Skip to content

[AMDGPU] Allow dpp in v_pk_fmac_f16 for GFX9 and GFX10 #144782

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
merged 7 commits into from
Jun 24, 2025
Merged

[AMDGPU] Allow dpp in v_pk_fmac_f16 for GFX9 and GFX10 #144782

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
7 commits
Select commit Hold shift + click to select a range
c51f487
[AMDGPU] Allow dpp in v_pk_fmac_f16 for GFX9 and GFX10
jwanggit86 Jun 18, 2025
76a36b2
for gfx9, use existing VOP2_Real_e32e64_gfx9
jwanggit86 Jun 19, 2025
e16fa23
for gfx10, use existing VOP2_Real_gfx10
jwanggit86 Jun 19, 2025
bccf7eb
Revert "for gfx10, use existing VOP2_Real_gfx10"
jwanggit86 Jun 23, 2025
e9acb6c
GFX9 has SDWA instruction while GFX10 does not. Add tests to
jwanggit86 Jun 23, 2025
ff2c368
(1) A new multiclass for gfx10 (2) tests for disassembler
jwanggit86 Jun 23, 2025
39fcf97
Combine two defs into one
jwanggit86 Jun 24, 2025
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
14 changes: 10 additions & 4 deletions llvm/lib/Target/AMDGPU/VOP2Instructions.td
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -1945,6 +1945,14 @@ let AssemblerPredicate = isGFX10Only, DecoderNamespace = "GFX10" in {
if !cast<VOP2_Pseudo>(NAME#"_e32").Pfl.HasExt32BitDPP then
def _dpp8_gfx10 : VOP2_DPP8<op, !cast<VOP2_Pseudo>(NAME#"_e32")>;
}
multiclass VOP2Only_Real_e32_gfx10<bits<6> op> {
let IsSingle = 1 in
defm NAME: VOP2_Real_e32_gfx10<op>;
}
multiclass VOP2_Real_e32_dpp_dpp8_gfx10<bits<6> op> :
VOP2Only_Real_e32_gfx10<op>,
VOP2_Real_dpp_gfx10<op>,
VOP2_Real_dpp8_gfx10<op>;

//===------------------------- VOP2 (with name) -------------------------===//
multiclass VOP2_Real_e32_gfx10_with_name<bits<6> op, string opName,
Expand Down Expand Up @@ -2168,10 +2176,7 @@ defm V_FMAAK_F16 : VOP2Only_Real_MADK_gfx10<0x038>;
defm V_MAX_F16 : VOP2_Real_gfx10<0x039>;
defm V_MIN_F16 : VOP2_Real_gfx10<0x03a>;
defm V_LDEXP_F16 : VOP2_Real_gfx10<0x03b>;

let IsSingle = 1 in {
defm V_PK_FMAC_F16 : VOP2_Real_e32_gfx10<0x03c>;
}
defm V_PK_FMAC_F16 : VOP2_Real_e32_dpp_dpp8_gfx10<0x03c>;

// VOP2 no carry-in, carry-out.
defm V_ADD_NC_U32 :
Expand Down Expand Up @@ -2560,6 +2565,7 @@ defm V_SUBBREV_CO_U32 : VOP2be_Real_e32e64_gfx9 <0x1e, "V_SUBBREV_U32", "v_s
defm V_ADD_U32 : VOP2_Real_e32e64_gfx9 <0x34>;
defm V_SUB_U32 : VOP2_Real_e32e64_gfx9 <0x35>;
defm V_SUBREV_U32 : VOP2_Real_e32e64_gfx9 <0x36>;
defm V_PK_FMAC_F16 : VOP2_Real_e32e64_gfx9<0x03c>;
} // End AssemblerPredicate = isGFX9Only

defm V_BFM_B32 : VOP2_Real_e64only_vi <0x293>;
Expand Down
12 changes: 12 additions & 0 deletions llvm/test/MC/AMDGPU/gfx10_asm_vop2.s
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -13185,3 +13185,15 @@ v_pk_fmac_f16 v5, -4.0, v2

v_pk_fmac_f16 v5, v1, v255
// GFX10: encoding: [0x01,0xff,0x0b,0x78]

v_pk_fmac_f16 v5, v1, v2
// GFX10: encoding: [0x01,0x05,0x0a,0x78]

v_pk_fmac_f16_dpp v5, v1, v2 quad_perm:[0,1,2,3]
// GFX10: encoding: [0xfa,0x04,0x0a,0x78,0x01,0xe4,0x00,0xff]

v_pk_fmac_f16_dpp v5, v1, v2 quad_perm:[0,1,2,3] row_mask:0x0 bank_mask:0x0
// GFX10: encoding: [0xfa,0x04,0x0a,0x78,0x01,0xe4,0x00,0x00]

v_pk_fmac_f16_dpp v5, v1, v2 dpp8:[7,6,5,4,3,2,1,0]
// GFX10: encoding: [0xe9,0x04,0x0a,0x78,0x01,0x77,0x39,0x05]
3 changes: 3 additions & 0 deletions llvm/test/MC/AMDGPU/gfx10_unsupported_sdwa.s
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -32,6 +32,9 @@ v_min_u16_sdwa v255, v1, v2 dst_sel:DWORD dst_unused:UNUSED_PAD src0_sel:DWORD s
v_mul_lo_u16_sdwa v255, v1, v2 dst_sel:DWORD dst_unused:UNUSED_PAD src0_sel:DWORD src1_sel:DWORD
// CHECK: :[[@LINE-1]]:{{[0-9]+}}: error: sdwa variant of this instruction is not supported

v_pk_fmac_f16_sdwa v255, v1, v2 dst_sel:DWORD dst_unused:UNUSED_PAD src0_sel:DWORD src1_sel:DWORD
// CHECK: :[[@LINE-1]]:{{[0-9]+}}: error: sdwa variant of this instruction is not supported

v_sub_co_u32_sdwa v0, v0, v4 dst_sel:DWORD dst_unused:UNUSED_PAD src0_sel:DWORD src1_sel:BYTE_0
// CHECK: :[[@LINE-1]]:{{[0-9]+}}: error: sdwa variant of this instruction is not supported

Expand Down
91 changes: 91 additions & 0 deletions llvm/test/MC/AMDGPU/gfx9_asm_vop2_features.s
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -0,0 +1,91 @@
// RUN: llvm-mc -triple=amdgcn -mcpu=gfx908 -show-encoding %s | FileCheck --check-prefix=CHECK-MI %s
// RUN: llvm-mc -triple=amdgcn -mcpu=gfx90a -show-encoding %s | FileCheck --check-prefix=CHECK-MI %s
// RUN: llvm-mc -triple=amdgcn -mcpu=gfx942 -show-encoding %s | FileCheck --check-prefix=CHECK-MI %s
// RUN: llvm-mc -triple=amdgcn -mcpu=gfx950 -show-encoding %s | FileCheck --check-prefix=CHECK-MI %s

v_pk_fmac_f16 v5, v1, v2
// CHECK-MI: [0x01,0x05,0x0a,0x78]

v_pk_fmac_f16 v5, v1, v2 quad_perm:[0,1,2,3]
// CHECK-MI: [0xfa,0x04,0x0a,0x78,0x01,0xe4,0x00,0xff]

v_pk_fmac_f16 v5, v1, v2 quad_perm:[0,1,2,3] row_mask:0x0 bank_mask:0x0
// CHECK-MI: [0xfa,0x04,0x0a,0x78,0x01,0xe4,0x00,0x00]

v_pk_fmac_f16_sdwa v5, v1, v2 dst_sel:DWORD dst_unused:UNUSED_PAD src0_sel:DWORD src1_sel:DWORD
// CHECK-MI: [0xf9,0x04,0x0a,0x78,0x01,0x06,0x06,0x06]

v_pk_fmac_f16_sdwa v5, v1, v2 dst_sel:DWORD dst_unused:UNUSED_PAD src0_sel:DWORD src1_sel:DWORD
// CHECK-MI: [0xf9,0x04,0x0a,0x78,0x01,0x06,0x06,0x06]

v_pk_fmac_f16_sdwa v5, v1, v2 dst_sel:BYTE_0 dst_unused:UNUSED_PAD src0_sel:DWORD src1_sel:DWORD
// CHECK-MI: [0xf9,0x04,0x0a,0x78,0x01,0x00,0x06,0x06]

v_pk_fmac_f16_sdwa v5, v1, v2 dst_sel:BYTE_1 dst_unused:UNUSED_PAD src0_sel:DWORD src1_sel:DWORD
// CHECK-MI: [0xf9,0x04,0x0a,0x78,0x01,0x01,0x06,0x06]

v_pk_fmac_f16_sdwa v5, v1, v2 dst_sel:BYTE_2 dst_unused:UNUSED_PAD src0_sel:DWORD src1_sel:DWORD
// CHECK-MI: [0xf9,0x04,0x0a,0x78,0x01,0x02,0x06,0x06]

v_pk_fmac_f16_sdwa v5, v1, v2 dst_sel:BYTE_3 dst_unused:UNUSED_PAD src0_sel:DWORD src1_sel:DWORD
// CHECK-MI: [0xf9,0x04,0x0a,0x78,0x01,0x03,0x06,0x06]

v_pk_fmac_f16_sdwa v5, v1, v2 dst_sel:WORD_0 dst_unused:UNUSED_PAD src0_sel:DWORD src1_sel:DWORD
// CHECK-MI: [0xf9,0x04,0x0a,0x78,0x01,0x04,0x06,0x06]

v_pk_fmac_f16_sdwa v5, v1, v2 dst_sel:WORD_1 dst_unused:UNUSED_PAD src0_sel:DWORD src1_sel:DWORD
// CHECK-MI: [0xf9,0x04,0x0a,0x78,0x01,0x05,0x06,0x06]

v_pk_fmac_f16_sdwa v5, v1, v2 dst_sel:DWORD dst_unused:UNUSED_SEXT src0_sel:DWORD src1_sel:DWORD
// CHECK-MI: [0xf9,0x04,0x0a,0x78,0x01,0x0e,0x06,0x06]

v_pk_fmac_f16_sdwa v5, v1, v2 dst_sel:DWORD dst_unused:UNUSED_PRESERVE src0_sel:DWORD src1_sel:DWORD
// CHECK-MI: [0xf9,0x04,0x0a,0x78,0x01,0x16,0x06,0x06]

v_pk_fmac_f16_sdwa v5, v1, v2 dst_sel:DWORD dst_unused:UNUSED_PRESERVE src0_sel:DWORD src1_sel:DWORD
// CHECK-MI: [0xf9,0x04,0x0a,0x78,0x01,0x16,0x06,0x06]

v_pk_fmac_f16_sdwa v5, v1, v2 dst_sel:DWORD dst_unused:UNUSED_PAD src0_sel:DWORD src1_sel:DWORD
// CHECK-MI: [0xf9,0x04,0x0a,0x78,0x01,0x06,0x06,0x06]

v_pk_fmac_f16_sdwa v5, v1, v2 dst_sel:DWORD dst_unused:UNUSED_PAD src0_sel:BYTE_0 src1_sel:DWORD
// CHECK-MI: [0xf9,0x04,0x0a,0x78,0x01,0x06,0x00,0x06]

v_pk_fmac_f16_sdwa v5, v1, v2 dst_sel:DWORD dst_unused:UNUSED_PAD src0_sel:BYTE_1 src1_sel:DWORD
// CHECK-MI: [0xf9,0x04,0x0a,0x78,0x01,0x06,0x01,0x06]

v_pk_fmac_f16_sdwa v5, v1, v2 dst_sel:DWORD dst_unused:UNUSED_PAD src0_sel:BYTE_2 src1_sel:DWORD
// CHECK-MI: [0xf9,0x04,0x0a,0x78,0x01,0x06,0x02,0x06]

v_pk_fmac_f16_sdwa v5, v1, v2 dst_sel:DWORD dst_unused:UNUSED_PAD src0_sel:BYTE_3 src1_sel:DWORD
// CHECK-MI: [0xf9,0x04,0x0a,0x78,0x01,0x06,0x03,0x06]

v_pk_fmac_f16_sdwa v5, v1, v2 dst_sel:DWORD dst_unused:UNUSED_PAD src0_sel:WORD_0 src1_sel:DWORD
// CHECK-MI: [0xf9,0x04,0x0a,0x78,0x01,0x06,0x04,0x06]

v_pk_fmac_f16_sdwa v5, v1, v2 dst_sel:DWORD dst_unused:UNUSED_PAD src0_sel:WORD_1 src1_sel:DWORD
// CHECK-MI: [0xf9,0x04,0x0a,0x78,0x01,0x06,0x05,0x06]

v_pk_fmac_f16_sdwa v5, v1, v2 dst_sel:DWORD dst_unused:UNUSED_PAD src0_sel:DWORD src1_sel:DWORD
// CHECK-MI: [0xf9,0x04,0x0a,0x78,0x01,0x06,0x06,0x06]

v_pk_fmac_f16_sdwa v5, v1, v2 dst_sel:DWORD dst_unused:UNUSED_PAD src0_sel:DWORD src1_sel:BYTE_0
// CHECK-MI: [0xf9,0x04,0x0a,0x78,0x01,0x06,0x06,0x00]

v_pk_fmac_f16_sdwa v5, v1, v2 dst_sel:DWORD dst_unused:UNUSED_PAD src0_sel:DWORD src1_sel:BYTE_1
// CHECK-MI: [0xf9,0x04,0x0a,0x78,0x01,0x06,0x06,0x01]

v_pk_fmac_f16_sdwa v5, v1, v2 dst_sel:DWORD dst_unused:UNUSED_PAD src0_sel:DWORD src1_sel:BYTE_2
// CHECK-MI: [0xf9,0x04,0x0a,0x78,0x01,0x06,0x06,0x02]

v_pk_fmac_f16_sdwa v5, v1, v2 dst_sel:DWORD dst_unused:UNUSED_PAD src0_sel:DWORD src1_sel:BYTE_3
// CHECK-MI: [0xf9,0x04,0x0a,0x78,0x01,0x06,0x06,0x03]

v_pk_fmac_f16_sdwa v5, v1, v2 dst_sel:DWORD dst_unused:UNUSED_PAD src0_sel:DWORD src1_sel:WORD_0
// CHECK-MI: [0xf9,0x04,0x0a,0x78,0x01,0x06,0x06,0x04]

v_pk_fmac_f16_sdwa v5, v1, v2 dst_sel:DWORD dst_unused:UNUSED_PAD src0_sel:DWORD src1_sel:WORD_1
// CHECK-MI: [0xf9,0x04,0x0a,0x78,0x01,0x06,0x06,0x05]

v_pk_fmac_f16_sdwa v5, v1, sext(v2) dst_sel:DWORD dst_unused:UNUSED_PAD src0_sel:DWORD src1_sel:DWORD
// CHECK-MI: [0xf9,0x04,0x0a,0x78,0x01,0x06,0x06,0x16]
7 changes: 7 additions & 0 deletions llvm/test/MC/Disassembler/AMDGPU/gfx10_vop2_dpp16.txt
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -2476,3 +2476,10 @@
# W32: v_cndmask_b32_dpp v5, -|v1|, -|v2|, vcc_lo quad_perm:[0,1,2,3] row_mask:0xf bank_mask:0xf ; encoding: [0xfa,0x04,0x0a,0x02,0x01,0xe4,0xf0,0xff]
# W64: v_cndmask_b32_dpp v5, -|v1|, -|v2|, vcc quad_perm:[0,1,2,3] row_mask:0xf bank_mask:0xf ; encoding: [0xfa,0x04,0x0a,0x02,0x01,0xe4,0xf0,0xff]
0xfa,0x04,0x0a,0x02,0x01,0xe4,0xf0,0xff

# GFX10: v_pk_fmac_f16_dpp v5, v1, v2 quad_perm:[0,1,2,3] row_mask:0xf bank_mask:0xf ; encoding: [0xfa,0x04,0x0a,0x78,0x01,0xe4,0x00,0xff]
0xfa,0x04,0x0a,0x78,0x01,0xe4,0x00,0xff

# GFX10: v_pk_fmac_f16_dpp v5, v1, v2 quad_perm:[0,1,2,3] row_mask:0x0 bank_mask:0x0 ; encoding: [0xfa,0x04,0x0a,0x78,0x01,0xe4,0x00,0x00]
0xfa,0x04,0x0a,0x78,0x01,0xe4,0x00,0x00

3 changes: 3 additions & 0 deletions llvm/test/MC/Disassembler/AMDGPU/gfx10_vop2_dpp8.txt
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -222,3 +222,6 @@
# W32: v_cndmask_b32_dpp v0, v1, v2, vcc_lo dpp8:[0,1,2,3,4,5,6,7] fi:1 ; encoding: [0xea,0x04,0x00,0x02,0x01,0x88,0xc6,0xfa]
# W64: v_cndmask_b32_dpp v0, v1, v2, vcc dpp8:[0,1,2,3,4,5,6,7] fi:1 ; encoding: [0xea,0x04,0x00,0x02,0x01,0x88,0xc6,0xfa]
0xea,0x04,0x00,0x02,0x01,0x88,0xc6,0xfa

# GFX10: v_pk_fmac_f16_dpp v5, v1, v2 dpp8:[7,6,5,4,3,2,1,0] ; encoding: [0xe9,0x04,0x0a,0x78,0x01,0x77,0x39,0x05]
0xe9,0x04,0x0a,0x78,0x01,0x77,0x39,0x05
92 changes: 92 additions & 0 deletions llvm/test/MC/Disassembler/AMDGPU/gfx9_vop2_features.txt
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -0,0 +1,92 @@
# RUN: llvm-mc -triple=amdgcn -mcpu=gfx908 -disassemble -show-encoding < %s | FileCheck -check-prefix=CHECK-MI %s
# RUN: llvm-mc -triple=amdgcn -mcpu=gfx90a -disassemble -show-encoding < %s | FileCheck -check-prefix=CHECK-MI %s
# RUN: llvm-mc -triple=amdgcn -mcpu=gfx942 -disassemble -show-encoding < %s | FileCheck -check-prefix=CHECK-MI %s
# RUN: llvm-mc -triple=amdgcn -mcpu=gfx950 -disassemble -show-encoding < %s | FileCheck -check-prefix=CHECK-MI %s

# CHECK-MI: v_pk_fmac_f16_e32 v5, v1, v2
0x01,0x05,0x0a,0x78

# CHECK-MI: v_pk_fmac_f16_dpp v5, v1, v2 quad_perm:[0,1,2,3] row_mask:0xf bank_mask:0xf
0xfa,0x04,0x0a,0x78,0x01,0xe4,0x00,0xff

# CHECK-MI: v_pk_fmac_f16_dpp v5, v1, v2 quad_perm:[0,1,2,3] row_mask:0x0 bank_mask:0x0
Copy link
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

sdwa disasm tests needed.

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Done.

0xfa,0x04,0x0a,0x78,0x01,0xe4,0x00,0x00

# CHECK-MI: v_pk_fmac_f16_sdwa v5, v1, v2 dst_sel:DWORD dst_unused:UNUSED_PAD src0_sel:DWORD src1_sel:DWORD
0xf9,0x04,0x0a,0x78,0x01,0x06,0x06,0x06

# CHECK-MI: v_pk_fmac_f16_sdwa v5, v1, v2 dst_sel:DWORD dst_unused:UNUSED_PAD src0_sel:DWORD src1_sel:DWORD
0xf9,0x04,0x0a,0x78,0x01,0x06,0x06,0x06

# CHECK-MI: v_pk_fmac_f16_sdwa v5, v1, v2 dst_sel:BYTE_0 dst_unused:UNUSED_PAD src0_sel:DWORD src1_sel:DWORD
0xf9,0x04,0x0a,0x78,0x01,0x00,0x06,0x06

# CHECK-MI: v_pk_fmac_f16_sdwa v5, v1, v2 dst_sel:BYTE_1 dst_unused:UNUSED_PAD src0_sel:DWORD src1_sel:DWORD
0xf9,0x04,0x0a,0x78,0x01,0x01,0x06,0x06

# CHECK-MI: v_pk_fmac_f16_sdwa v5, v1, v2 dst_sel:BYTE_2 dst_unused:UNUSED_PAD src0_sel:DWORD src1_sel:DWORD
0xf9,0x04,0x0a,0x78,0x01,0x02,0x06,0x06

# CHECK-MI: v_pk_fmac_f16_sdwa v5, v1, v2 dst_sel:BYTE_3 dst_unused:UNUSED_PAD src0_sel:DWORD src1_sel:DWORD
0xf9,0x04,0x0a,0x78,0x01,0x03,0x06,0x06

# CHECK-MI: v_pk_fmac_f16_sdwa v5, v1, v2 dst_sel:WORD_0 dst_unused:UNUSED_PAD src0_sel:DWORD src1_sel:DWORD
0xf9,0x04,0x0a,0x78,0x01,0x04,0x06,0x06

# CHECK-MI: v_pk_fmac_f16_sdwa v5, v1, v2 dst_sel:WORD_1 dst_unused:UNUSED_PAD src0_sel:DWORD src1_sel:DWORD
0xf9,0x04,0x0a,0x78,0x01,0x05,0x06,0x06

# CHECK-MI: v_pk_fmac_f16_sdwa v5, v1, v2 dst_sel:DWORD dst_unused:UNUSED_SEXT src0_sel:DWORD src1_sel:DWORD
0xf9,0x04,0x0a,0x78,0x01,0x0e,0x06,0x06

# CHECK-MI: v_pk_fmac_f16_sdwa v5, v1, v2 dst_sel:DWORD dst_unused:UNUSED_PRESERVE src0_sel:DWORD src1_sel:DWORD
0xf9,0x04,0x0a,0x78,0x01,0x16,0x06,0x06

# CHECK-MI: v_pk_fmac_f16_sdwa v5, v1, v2 dst_sel:DWORD dst_unused:UNUSED_PRESERVE src0_sel:DWORD src1_sel:DWORD
0xf9,0x04,0x0a,0x78,0x01,0x16,0x06,0x06

# CHECK-MI: v_pk_fmac_f16_sdwa v5, v1, v2 dst_sel:DWORD dst_unused:UNUSED_PAD src0_sel:DWORD src1_sel:DWORD
0xf9,0x04,0x0a,0x78,0x01,0x06,0x06,0x06

# CHECK-MI: v_pk_fmac_f16_sdwa v5, v1, v2 dst_sel:DWORD dst_unused:UNUSED_PAD src0_sel:BYTE_0 src1_sel:DWORD
0xf9,0x04,0x0a,0x78,0x01,0x06,0x00,0x06

# CHECK-MI: v_pk_fmac_f16_sdwa v5, v1, v2 dst_sel:DWORD dst_unused:UNUSED_PAD src0_sel:BYTE_1 src1_sel:DWORD
0xf9,0x04,0x0a,0x78,0x01,0x06,0x01,0x06

# CHECK-MI: v_pk_fmac_f16_sdwa v5, v1, v2 dst_sel:DWORD dst_unused:UNUSED_PAD src0_sel:BYTE_2 src1_sel:DWORD
0xf9,0x04,0x0a,0x78,0x01,0x06,0x02,0x06

# CHECK-MI: v_pk_fmac_f16_sdwa v5, v1, v2 dst_sel:DWORD dst_unused:UNUSED_PAD src0_sel:BYTE_3 src1_sel:DWORD
0xf9,0x04,0x0a,0x78,0x01,0x06,0x03,0x06

# CHECK-MI: v_pk_fmac_f16_sdwa v5, v1, v2 dst_sel:DWORD dst_unused:UNUSED_PAD src0_sel:WORD_0 src1_sel:DWORD
0xf9,0x04,0x0a,0x78,0x01,0x06,0x04,0x06

# CHECK-MI: v_pk_fmac_f16_sdwa v5, v1, v2 dst_sel:DWORD dst_unused:UNUSED_PAD src0_sel:WORD_1 src1_sel:DWORD
0xf9,0x04,0x0a,0x78,0x01,0x06,0x05,0x06

# CHECK-MI: v_pk_fmac_f16_sdwa v5, v1, v2 dst_sel:DWORD dst_unused:UNUSED_PAD src0_sel:DWORD src1_sel:DWORD
0xf9,0x04,0x0a,0x78,0x01,0x06,0x06,0x06

# CHECK-MI: v_pk_fmac_f16_sdwa v5, v1, v2 dst_sel:DWORD dst_unused:UNUSED_PAD src0_sel:DWORD src1_sel:BYTE_0
0xf9,0x04,0x0a,0x78,0x01,0x06,0x06,0x00

# CHECK-MI: v_pk_fmac_f16_sdwa v5, v1, v2 dst_sel:DWORD dst_unused:UNUSED_PAD src0_sel:DWORD src1_sel:BYTE_1
0xf9,0x04,0x0a,0x78,0x01,0x06,0x06,0x01

# CHECK-MI: v_pk_fmac_f16_sdwa v5, v1, v2 dst_sel:DWORD dst_unused:UNUSED_PAD src0_sel:DWORD src1_sel:BYTE_2
0xf9,0x04,0x0a,0x78,0x01,0x06,0x06,0x02

# CHECK-MI: v_pk_fmac_f16_sdwa v5, v1, v2 dst_sel:DWORD dst_unused:UNUSED_PAD src0_sel:DWORD src1_sel:BYTE_3
0xf9,0x04,0x0a,0x78,0x01,0x06,0x06,0x03

# CHECK-MI: v_pk_fmac_f16_sdwa v5, v1, v2 dst_sel:DWORD dst_unused:UNUSED_PAD src0_sel:DWORD src1_sel:WORD_0
0xf9,0x04,0x0a,0x78,0x01,0x06,0x06,0x04

# CHECK-MI: v_pk_fmac_f16_sdwa v5, v1, v2 dst_sel:DWORD dst_unused:UNUSED_PAD src0_sel:DWORD src1_sel:WORD_1
0xf9,0x04,0x0a,0x78,0x01,0x06,0x06,0x05

# CHECK-MI: v_pk_fmac_f16_sdwa v5, v1, sext(v2) dst_sel:DWORD dst_unused:UNUSED_PAD src0_sel:DWORD src1_sel:DWORD
0xf9,0x04,0x0a,0x78,0x01,0x06,0x06,0x16