xf bank_mask:0xf ; encoding: [0x05,0x00,0x11,0xd7,0xfa,0x04,0x02,0x00,0x01,0x2f,0x01,0xff]
+v_pack_b32_f16_e64_dpp v5, v1.l, v2.l row_ror:15
+// GFX11: v_pack_b32_f16_e64_dpp v5, v1.l, v2.l row_ror:15 row_mask:0xf bank_mask:0xf ; encoding: [0x05,0x00,0x11,0xd7,0xfa,0x04,0x02,0x00,0x01,0x2f,0x01,0xff]
 
-v_pack_b32_f16_e64_dpp v5, v1, v2 row_share:0 row_mask:0xf bank_mask:0xf
-// GFX11: v_pack_b32_f16_e64_dpp v5, v1, v2 row_share:0 row_mask:0xf bank_mask:0xf ; encoding: [0x05,0x00,0x11,0xd7,0xfa,0x04,0x02,0x00,0x01,0x50,0x01,0xff]
+v_pack_b32_f16_e64_dpp v5, v1.l, v2.l row_share:0 row_mask:0xf bank_mask:0xf
+// GFX11: v_pack_b32_f16_e64_dpp v5, v1.l, v2.l row_share:0 row_mask:0xf bank_mask:0xf ; encoding: [0x05,0x00,0x11,0xd7,0xfa,0x04,0x02,0x00,0x01,0x50,0x01,0xff]
 
-v_pack_b32_f16_e64_dpp v5, |v1|, -v2 row_share:15 row_mask:0x0 bank_mask:0x1
-// GFX11: v_pack_b32_f16_e64_dpp v5, |v1|, -v2 row_share:15 row_mask:0x0 bank_mask:0x1 ; encoding: [0x05,0x01,0x11,0xd7,0xfa,0x04,0x02,0x40,0x01,0x5f,0x01,0x01]
+v_pack_b32_f16_e64_dpp v5, |v1.l|, -v2.l row_share:15 row_mask:0x0 bank_mask:0x1
+// GFX11: v_pack_b32_f16_e64_dpp v5, |v1.l|, -v2.l row_share:15 row_mask:0x0 bank_mask:0x1 ; encoding: [0x05,0x01,0x11,0xd7,0xfa,0x04,0x02,0x40,0x01,0x5f,0x01,0x01]
 
-v_pack_b32_f16_e64_dpp v5, -v1, |v2| row_xmask:0 row_mask:0x1 bank_mask:0x3 bound_ctrl:1 fi:0
-// GFX11: v_pack_b32_f16_e64_dpp v5, -v1, |v2| row_xmask:0 row_mask:0x1 bank_mask:0x3 bound_ctrl:1 ; encoding: [0x05,0x02,0x11,0xd7,0xfa,0x04,0x02,0x20,0x01,0x60,0x09,0x13]
+v_pack_b32_f16_e64_dpp v5, -v1.l, |v2.l| row_xmask:0 row_mask:0x1 bank_mask:0x3 bound_ctrl:1
+// GFX11: v_pack_b32_f16_e64_dpp v5, -v1.l, |v2.l| row_xmask:0 row_mask:0x1 bank_mask:0x3 bound_ctrl:1 ; encoding: [0x05,0x02,0x11,0xd7,0xfa,0x04,0x02,0x20,0x01,0x60,0x09,0x13]
 
-v_pack_b32_f16_e64_dpp v255, -|v255|, -|v255| row_xmask:15 row_mask:0x3 bank_mask:0x0 bound_ctrl:0 fi:1
-// GFX11: v_pack_b32_f16_e64_dpp v255, -|v255|, -|v255| row_xmask:15 row_mask:0x3 bank_mask:0x0 fi:1 ; encoding: [0xff,0x03,0x11,0xd7,0xfa,0xfe,0x03,0x60,0xff,0x6f,0x05,0x30]
+v_pack_b32_f16_e64_dpp v255, -|v255.l|, -|v255.l| row_xmask:15 row_mask:0x3 bank_mask:0x0 fi:1
+// GFX11: v_pack_b32_f16_e64_dpp v255, -|v255.l|, -|v255.l| row_xmask:15 row_mask:0x3 bank_mask:0x0 fi:1 ; encoding: [0xff,0x03,0x11,0xd7,0xfa,0xfe,0x03,0x60,0xff,0x6f,0x05,0x30]
+
+v_pack_b32_f16_e64_dpp v5, -v1.h, |v2.l| row_xmask:0 row_mask:0x1 bank_mask:0x3 bound_ctrl:1 fi:0
+// GFX11: v_pack_b32_f16_e64_dpp v5, -v1.h, |v2.l| op_sel:[1,0,0] row_xmask:0 row_mask:0x1 bank_mask:0x3 bound_ctrl:1 ; encoding: [0x05,0x0a,0x11,0xd7,0xfa,0x04,0x02,0x20,0x01,0x60,0x09,0x13]
+
+v_pack_b32_f16_e64_dpp v255, -|v255.l|, -|v255.h| row_xmask:15 row_mask:0x3 bank_mask:0x0 bound_ctrl:0 fi:1
+// GFX11: v_pack_b32_f16_e64_dpp v255, -|v255.l|, -|v255.h| op_sel:[0,1,0] row_xmask:15 row_mask:0x3 bank_mask:0x0 fi:1 ; encoding: [0xff,0x13,0x11,0xd7,0xfa,0xfe,0x03,0x60,0xff,0x6f,0x05,0x30]
 
 v_perm_b32_e64_dpp v5, v1, v2, v3 quad_perm:[3,2,1,0]
 // GFX11: v_perm_b32_e64_dpp v5, v1, v2, v3 quad_perm:[3,2,1,0] row_mask:0xf bank_mask:0xf ; encoding: [0x05,0x00,0x44,0xd6,0xfa,0x04,0x0e,0x04,0x01,0x1b,0x00,0xff]
@@ -4611,11 +4617,11 @@ v_min3_u16_e64_dpp v5, v1, v2, -1 op_sel:[0,0,1,0] row_xmask:0 row_mask:0x1 bank
 v_min3_u16_e64_dpp v255, v255, v255, src_scc op_sel:[0,0,0,1] row_xmask:15 row_mask:0x3 bank_mask:0x0 bound_ctrl:1 fi:1
 // GFX11: v_min3_u16_e64_dpp v255, v255, v255, src_scc op_sel:[0,0,0,1] row_xmask:15 row_mask:0x3 bank_mask:0x0 bound_ctrl:1 fi:1 ; encoding: [0xff,0x40,0x4b,0xd6,0xfa,0xfe,0xf7,0x03,0xff,0x6f,0x0d,0x30]
 
-v_pack_b32_f16_e64_dpp v5, -v1, |v2| op_sel:[1,0,0] row_xmask:0 row_mask:0x1 bank_mask:0x3
-// GFX11: v_pack_b32_f16_e64_dpp v5, -v1, |v2| op_sel:[1,0,0] row_xmask:0 row_mask:0x1 bank_mask:0x3 ; encoding: [0x05,0x0a,0x11,0xd7,0xfa,0x04,0x02,0x20,0x01,0x60,0x01,0x13]
+v_pack_b32_f16_e64_dpp v5, -v1.h, |v2.l| op_sel:[1,0,0] row_xmask:0 row_mask:0x1 bank_mask:0x3
+// GFX11: v_pack_b32_f16_e64_dpp v5, -v1.h, |v2.l| op_sel:[1,0,0] row_xmask:0 row_mask:0x1 bank_mask:0x3 ; encoding: [0x05,0x0a,0x11,0xd7,0xfa,0x04,0x02,0x20,0x01,0x60,0x01,0x13]
 
-v_pack_b32_f16_e64_dpp v255, -|v255|, -|v255| op_sel:[0,1,0] row_xmask:15 row_mask:0x3 bank_mask:0x0 bound_ctrl:1 fi:1
-// GFX11: v_pack_b32_f16_e64_dpp v255, -|v255|, -|v255| op_sel:[0,1,0] row_xmask:15 row_mask:0x3 bank_mask:0x0 bound_ctrl:1 fi:1 ; encoding: [0xff,0x13,0x11,0xd7,0xfa,0xfe,0x03,0x60,0xff,0x6f,0x0d,0x30]
+v_pack_b32_f16_e64_dpp v255, -|v255.l|, -|v255.h| op_sel:[0,1,0] row_xmask:15 row_mask:0x3 bank_mask:0x0 bound_ctrl:1 fi:1
+// GFX11: v_pack_b32_f16_e64_dpp v255, -|v255.l|, -|v255.h| op_sel:[0,1,0] row_xmask:15 row_mask:0x3 bank_mask:0x0 bound_ctrl:1 fi:1 ; encoding: [0xff,0x13,0x11,0xd7,0xfa,0xfe,0x03,0x60,0xff,0x6f,0x0d,0x30]
 
 v_dot2_f16_f16_e64_dpp v0.l, v1, v2, v3.l quad_perm:[0,1,2,3] row_mask:0x0 bank_mask:0x0 fi:1
 // GFX11: v_dot2_f16_f16_e64_dpp v0.l, v1, v2, v3.l quad_perm:[0,1,2,3] row_mask:0x0 bank_mask:0x0 fi:1 ; encoding: [0x00,0x00,0x66,0xd6,0xfa,0x04,0x0e,0x04,0x01,0xe4,0x04,0x00]

llvm/test/MC/AMDGPU/gfx11_asm_vop3_dpp8.s

@@ -2359,17 +2359,23 @@ v_or_b16_e64_dpp v5, v1, v2 dpp8:[7,6,5,4,3,2,1,0] fi:1
 v_or_b16_e64_dpp v255, v255, v255 dpp8:[0,0,0,0,0,0,0,0] fi:0
 // GFX11: v_or_b16_e64_dpp v255, v255, v255 dpp8:[0,0,0,0,0,0,0,0] ; encoding: [0xff,0x00,0x63,0xd7,0xe9,0xfe,0x03,0x00,0xff,0x00,0x00,0x00]
 
-v_pack_b32_f16_e64_dpp v5, v1, v2 dpp8:[7,6,5,4,3,2,1,0]
-// GFX11: v_pack_b32_f16_e64_dpp v5, v1, v2 dpp8:[7,6,5,4,3,2,1,0] ; encoding: [0x05,0x00,0x11,0xd7,0xe9,0x04,0x02,0x00,0x01,0x77,0x39,0x05]
+v_pack_b32_f16_e64_dpp v5, v1.l, v2.l dpp8:[7,6,5,4,3,2,1,0]
+// GFX11: v_pack_b32_f16_e64_dpp v5, v1.l, v2.l dpp8:[7,6,5,4,3,2,1,0] ; encoding: [0x05,0x00,0x11,0xd7,0xe9,0x04,0x02,0x00,0x01,0x77,0x39,0x05]
 
-v_pack_b32_f16_e64_dpp v5, |v1|, -v2 dpp8:[7,6,5,4,3,2,1,0]
-// GFX11: v_pack_b32_f16_e64_dpp v5, |v1|, -v2 dpp8:[7,6,5,4,3,2,1,0] ; encoding: [0x05,0x01,0x11,0xd7,0xe9,0x04,0x02,0x40,0x01,0x77,0x39,0x05]
+v_pack_b32_f16_e64_dpp v5, |v1.l|, -v2.l dpp8:[7,6,5,4,3,2,1,0]
+// GFX11: v_pack_b32_f16_e64_dpp v5, |v1.l|, -v2.l dpp8:[7,6,5,4,3,2,1,0] ; encoding: [0x05,0x01,0x11,0xd7,0xe9,0x04,0x02,0x40,0x01,0x77,0x39,0x05]
 
-v_pack_b32_f16_e64_dpp v5, -v1, |v2| dpp8:[7,6,5,4,3,2,1,0] fi:1
-// GFX11: v_pack_b32_f16_e64_dpp v5, -v1, |v2| dpp8:[7,6,5,4,3,2,1,0] fi:1 ; encoding: [0x05,0x02,0x11,0xd7,0xea,0x04,0x02,0x20,0x01,0x77,0x39,0x05]
+v_pack_b32_f16_e64_dpp v5, -v1.l, |v2.l| dpp8:[7,6,5,4,3,2,1,0] fi:1
+// GFX11: v_pack_b32_f16_e64_dpp v5, -v1.l, |v2.l| dpp8:[7,6,5,4,3,2,1,0] fi:1 ; encoding: [0x05,0x02,0x11,0xd7,0xea,0x04,0x02,0x20,0x01,0x77,0x39,0x05]
 
-v_pack_b32_f16_e64_dpp v255, -|v255|, -|v255| dpp8:[0,0,0,0,0,0,0,0] fi:0
-// GFX11: v_pack_b32_f16_e64_dpp v255, -|v255|, -|v255| dpp8:[0,0,0,0,0,0,0,0] ; encoding: [0xff,0x03,0x11,0xd7,0xe9,0xfe,0x03,0x60,0xff,0x00,0x00,0x00]
+v_pack_b32_f16_e64_dpp v255, -|v255.l|, -|v255.l| dpp8:[0,0,0,0,0,0,0,0]
+// GFX11: v_pack_b32_f16_e64_dpp v255, -|v255.l|, -|v255.l| dpp8:[0,0,0,0,0,0,0,0] ; encoding: [0xff,0x03,0x11,0xd7,0xe9,0xfe,0x03,0x60,0xff,0x00,0x00,0x00]
+
+v_pack_b32_f16_e64_dpp v5, -v1.h, |v2.l| dpp8:[7,6,5,4,3,2,1,0] fi:1
+// GFX11: v_pack_b32_f16_e64_dpp v5, -v1.h, |v2.l| op_sel:[1,0,0] dpp8:[7,6,5,4,3,2,1,0] fi:1 ; encoding: [0x05,0x0a,0x11,0xd7,0xea,0x04,0x02,0x20,0x01,0x77,0x39,0x05]
+
+v_pack_b32_f16_e64_dpp v255, -|v255.l|, -|v255.h| dpp8:[0,0,0,0,0,0,0,0] fi:0
+// GFX11: v_pack_b32_f16_e64_dpp v255, -|v255.l|, -|v255.h| op_sel:[0,1,0] dpp8:[0,0,0,0,0,0,0,0] ; encoding: [0xff,0x13,0x11,0xd7,0xe9,0xfe,0x03,0x60,0xff,0x00,0x00,0x00]
 
 v_perm_b32_e64_dpp v5, v1, v2, v3 dpp8:[7,6,5,4,3,2,1,0]
 // GFX11: v_perm_b32_e64_dpp v5, v1, v2, v3 dpp8:[7,6,5,4,3,2,1,0] ; encoding: [0x05,0x00,0x44,0xd6,0xe9,0x04,0x0e,0x04,0x01,0x77,0x39,0x05]
@@ -2968,11 +2974,11 @@ v_min3_u16_e64_dpp v5, v1, v2, -1 op_sel:[0,0,1,0] dpp8:[7,6,5,4,3,2,1,0]
 v_min3_u16_e64_dpp v255, v255, v255, src_scc op_sel:[0,0,0,1] dpp8:[0,0,0,0,0,0,0,0] fi:1
 // GFX11: v_min3_u16_e64_dpp v255, v255, v255, src_scc op_sel:[0,0,0,1] dpp8:[0,0,0,0,0,0,0,0] fi:1 ; encoding: [0xff,0x40,0x4b,0xd6,0xea,0xfe,0xf7,0x03,0xff,0x00,0x00,0x00]
 
-v_pack_b32_f16_e64_dpp v5, -v1, |v2| op_sel:[1,0,0] dpp8:[7,6,5,4,3,2,1,0]
-// GFX11: v_pack_b32_f16_e64_dpp v5, -v1, |v2| op_sel:[1,0,0] dpp8:[7,6,5,4,3,2,1,0] ; encoding: [0x05,0x0a,0x11,0xd7,0xe9,0x04,0x02,0x20,0x01,0x77,0x39,0x05]
+v_pack_b32_f16_e64_dpp v5, -v1.h, |v2.l| op_sel:[1,0,0] dpp8:[7,6,5,4,3,2,1,0]
+// GFX11: v_pack_b32_f16_e64_dpp v5, -v1.h, |v2.l| op_sel:[1,0,0] dpp8:[7,6,5,4,3,2,1,0] ; encoding: [0x05,0x0a,0x11,0xd7,0xe9,0x04,0x02,0x20,0x01,0x77,0x39,0x05]
 
-v_pack_b32_f16_e64_dpp v255, -|v255|, -|v255| op_sel:[0,1,0] dpp8:[0,0,0,0,0,0,0,0] fi:1
-// GFX11: v_pack_b32_f16_e64_dpp v255, -|v255|, -|v255| op_sel:[0,1,0] dpp8:[0,0,0,0,0,0,0,0] fi:1 ; encoding: [0xff,0x13,0x11,0xd7,0xea,0xfe,0x03,0x60,0xff,0x00,0x00,0x00]
+v_pack_b32_f16_e64_dpp v255, -|v255.l|, -|v255.h| op_sel:[0,1,0] dpp8:[0,0,0,0,0,0,0,0] fi:1
+// GFX11: v_pack_b32_f16_e64_dpp v255, -|v255.l|, -|v255.h| op_sel:[0,1,0] dpp8:[0,0,0,0,0,0,0,0] fi:1 ; encoding: [0xff,0x13,0x11,0xd7,0xea,0xfe,0x03,0x60,0xff,0x00,0x00,0x00]
 
 v_dot2_f16_f16_e64_dpp v0.l, v1, v2, v3.l dpp8:[0,1,2,3,4,4,4,4]
 // GFX11: v_dot2_f16_f16_e64_dpp v0.l, v1, v2, v3.l dpp8:[0,1,2,3,4,4,4,4] ; encoding: [0x00,0x00,0x66,0xd6,0xe9,0x04,0x0e,0x04,0x01,0x88,0x46,0x92]

llvm/test/MC/AMDGPU/gfx12_asm_vop3.s

@@ -5030,11 +5030,11 @@ v_or_b16 v5, src_scc, vcc_lo
 v_or_b16 v255, 0xfe0b, vcc_hi
 // GFX12: v_or_b16 v255, 0xfe0b, vcc_hi           ; encoding: [0xff,0x00,0x63,0xd7,0xff,0xd6,0x00,0x00,0x0b,0xfe,0x00,0x00]
 
-v_pack_b32_f16 v5, v1, v2
-// GFX12: v_pack_b32_f16 v5, v1, v2               ; encoding: [0x05,0x00,0x11,0xd7,0x01,0x05,0x02,0x00]
+v_pack_b32_f16 v5, v1.l, v2.l
+// GFX12: v_pack_b32_f16 v5, v1.l, v2.l           ; encoding: [0x05,0x00,0x11,0xd7,0x01,0x05,0x02,0x00]
 
-v_pack_b32_f16 v5, v255, v255
-// GFX12: v_pack_b32_f16 v5, v255, v255           ; encoding: [0x05,0x00,0x11,0xd7,0xff,0xff,0x03,0x00]
+v_pack_b32_f16 v5, v255.l, v255.l
+// GFX12: v_pack_b32_f16 v5, v255.l, v255.l       ; encoding: [0x05,0x00,0x11,0xd7,0xff,0xff,0x03,0x00]
 
 v_pack_b32_f16 v5, s1, s2
 // GFX12: v_pack_b32_f16 v5, s1, s2               ; encoding: [0x05,0x00,0x11,0xd7,0x01,0x04,0x00,0x00]
@@ -5066,7 +5066,7 @@ v_pack_b32_f16 v5, null, exec_lo
 v_pack_b32_f16 v5, -1, exec_hi
 // GFX12: v_pack_b32_f16 v5, -1, exec_hi          ; encoding: [0x05,0x00,0x11,0xd7,0xc1,0xfe,0x00,0x00]
 
-v_pack_b32_f16 v5, 0.5, -m0 op_sel:[0,0,0]
+v_pack_b32_f16 v5, 0.5, -m0
 // GFX12: v_pack_b32_f16 v5, 0.5, -m0             ; encoding: [0x05,0x00,0x11,0xd7,0xf0,0xfa,0x00,0x40]
 
 v_pack_b32_f16 v5, -src_scc, |vcc_lo| op_sel:[1,0,0]
@@ -5075,6 +5075,12 @@ v_pack_b32_f16 v5, -src_scc, |vcc_lo| op_sel:[1,0,0]
 v_pack_b32_f16 v255, -|0xfe0b|, -|vcc_hi| op_sel:[0,1,0]
 // GFX12: v_pack_b32_f16 v255, -|0xfe0b|, -|vcc_hi| op_sel:[0,1,0] ; encoding: [0xff,0x13,0x11,0xd7,0xff,0xd6,0x00,0x60,0x0b,0xfe,0x00,0x00]
 
+v_pack_b32_f16 v5, v1.h, v2.l
+// GFX12: v_pack_b32_f16 v5, v1.h, v2.l op_sel:[1,0,0] ; encoding: [0x05,0x08,0x11,0xd7,0x01,0x05,0x02,0x00]
+
+v_pack_b32_f16 v5, v255.l, v255.h
+// GFX12: v_pack_b32_f16 v5, v255.l, v255.h op_sel:[0,1,0] ; encoding: [0x05,0x10,0x11,0xd7,0xff,0xff,0x03,0x00]
+
 v_perm_b32 v5, v1, v2, s3
 // GFX12: v_perm_b32 v5, v1, v2, s3               ; encoding: [0x05,0x00,0x44,0xd6,0x01,0x05,0x0e,0x00]