[llvm] [AMDGPU][True16][MC] VOP3 profile in True16 format (PR #109031)

[Brox Chen via llvm-commits]

================
@@ -194,47 +194,59 @@ v_add_lshl_u32_e64_dpp v5, v1, v2, 0.5 row_xmask:0 row_mask:0x1 bank_mask:0x3 bo
 v_add_lshl_u32_e64_dpp v255, v255, v255, src_scc row_xmask:15 row_mask:0x3 bank_mask:0x0 bound_ctrl:0 fi:1
 // GFX11: [0xff,0x00,0x47,0xd6,0xfa,0xfe,0xf7,0x03,0xff,0x6f,0x05,0x30]
 
-v_add_nc_i16_e64_dpp v5, v1, v2 quad_perm:[3,2,1,0]
+v_add_nc_i16_e64_dpp v5.l, v1.l, v2.l quad_perm:[3,2,1,0]
 // GFX11: [0x05,0x00,0x0d,0xd7,0xfa,0x04,0x02,0x00,0x01,0x1b,0x00,0xff]
 
-v_add_nc_i16_e64_dpp v5, v1, v2 quad_perm:[0,1,2,3]
+v_add_nc_i16_e64_dpp v5.l, v1.l, v2.l quad_perm:[0,1,2,3]
 // GFX11: [0x05,0x00,0x0d,0xd7,0xfa,0x04,0x02,0x00,0x01,0xe4,0x00,0xff]
 
-v_add_nc_i16_e64_dpp v5, v1, v2 row_mirror
+v_add_nc_i16_e64_dpp v5.l, v1.l, v2.l row_mirror
 // GFX11: [0x05,0x00,0x0d,0xd7,0xfa,0x04,0x02,0x00,0x01,0x40,0x01,0xff]
 
-v_add_nc_i16_e64_dpp v5, v1, v2 row_half_mirror
+v_add_nc_i16_e64_dpp v5.l, v1.l, v2.l row_half_mirror
 // GFX11: [0x05,0x00,0x0d,0xd7,0xfa,0x04,0x02,0x00,0x01,0x41,0x01,0xff]
 
-v_add_nc_i16_e64_dpp v5, v1, v2 row_shl:1
+v_add_nc_i16_e64_dpp v5.l, v1.l, v2.l row_shl:1
 // GFX11: [0x05,0x00,0x0d,0xd7,0xfa,0x04,0x02,0x00,0x01,0x01,0x01,0xff]
 
-v_add_nc_i16_e64_dpp v5, v1, v2 row_shl:15
+v_add_nc_i16_e64_dpp v5.l, v1.l, v2.l row_shl:15
 // GFX11: [0x05,0x00,0x0d,0xd7,0xfa,0x04,0x02,0x00,0x01,0x0f,0x01,0xff]
 
-v_add_nc_i16_e64_dpp v5, v1, v2 row_shr:1
+v_add_nc_i16_e64_dpp v5.l, v1.l, v2.l row_shr:1
 // GFX11: [0x05,0x00,0x0d,0xd7,0xfa,0x04,0x02,0x00,0x01,0x11,0x01,0xff]
 
-v_add_nc_i16_e64_dpp v5, v1, v2 row_shr:15
+v_add_nc_i16_e64_dpp v5.l, v1.l, v2.l row_shr:15
 // GFX11: [0x05,0x00,0x0d,0xd7,0xfa,0x04,0x02,0x00,0x01,0x1f,0x01,0xff]
 
-v_add_nc_i16_e64_dpp v5, v1, v2 row_ror:1
+v_add_nc_i16_e64_dpp v5.l, v1.l, v2.l row_ror:1
 // GFX11: [0x05,0x00,0x0d,0xd7,0xfa,0x04,0x02,0x00,0x01,0x21,0x01,0xff]
 
-v_add_nc_i16_e64_dpp v5, v1, v2 row_ror:15
+v_add_nc_i16_e64_dpp v5.l, v1.l, v2.l row_ror:15
 // GFX11: [0x05,0x00,0x0d,0xd7,0xfa,0x04,0x02,0x00,0x01,0x2f,0x01,0xff]
 
-v_add_nc_i16_e64_dpp v5, v1, v2 row_share:0 row_mask:0xf bank_mask:0xf
-// GFX11: [0x05,0x00,0x0d,0xd7,0xfa,0x04,0x02,0x00,0x01,0x50,0x01,0xff]
+v_add_nc_i16_e64_dpp v5.h, v1.h, v2.h row_share:0 row_mask:0xf bank_mask:0xf
+// GFX11: [0x05,0x58,0x0d,0xd7,0xfa,0x04,0x02,0x00,0x01,0x50,0x01,0xff]
+
+v_add_nc_i16_e64_dpp v5.l, v1.h, v2.l row_share:15 row_mask:0x0 bank_mask:0x1
+// GFX11: [0x05,0x08,0x0d,0xd7,0xfa,0x04,0x02,0x00,0x01,0x5f,0x01,0x01]
+
+v_add_nc_i16_e64_dpp v5.l, v1.l, v2.h row_xmask:0 row_mask:0x1 bank_mask:0x3 bound_ctrl:1 fi:0
+// GFX11: [0x05,0x10,0x0d,0xd7,0xfa,0x04,0x02,0x00,0x01,0x60,0x09,0x13]
+
+v_add_nc_i16_e64_dpp v255.h, v255.l, v255.l clamp row_xmask:15 row_mask:0x3 bank_mask:0x0 bound_ctrl:0 fi:1
+// GFX11: [0xff,0xc0,0x0d,0xd7,0xfa,0xfe,0x03,0x00,0xff,0x6f,0x05,0x30]
+
+v_add_nc_i16_e64_dpp v5.h, v1.h, v2.h row_share:0 row_mask:0xf bank_mask:0xf
----------------
broxigarchen wrote:

No I think this is running the script twice. The whole block here are duplicated. Let me remove this

https://github.com/llvm/llvm-project/pull/109031