Skip to content

[AMDGPU][True16][MC] update a few mc test for true16 #135816

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

Jump to bottom
Merged
merged 2 commits into from
Apr 16, 2025
Merged

[AMDGPU][True16][MC] update a few mc test for true16 #135816

merged 2 commits into from
Apr 16, 2025

Conversation

@broxigarchen
Copy link
Contributor

@broxigarchen broxigarchen commented Apr 15, 2025 •
edited
Loading

This is another NFC patch.

Update mc test for a few true16 instructions by duplicating the file to fake16 versions and udpate mattr flag with +/-real-true16. Also added some fake16 file that are not properly created before

@broxigarchen broxigarchen changed the title clean up of true16 mc changes [AMDGPU][True16][MC] update a few mc test in bmm and gfx11 Apr 15, 2025
@broxigarchen broxigarchen changed the title [AMDGPU][True16][MC] update a few mc test in bmm and gfx11 [AMDGPU][True16][MC] update a few mc test for true16 Apr 15, 2025
@broxigarchen broxigarchen marked this pull request as ready for review April 15, 2025 17:02
@broxigarchen broxigarchen requested review from Sisyph and shiltian April 15, 2025 17:02
@llvmbot llvmbot added backend:AMDGPU llvm:mc Machine (object) code labels Apr 15, 2025
@llvmbot
Copy link
Member

llvmbot commented Apr 15, 2025

@llvm/pr-subscribers-backend-amdgpu

Author: Brox Chen (broxigarchen)

Changes

This is another NFC patch.

Update mc test for a few true16 instructions by duplicating the file to fake16 versions and udpate mattr flag with +/-real-true16

Patch is 46.95 KiB, truncated to 20.00 KiB below, full version: https://github.com/llvm/llvm-project/pull/135816.diff

8 Files Affected:

  • (added) llvm/test/MC/AMDGPU/bf16_imm-fake16.s (+114)
  • (modified) llvm/test/MC/AMDGPU/bf16_imm.s (+32-32)
  • (added) llvm/test/MC/AMDGPU/gfx11-promotions-fake16.s (+353)
  • (modified) llvm/test/MC/AMDGPU/gfx11-promotions.s (+33-33)
  • (added) llvm/test/MC/AMDGPU/gfx1150_asm_features-fake16.s (+48)
  • (modified) llvm/test/MC/AMDGPU/gfx1150_asm_features.s (+10-10)
  • (added) llvm/test/MC/AMDGPU/gfx11_asm_t16.s (+59)
  • (modified) llvm/test/MC/AMDGPU/gfx11_asm_vop3_alias.s (+4-4)
diff --git a/llvm/test/MC/AMDGPU/bf16_imm-fake16.s b/llvm/test/MC/AMDGPU/bf16_imm-fake16.s
new file mode 100644
index 0000000000000..ee697bee6ab2d
--- /dev/null
+++ b/llvm/test/MC/AMDGPU/bf16_imm-fake16.s
@@ -0,0 +1,114 @@
+// NOTE: Assertions have been autogenerated by utils/update_mc_test_checks.py UTC_ARGS: --unique --version 5
+// RUN: llvm-mc -triple=amdgcn -mcpu=gfx1100 -mattr=-real-true16 -show-encoding %s | FileCheck %s
+// RUN: llvm-mc -triple=amdgcn -mcpu=gfx1200 -mattr=-real-true16 -show-encoding %s | FileCheck %s
+
+v_dot2_bf16_bf16 v5, v1, v2, 100.0
+// CHECK: v_dot2_bf16_bf16 v5, v1, v2, 0x42c8     ; encoding: [0x05,0x00,0x67,0xd6,0x01,0x05,0xfe,0x03,0xc8,0x42,0x00,0x00]
+
+v_dot2_bf16_bf16 v2, v0, 1.0, v2
+// CHECK: v_dot2_bf16_bf16 v2, v0, 1.0, v2        ; encoding: [0x02,0x00,0x67,0xd6,0x00,0xe5,0x09,0x04]
+
+v_dot2_bf16_bf16 v2, 1.0, v0, v2
+// CHECK: v_dot2_bf16_bf16 v2, 1.0, v0, v2        ; encoding: [0x02,0x00,0x67,0xd6,0xf2,0x00,0x0a,0x04]
+
+v_dot2_bf16_bf16 v5, v1, v2, 1.0
+// CHECK: v_dot2_bf16_bf16 v5, v1, v2, 1.0        ; encoding: [0x05,0x00,0x67,0xd6,0x01,0x05,0xca,0x03]
+
+v_dot2_bf16_bf16 v2, v0, -1.0, v2
+// CHECK: v_dot2_bf16_bf16 v2, v0, -1.0, v2       ; encoding: [0x02,0x00,0x67,0xd6,0x00,0xe7,0x09,0x04]
+
+v_dot2_bf16_bf16 v2, v0, 0.5, v2
+// CHECK: v_dot2_bf16_bf16 v2, v0, 0.5, v2        ; encoding: [0x02,0x00,0x67,0xd6,0x00,0xe1,0x09,0x04]
+
+v_dot2_bf16_bf16 v2, v0, -0.5, v2
+// CHECK: v_dot2_bf16_bf16 v2, v0, -0.5, v2       ; encoding: [0x02,0x00,0x67,0xd6,0x00,0xe3,0x09,0x04]
+
+v_dot2_bf16_bf16 v2, v0, 2.0, v2
+// CHECK: v_dot2_bf16_bf16 v2, v0, 2.0, v2        ; encoding: [0x02,0x00,0x67,0xd6,0x00,0xe9,0x09,0x04]
+
+v_dot2_bf16_bf16 v2, v0, -2.0, v2
+// CHECK: v_dot2_bf16_bf16 v2, v0, -2.0, v2       ; encoding: [0x02,0x00,0x67,0xd6,0x00,0xeb,0x09,0x04]
+
+v_dot2_bf16_bf16 v2, v0, 4.0, v2
+// CHECK: v_dot2_bf16_bf16 v2, v0, 4.0, v2        ; encoding: [0x02,0x00,0x67,0xd6,0x00,0xed,0x09,0x04]
+
+v_dot2_bf16_bf16 v2, v0, -4.0, v2
+// CHECK: v_dot2_bf16_bf16 v2, v0, -4.0, v2       ; encoding: [0x02,0x00,0x67,0xd6,0x00,0xef,0x09,0x04]
+
+// Check 1/(2*pi) rounded value and ideomatic fp32 0.15915494 value
+// which cannot be accurately represented in bf16.
+
+v_dot2_bf16_bf16 v2, v0, 0.158203125, v2
+// CHECK: v_dot2_bf16_bf16 v2, v0, 0.15915494, v2 ; encoding: [0x02,0x00,0x67,0xd6,0x00,0xf1,0x09,0x04]
+
+v_dot2_bf16_bf16 v2, v0, 0.15915494, v2
+// CHECK: v_dot2_bf16_bf16 v2, v0, 0.15915494, v2 ; encoding: [0x02,0x00,0x67,0xd6,0x00,0xf1,0x09,0x04]
+
+v_dot2_bf16_bf16 v2, v0, 0x3e22, v2
+// CHECK: v_dot2_bf16_bf16 v2, v0, 0.15915494, v2 ; encoding: [0x02,0x00,0x67,0xd6,0x00,0xf1,0x09,0x04]
+
+v_dot2_bf16_bf16 v2, v0, v2, 0.15915494
+// CHECK: v_dot2_bf16_bf16 v2, v0, v2, 0.15915494 ; encoding: [0x02,0x00,0x67,0xd6,0x00,0x05,0xe2,0x03]
+
+v_dot2_f32_bf16 v2, v1, 0, v2
+// CHECK: v_dot2_f32_bf16 v2, v1, 0, v2           ; encoding: [0x02,0x40,0x1a,0xcc,0x01,0x01,0x09,0x1c]
+
+v_dot2_f32_bf16 v2, v1, 0.5, v2
+// CHECK: v_dot2_f32_bf16 v2, v1, 0.5, v2         ; encoding: [0x02,0x40,0x1a,0xcc,0x01,0xe1,0x09,0x1c]
+
+v_dot2_f32_bf16 v2, v1, -0.5, v2
+// CHECK: v_dot2_f32_bf16 v2, v1, -0.5, v2        ; encoding: [0x02,0x40,0x1a,0xcc,0x01,0xe3,0x09,0x1c]
+
+v_dot2_f32_bf16 v2, v1, 1.0, v2
+// CHECK: v_dot2_f32_bf16 v2, v1, 1.0, v2         ; encoding: [0x02,0x40,0x1a,0xcc,0x01,0xe5,0x09,0x1c]
+
+v_dot2_f32_bf16 v2, v1, -1.0, v2
+// CHECK: v_dot2_f32_bf16 v2, v1, -1.0, v2        ; encoding: [0x02,0x40,0x1a,0xcc,0x01,0xe7,0x09,0x1c]
+
+v_dot2_f32_bf16 v2, v1, 2.0, v2
+// CHECK: v_dot2_f32_bf16 v2, v1, 2.0, v2         ; encoding: [0x02,0x40,0x1a,0xcc,0x01,0xe9,0x09,0x1c]
+
+v_dot2_f32_bf16 v2, v1, -2.0, v2
+// CHECK: v_dot2_f32_bf16 v2, v1, -2.0, v2        ; encoding: [0x02,0x40,0x1a,0xcc,0x01,0xeb,0x09,0x1c]
+
+v_dot2_f32_bf16 v2, v1, 4.0, v2
+// CHECK: v_dot2_f32_bf16 v2, v1, 4.0, v2         ; encoding: [0x02,0x40,0x1a,0xcc,0x01,0xed,0x09,0x1c]
+
+v_dot2_f32_bf16 v2, v1, -4.0, v2
+// CHECK: v_dot2_f32_bf16 v2, v1, -4.0, v2        ; encoding: [0x02,0x40,0x1a,0xcc,0x01,0xef,0x09,0x1c]
+
+v_dot2_f32_bf16 v2, v1, 0.15915494, v2
+// CHECK: v_dot2_f32_bf16 v2, v1, 0.15915494, v2  ; encoding: [0x02,0x40,0x1a,0xcc,0x01,0xf1,0x09,0x1c]
+
+v_dot2_f32_bf16 v2, v1, 0x3e22, v2
+// CHECK: v_dot2_f32_bf16 v2, v1, 0.15915494, v2  ; encoding: [0x02,0x40,0x1a,0xcc,0x01,0xf1,0x09,0x1c]
+
+v_dot2_f32_bf16 v2, 0.5, v1, v2
+// CHECK: v_dot2_f32_bf16 v2, 0.5, v1, v2         ; encoding: [0x02,0x40,0x1a,0xcc,0xf0,0x02,0x0a,0x1c]
+
+v_dot2_f32_bf16 v2, -0.5, v1, v2
+// CHECK: v_dot2_f32_bf16 v2, -0.5, v1, v2        ; encoding: [0x02,0x40,0x1a,0xcc,0xf1,0x02,0x0a,0x1c]
+
+v_dot2_f32_bf16 v2, 1.0, v1, v2
+// CHECK: v_dot2_f32_bf16 v2, 1.0, v1, v2         ; encoding: [0x02,0x40,0x1a,0xcc,0xf2,0x02,0x0a,0x1c]
+
+v_dot2_f32_bf16 v2, -1.0, v1, v2
+// CHECK: v_dot2_f32_bf16 v2, -1.0, v1, v2        ; encoding: [0x02,0x40,0x1a,0xcc,0xf3,0x02,0x0a,0x1c]
+
+v_dot2_f32_bf16 v2, 2.0, v1, v2
+// CHECK: v_dot2_f32_bf16 v2, 2.0, v1, v2         ; encoding: [0x02,0x40,0x1a,0xcc,0xf4,0x02,0x0a,0x1c]
+
+v_dot2_f32_bf16 v2, -2.0, v1, v2
+// CHECK: v_dot2_f32_bf16 v2, -2.0, v1, v2        ; encoding: [0x02,0x40,0x1a,0xcc,0xf5,0x02,0x0a,0x1c]
+
+v_dot2_f32_bf16 v2, 4.0, v1, v2
+// CHECK: v_dot2_f32_bf16 v2, 4.0, v1, v2         ; encoding: [0x02,0x40,0x1a,0xcc,0xf6,0x02,0x0a,0x1c]
+
+v_dot2_f32_bf16 v2, -4.0, v1, v2
+// CHECK: v_dot2_f32_bf16 v2, -4.0, v1, v2        ; encoding: [0x02,0x40,0x1a,0xcc,0xf7,0x02,0x0a,0x1c]
+
+v_dot2_f32_bf16 v2, 100.0, v1, v2
+// CHECK: v_dot2_f32_bf16 v2, 0x42c8, v1, v2      ; encoding: [0x02,0x40,0x1a,0xcc,0xff,0x02,0x0a,0x1c,0xc8,0x42,0x00,0x00]
+
+v_dot2_f32_bf16 v2, v1, 100.0, v2
+// CHECK: v_dot2_f32_bf16 v2, v1, 0x42c8, v2      ; encoding: [0x02,0x40,0x1a,0xcc,0x01,0xff,0x09,0x1c,0xc8,0x42,0x00,0x00]
diff --git a/llvm/test/MC/AMDGPU/bf16_imm.s b/llvm/test/MC/AMDGPU/bf16_imm.s
index 7cf18103adfe5..d79649073aa89 100644
--- a/llvm/test/MC/AMDGPU/bf16_imm.s
+++ b/llvm/test/MC/AMDGPU/bf16_imm.s
@@ -1,54 +1,54 @@
 // NOTE: Assertions have been autogenerated by utils/update_mc_test_checks.py UTC_ARGS: --unique --version 5
-// RUN: llvm-mc -triple=amdgcn -mcpu=gfx1100 -show-encoding %s | FileCheck %s
-// RUN: llvm-mc -triple=amdgcn -mcpu=gfx1200 -show-encoding %s | FileCheck %s
+// RUN: llvm-mc -triple=amdgcn -mcpu=gfx1100 -mattr=+real-true16 -show-encoding %s | FileCheck %s
+// RUN: llvm-mc -triple=amdgcn -mcpu=gfx1200 -mattr=+real-true16 -show-encoding %s | FileCheck %s
 
-v_dot2_bf16_bf16 v5, v1, v2, 100.0
-// CHECK: v_dot2_bf16_bf16 v5, v1, v2, 0x42c8     ; encoding: [0x05,0x00,0x67,0xd6,0x01,0x05,0xfe,0x03,0xc8,0x42,0x00,0x00]
+v_dot2_bf16_bf16 v5.l, v1, v2, 100.0
+// CHECK: v_dot2_bf16_bf16 v5.l, v1, v2, 0x42c8   ; encoding: [0x05,0x00,0x67,0xd6,0x01,0x05,0xfe,0x03,0xc8,0x42,0x00,0x00]
 
-v_dot2_bf16_bf16 v2, v0, 1.0, v2
-// CHECK: v_dot2_bf16_bf16 v2, v0, 1.0, v2        ; encoding: [0x02,0x00,0x67,0xd6,0x00,0xe5,0x09,0x04]
+v_dot2_bf16_bf16 v2.l, v0, 1.0, v2.l
+// CHECK: v_dot2_bf16_bf16 v2.l, v0, 1.0, v2.l    ; encoding: [0x02,0x00,0x67,0xd6,0x00,0xe5,0x09,0x04]
 
-v_dot2_bf16_bf16 v2, 1.0, v0, v2
-// CHECK: v_dot2_bf16_bf16 v2, 1.0, v0, v2        ; encoding: [0x02,0x00,0x67,0xd6,0xf2,0x00,0x0a,0x04]
+v_dot2_bf16_bf16 v2.l, 1.0, v0, v2.l
+// CHECK: v_dot2_bf16_bf16 v2.l, 1.0, v0, v2.l    ; encoding: [0x02,0x00,0x67,0xd6,0xf2,0x00,0x0a,0x04]
 
-v_dot2_bf16_bf16 v5, v1, v2, 1.0
-// CHECK: v_dot2_bf16_bf16 v5, v1, v2, 1.0        ; encoding: [0x05,0x00,0x67,0xd6,0x01,0x05,0xca,0x03]
+v_dot2_bf16_bf16 v5.l, v1, v2, 1.0
+// CHECK: v_dot2_bf16_bf16 v5.l, v1, v2, 1.0      ; encoding: [0x05,0x00,0x67,0xd6,0x01,0x05,0xca,0x03]
 
-v_dot2_bf16_bf16 v2, v0, -1.0, v2
-// CHECK: v_dot2_bf16_bf16 v2, v0, -1.0, v2       ; encoding: [0x02,0x00,0x67,0xd6,0x00,0xe7,0x09,0x04]
+v_dot2_bf16_bf16 v2.l, v0, -1.0, v2.l
+// CHECK: v_dot2_bf16_bf16 v2.l, v0, -1.0, v2.l   ; encoding: [0x02,0x00,0x67,0xd6,0x00,0xe7,0x09,0x04]
 
-v_dot2_bf16_bf16 v2, v0, 0.5, v2
-// CHECK: v_dot2_bf16_bf16 v2, v0, 0.5, v2        ; encoding: [0x02,0x00,0x67,0xd6,0x00,0xe1,0x09,0x04]
+v_dot2_bf16_bf16 v2.l, v0, 0.5, v2.l
+// CHECK: v_dot2_bf16_bf16 v2.l, v0, 0.5, v2.l    ; encoding: [0x02,0x00,0x67,0xd6,0x00,0xe1,0x09,0x04]
 
-v_dot2_bf16_bf16 v2, v0, -0.5, v2
-// CHECK: v_dot2_bf16_bf16 v2, v0, -0.5, v2       ; encoding: [0x02,0x00,0x67,0xd6,0x00,0xe3,0x09,0x04]
+v_dot2_bf16_bf16 v2.l, v0, -0.5, v2.l
+// CHECK: v_dot2_bf16_bf16 v2.l, v0, -0.5, v2.l   ; encoding: [0x02,0x00,0x67,0xd6,0x00,0xe3,0x09,0x04]
 
-v_dot2_bf16_bf16 v2, v0, 2.0, v2
-// CHECK: v_dot2_bf16_bf16 v2, v0, 2.0, v2        ; encoding: [0x02,0x00,0x67,0xd6,0x00,0xe9,0x09,0x04]
+v_dot2_bf16_bf16 v2.l, v0, 2.0, v2.l
+// CHECK: v_dot2_bf16_bf16 v2.l, v0, 2.0, v2.l    ; encoding: [0x02,0x00,0x67,0xd6,0x00,0xe9,0x09,0x04]
 
-v_dot2_bf16_bf16 v2, v0, -2.0, v2
-// CHECK: v_dot2_bf16_bf16 v2, v0, -2.0, v2       ; encoding: [0x02,0x00,0x67,0xd6,0x00,0xeb,0x09,0x04]
+v_dot2_bf16_bf16 v2.l, v0, -2.0, v2.l
+// CHECK: v_dot2_bf16_bf16 v2.l, v0, -2.0, v2.l   ; encoding: [0x02,0x00,0x67,0xd6,0x00,0xeb,0x09,0x04]
 
-v_dot2_bf16_bf16 v2, v0, 4.0, v2
-// CHECK: v_dot2_bf16_bf16 v2, v0, 4.0, v2        ; encoding: [0x02,0x00,0x67,0xd6,0x00,0xed,0x09,0x04]
+v_dot2_bf16_bf16 v2.l, v0, 4.0, v2.l
+// CHECK: v_dot2_bf16_bf16 v2.l, v0, 4.0, v2.l    ; encoding: [0x02,0x00,0x67,0xd6,0x00,0xed,0x09,0x04]
 
-v_dot2_bf16_bf16 v2, v0, -4.0, v2
-// CHECK: v_dot2_bf16_bf16 v2, v0, -4.0, v2       ; encoding: [0x02,0x00,0x67,0xd6,0x00,0xef,0x09,0x04]
+v_dot2_bf16_bf16 v2.l, v0, -4.0, v2.l
+// CHECK: v_dot2_bf16_bf16 v2.l, v0, -4.0, v2.l   ; encoding: [0x02,0x00,0x67,0xd6,0x00,0xef,0x09,0x04]
 
 // Check 1/(2*pi) rounded value and ideomatic fp32 0.15915494 value
 // which cannot be accurately represented in bf16.
 
-v_dot2_bf16_bf16 v2, v0, 0.158203125, v2
-// CHECK: v_dot2_bf16_bf16 v2, v0, 0.15915494, v2 ; encoding: [0x02,0x00,0x67,0xd6,0x00,0xf1,0x09,0x04]
+v_dot2_bf16_bf16 v2.l, v0, 0.158203125, v2.l
+// CHECK: v_dot2_bf16_bf16 v2.l, v0, 0.15915494, v2.l ; encoding: [0x02,0x00,0x67,0xd6,0x00,0xf1,0x09,0x04]
 
-v_dot2_bf16_bf16 v2, v0, 0.15915494, v2
-// CHECK: v_dot2_bf16_bf16 v2, v0, 0.15915494, v2 ; encoding: [0x02,0x00,0x67,0xd6,0x00,0xf1,0x09,0x04]
+v_dot2_bf16_bf16 v2.l, v0, 0.15915494, v2.l
+// CHECK: v_dot2_bf16_bf16 v2.l, v0, 0.15915494, v2.l ; encoding: [0x02,0x00,0x67,0xd6,0x00,0xf1,0x09,0x04]
 
-v_dot2_bf16_bf16 v2, v0, 0x3e22, v2
-// CHECK: v_dot2_bf16_bf16 v2, v0, 0.15915494, v2 ; encoding: [0x02,0x00,0x67,0xd6,0x00,0xf1,0x09,0x04]
+v_dot2_bf16_bf16 v2.l, v0, 0x3e22, v2.l
+// CHECK: v_dot2_bf16_bf16 v2.l, v0, 0.15915494, v2.l ; encoding: [0x02,0x00,0x67,0xd6,0x00,0xf1,0x09,0x04]
 
-v_dot2_bf16_bf16 v2, v0, v2, 0.15915494
-// CHECK: v_dot2_bf16_bf16 v2, v0, v2, 0.15915494 ; encoding: [0x02,0x00,0x67,0xd6,0x00,0x05,0xe2,0x03]
+v_dot2_bf16_bf16 v2.l, v0, v2, 0.15915494
+// CHECK: v_dot2_bf16_bf16 v2.l, v0, v2, 0.15915494 ; encoding: [0x02,0x00,0x67,0xd6,0x00,0x05,0xe2,0x03]
 
 v_dot2_f32_bf16 v2, v1, 0, v2
 // CHECK: v_dot2_f32_bf16 v2, v1, 0, v2           ; encoding: [0x02,0x40,0x1a,0xcc,0x01,0x01,0x09,0x1c]
diff --git a/llvm/test/MC/AMDGPU/gfx11-promotions-fake16.s b/llvm/test/MC/AMDGPU/gfx11-promotions-fake16.s
new file mode 100644
index 0000000000000..95a52ffe103fa
--- /dev/null
+++ b/llvm/test/MC/AMDGPU/gfx11-promotions-fake16.s
@@ -0,0 +1,353 @@
+// RUN: llvm-mc -triple=amdgcn -show-encoding -mcpu=gfx1100 -mattr=+wavefrontsize32,-real-true16 %s | FileCheck --check-prefix=GFX11 %s
+
+// Check opcode promotions and forced suffices.
+// 1. When a suffix is optional, check that it may be omitted.
+// 2. When a suffix is optional, check that it may be specified w/o any effect.
+// 3. When a suffix is required, check that specifying it enforces opcode promotion.
+// 4. When a suffix is required, check that omitting the suffix results in a different encoding.
+
+//===----------------------------------------------------------------------===//
+// VOP1.
+//===----------------------------------------------------------------------===//
+
+v_mov_b32 v0, v1
+// GFX11: v_mov_b32_e32 v0, v1                    ; encoding: [0x01,0x03,0x00,0x7e]
+
+v_mov_b32_e32 v0, v1
+// GFX11: v_mov_b32_e32 v0, v1                    ; encoding: [0x01,0x03,0x00,0x7e]
+
+//===----------------------------------------------------------------------===//
+// VOP2.
+//===----------------------------------------------------------------------===//
+
+v_add_f16 v5, v1, v2
+// GFX11: v_add_f16_e32 v5, v1, v2                ; encoding: [0x01,0x05,0x0a,0x64]
+
+v_add_f16_e32 v5, v1, v2
+// GFX11: v_add_f16_e32 v5, v1, v2                ; encoding: [0x01,0x05,0x0a,0x64]
+
+//===----------------------------------------------------------------------===//
+// VOPC.
+//===----------------------------------------------------------------------===//
+
+v_cmp_lt_f32 vcc_lo, v1, v2
+// GFX11: v_cmp_lt_f32_e32 vcc_lo, v1, v2         ; encoding: [0x01,0x05,0x22,0x7c]
+
+v_cmp_lt_f32_e32 vcc_lo, v1, v2
+// GFX11: v_cmp_lt_f32_e32 vcc_lo, v1, v2         ; encoding: [0x01,0x05,0x22,0x7c]
+
+//===----------------------------------------------------------------------===//
+// VOPCX.
+//===----------------------------------------------------------------------===//
+
+v_cmpx_class_f16 v1, v2
+// GFX11: v_cmpx_class_f16_e32 v1, v2             ; encoding: [0x01,0x05,0xfa,0x7d]
+
+v_cmpx_class_f16_e32 v1, v2
+// GFX11: v_cmpx_class_f16_e32 v1, v2             ; encoding: [0x01,0x05,0xfa,0x7d]
+
+//===----------------------------------------------------------------------===//
+// VOP1.DPP8.
+//===----------------------------------------------------------------------===//
+
+v_bfrev_b32 v5, v1 dpp8:[7,6,5,4,3,2,1,0]
+// GFX11: v_bfrev_b32_dpp v5, v1 dpp8:[7,6,5,4,3,2,1,0] ; encoding: [0xe9,0x70,0x0a,0x7e,0x01,0x77,0x39,0x05]
+
+v_bfrev_b32_dpp v5, v1 dpp8:[7,6,5,4,3,2,1,0]
+// GFX11: v_bfrev_b32_dpp v5, v1 dpp8:[7,6,5,4,3,2,1,0] ; encoding: [0xe9,0x70,0x0a,0x7e,0x01,0x77,0x39,0x05]
+
+//===----------------------------------------------------------------------===//
+// VOP1.DPP16.
+//===----------------------------------------------------------------------===//
+
+v_bfrev_b32 v5, v1 quad_perm:[3,2,1,0]
+// GFX11: v_bfrev_b32_dpp v5, v1 quad_perm:[3,2,1,0] row_mask:0xf bank_mask:0xf ; encoding: [0xfa,0x70,0x0a,0x7e,0x01,0x1b,0x00,0xff]
+
+v_bfrev_b32_dpp v5, v1 quad_perm:[3,2,1,0]
+// GFX11: v_bfrev_b32_dpp v5, v1 quad_perm:[3,2,1,0] row_mask:0xf bank_mask:0xf ; encoding: [0xfa,0x70,0x0a,0x7e,0x01,0x1b,0x00,0xff]
+
+//===----------------------------------------------------------------------===//
+// VOP2.DPP8.
+//===----------------------------------------------------------------------===//
+
+v_add_f16 v5, v1, v2 dpp8:[7,6,5,4,3,2,1,0]
+// GFX11: v_add_f16_dpp v5, v1, v2 dpp8:[7,6,5,4,3,2,1,0] ; encoding: [0xe9,0x04,0x0a,0x64,0x01,0x77,0x39,0x05]
+
+v_add_f16_dpp v5, v1, v2 dpp8:[7,6,5,4,3,2,1,0]
+// GFX11: v_add_f16_dpp v5, v1, v2 dpp8:[7,6,5,4,3,2,1,0] ; encoding: [0xe9,0x04,0x0a,0x64,0x01,0x77,0x39,0x05]
+
+//===----------------------------------------------------------------------===//
+// VOP2.DPP16.
+//===----------------------------------------------------------------------===//
+
+v_add_f16 v5, v1, v2 quad_perm:[3,2,1,0]
+// GFX11: v_add_f16_dpp v5, v1, v2 quad_perm:[3,2,1,0] row_mask:0xf bank_mask:0xf ; encoding: [0xfa,0x04,0x0a,0x64,0x01,0x1b,0x00,0xff]
+
+v_add_f16_dpp v5, v1, v2 quad_perm:[3,2,1,0]
+// GFX11: v_add_f16_dpp v5, v1, v2 quad_perm:[3,2,1,0] row_mask:0xf bank_mask:0xf ; encoding: [0xfa,0x04,0x0a,0x64,0x01,0x1b,0x00,0xff]
+
+//===----------------------------------------------------------------------===//
+// VOPC.DPP8.
+//===----------------------------------------------------------------------===//
+
+v_cmp_le_u16 v1, v2 dpp8:[7,7,7,3,4,4,6,7] fi:1
+// GFX11: v_cmp_le_u16 vcc_lo, v1, v2 dpp8:[7,7,7,3,4,4,6,7] fi:1 ; encoding: [0xea,0x04,0x76,0x7c,0x01,0xff,0x47,0xfa]
+
+v_cmp_le_u16_dpp v1, v2 dpp8:[7,7,7,3,4,4,6,7] fi:1
+// GFX11: v_cmp_le_u16 vcc_lo, v1, v2 dpp8:[7,7,7,3,4,4,6,7] fi:1 ; encoding: [0xea,0x04,0x76,0x7c,0x01,0xff,0x47,0xfa]
+
+//===----------------------------------------------------------------------===//
+// VOPC.DPP16.
+//===----------------------------------------------------------------------===//
+
+v_cmp_gt_u16 v1, v2 row_shl:0x7 row_mask:0x0 bank_mask:0x0 fi:1
+// GFX11: v_cmp_gt_u16 vcc_lo, v1, v2 row_shl:7 row_mask:0x0 bank_mask:0x0 fi:1 ; encoding: [0xfa,0x04,0x78,0x7c,0x01,0x07,0x05,0x00]
+
+v_cmp_gt_u16_dpp v1, v2 row_shl:0x7 row_mask:0x0 bank_mask:0x0 fi:1
+// GFX11: v_cmp_gt_u16 vcc_lo, v1, v2 row_shl:7 row_mask:0x0 bank_mask:0x0 fi:1 ; encoding: [0xfa,0x04,0x78,0x7c,0x01,0x07,0x05,0x00]
+
+//===----------------------------------------------------------------------===//
+// VOPCX.DPP8.
+//===----------------------------------------------------------------------===//
+
+v_cmpx_class_f16 v1, v2 dpp8:[7,6,5,4,3,2,1,0]
+// GFX11: v_cmpx_class_f16 v1, v2 dpp8:[7,6,5,4,3,2,1,0] ; encoding: [0xe9,0x04,0xfa,0x7d,0x01,0x77,0x39,0x05]
+
+v_cmpx_class_f16_dpp v1, v2 dpp8:[7,6,5,4,3,2,1,0]
+// GFX11: v_cmpx_class_f16 v1, v2 dpp8:[7,6,5,4,3,2,1,0] ; encoding: [0xe9,0x04,0xfa,0x7d,0x01,0x77,0x39,0x05]
+
+//===----------------------------------------------------------------------===//
+// VOPCX.DPP16.
+//===----------------------------------------------------------------------===//
+
+v_cmpx_class_f16 v1, v2 quad_perm:[3,2,1,0]
+// GFX11: v_cmpx_class_f16 v1, v2 quad_perm:[3,2,1,0] row_mask:0xf bank_mask:0xf ; encoding: [0xfa,0x04,0xfa,0x7d,0x01,0x1b,0x00,0xff]
+
+v_cmpx_class_f16_dpp v1, v2 quad_perm:[3,2,1,0]
+// GFX11: v_cmpx_class_f16 v1, v2 quad_perm:[3,2,1,0] row_mask:0xf bank_mask:0xf ; encoding: [0xfa,0x04,0xfa,0x7d,0x01,0x1b,0x00,0xff]
+
+//===----------------------------------------------------------------------===//
+// VOP1 -> VOP3.
+//===----------------------------------------------------------------------===//
+
+v_sin_f32 v5, 0.5 mul:2
+// GFX11: v_sin_f32_e64 v5, 0.5 mul:2             ; encoding: [0x05,0x00,0xb5,0xd5,0xf0,0x00,0x00,0x08]
+
+v_sin_f32_e64 v5, 0.5 mul:2
+// GFX11: v_sin_f32_e64 v5, 0.5 mul:2             ; encoding: [0x05,0x00,0xb5,0xd5,0xf0,0x00,0x00,0x08]
+
+v_sin_f32_e64 v5, v1
+// GFX11: v_sin_f32_e64 v5, v1                    ; encoding: [0x05,0x00,0xb5,0xd5,0x01,0x01,0x00,0x00]
+
+v_sin_f32 v5, v1
+// GFX11: v_sin_f32_e32 v5, v1                    ; encoding: [0x01,0x6b,0x0a,0x7e]
+
+//===----------------------------------------------------------------------===//
+// VOP2 -> VOP3.
+//===----------------------------------------------------------------------===//
+
+v_add_f32 v5, v1, -v2
+// GFX11: v_add_f32_e64 v5, v1, -v2               ; encoding: [0x05,0x00,0x03,0xd5,0x01,0x05,0x02,0x40]
+
+v_add_f32_e64 v5, v1, -v2
+// GFX11: v_add_f32_e64 v5, v1, -v2               ; encoding: [0x05,0x00,0x03,0xd5,0x01,0x05,0x02,0x40]
+
+v_add_f32_e64 v5, v1, v2
+// GFX11: v_add_f32_e64 v5, v1, v2                ; encoding: [0x05,0x00,0x03,0xd5,0x01,0x05,0x02,0x00]
+
+v_add_f32 v5, v1, v2
+// GFX11: v_add_f32_e32 v5, v1, v2                ; encoding: [0x01,0x05,0x0a,0x06]
+
+//===----------------------------------------------------------------------===//
+// VOPC -> VOP3.
+//===----------------------------------------------------------------------===//
+
+v_cmp_f_f32 s10, -v1, v2
+// GFX11: v_cmp_f_f32_e64 s10, -v1, v2            ; encoding: [0x0a,0x00,0x10,0xd4,0x01,0x05,0x02,0x20]
+
+v_cmp_f_f32_e64 s10, -v1, v2
+// GFX11: v_cmp_f_f32_e64 s10, -v1, v2            ; encoding: [0x0a,0x00,0x10,0xd4,0x01,0x05,0x02,0x20]
+
+v_cmp_f_f32_e64 vcc_lo, v1, v2
+// GFX11: v_cmp_f_f32_e64 vcc_lo, v1, v2          ; encoding: [0x6a,0x00,0x10,0xd4,0x01,0x05,0x02,0x00]
+
+v_cmp_f_f32 vcc_lo, v1, v2
+// GFX11: v_cmp_f_f32_e32 vcc_lo, v1, v2          ; encoding: [0x01,0x05,0x20,0x7c]
+
+//===----------------------------------------------------------------------===//
+// VOPCX -> VOP3.
+//===----------------------------------------------------------------------===//
+
+v_cmpx_f_f32 -v1, v2
+// GFX11: v_cmpx_f_f32_e64 -v1, v2                ; encoding: [0x7e,0x00,0x90,0xd4,0x01,0x05,0x02,0x20]
+
+v_cmpx_f_f32_e64 -v1, v2
+// GFX11: v_cmpx_f_f32_e64 -v1, v2                ; encoding: [0x7e,0x00,0x90,0xd4,0x01,0x05,0x02,0x20]
+
+v_cmpx_f_f32_e64 v1, v2
+// GFX11: v_cmpx_f_f32_e64 v1, v2                 ; encoding: [0x7e,0x00,0x90,0xd4,0x01,0x05,0x02,0x00]
+
+v_cmpx_f_f32 v1, v2
+// GFX11: v_cmpx_f_f32_e...
[truncated]

@llvmbot
Copy link
Member

llvmbot commented Apr 15, 2025

@llvm/pr-subscribers-mc

Author: Brox Chen (broxigarchen)

Changes

This is another NFC patch.

Update mc test for a few true16 instructions by duplicating the file to fake16 versions and udpate mattr flag with +/-real-true16

Patch is 46.95 KiB, truncated to 20.00 KiB below, full version: https://github.com/llvm/llvm-project/pull/135816.diff

8 Files Affected:

  • (added) llvm/test/MC/AMDGPU/bf16_imm-fake16.s (+114)
  • (modified) llvm/test/MC/AMDGPU/bf16_imm.s (+32-32)
  • (added) llvm/test/MC/AMDGPU/gfx11-promotions-fake16.s (+353)
  • (modified) llvm/test/MC/AMDGPU/gfx11-promotions.s (+33-33)
  • (added) llvm/test/MC/AMDGPU/gfx1150_asm_features-fake16.s (+48)
  • (modified) llvm/test/MC/AMDGPU/gfx1150_asm_features.s (+10-10)
  • (added) llvm/test/MC/AMDGPU/gfx11_asm_t16.s (+59)
  • (modified) llvm/test/MC/AMDGPU/gfx11_asm_vop3_alias.s (+4-4)
diff --git a/llvm/test/MC/AMDGPU/bf16_imm-fake16.s b/llvm/test/MC/AMDGPU/bf16_imm-fake16.s
new file mode 100644
index 0000000000000..ee697bee6ab2d
--- /dev/null
+++ b/llvm/test/MC/AMDGPU/bf16_imm-fake16.s
@@ -0,0 +1,114 @@
+// NOTE: Assertions have been autogenerated by utils/update_mc_test_checks.py UTC_ARGS: --unique --version 5
+// RUN: llvm-mc -triple=amdgcn -mcpu=gfx1100 -mattr=-real-true16 -show-encoding %s | FileCheck %s
+// RUN: llvm-mc -triple=amdgcn -mcpu=gfx1200 -mattr=-real-true16 -show-encoding %s | FileCheck %s
+
+v_dot2_bf16_bf16 v5, v1, v2, 100.0
+// CHECK: v_dot2_bf16_bf16 v5, v1, v2, 0x42c8     ; encoding: [0x05,0x00,0x67,0xd6,0x01,0x05,0xfe,0x03,0xc8,0x42,0x00,0x00]
+
+v_dot2_bf16_bf16 v2, v0, 1.0, v2
+// CHECK: v_dot2_bf16_bf16 v2, v0, 1.0, v2        ; encoding: [0x02,0x00,0x67,0xd6,0x00,0xe5,0x09,0x04]
+
+v_dot2_bf16_bf16 v2, 1.0, v0, v2
+// CHECK: v_dot2_bf16_bf16 v2, 1.0, v0, v2        ; encoding: [0x02,0x00,0x67,0xd6,0xf2,0x00,0x0a,0x04]
+
+v_dot2_bf16_bf16 v5, v1, v2, 1.0
+// CHECK: v_dot2_bf16_bf16 v5, v1, v2, 1.0        ; encoding: [0x05,0x00,0x67,0xd6,0x01,0x05,0xca,0x03]
+
+v_dot2_bf16_bf16 v2, v0, -1.0, v2
+// CHECK: v_dot2_bf16_bf16 v2, v0, -1.0, v2       ; encoding: [0x02,0x00,0x67,0xd6,0x00,0xe7,0x09,0x04]
+
+v_dot2_bf16_bf16 v2, v0, 0.5, v2
+// CHECK: v_dot2_bf16_bf16 v2, v0, 0.5, v2        ; encoding: [0x02,0x00,0x67,0xd6,0x00,0xe1,0x09,0x04]
+
+v_dot2_bf16_bf16 v2, v0, -0.5, v2
+// CHECK: v_dot2_bf16_bf16 v2, v0, -0.5, v2       ; encoding: [0x02,0x00,0x67,0xd6,0x00,0xe3,0x09,0x04]
+
+v_dot2_bf16_bf16 v2, v0, 2.0, v2
+// CHECK: v_dot2_bf16_bf16 v2, v0, 2.0, v2        ; encoding: [0x02,0x00,0x67,0xd6,0x00,0xe9,0x09,0x04]
+
+v_dot2_bf16_bf16 v2, v0, -2.0, v2
+// CHECK: v_dot2_bf16_bf16 v2, v0, -2.0, v2       ; encoding: [0x02,0x00,0x67,0xd6,0x00,0xeb,0x09,0x04]
+
+v_dot2_bf16_bf16 v2, v0, 4.0, v2
+// CHECK: v_dot2_bf16_bf16 v2, v0, 4.0, v2        ; encoding: [0x02,0x00,0x67,0xd6,0x00,0xed,0x09,0x04]
+
+v_dot2_bf16_bf16 v2, v0, -4.0, v2
+// CHECK: v_dot2_bf16_bf16 v2, v0, -4.0, v2       ; encoding: [0x02,0x00,0x67,0xd6,0x00,0xef,0x09,0x04]
+
+// Check 1/(2*pi) rounded value and ideomatic fp32 0.15915494 value
+// which cannot be accurately represented in bf16.
+
+v_dot2_bf16_bf16 v2, v0, 0.158203125, v2
+// CHECK: v_dot2_bf16_bf16 v2, v0, 0.15915494, v2 ; encoding: [0x02,0x00,0x67,0xd6,0x00,0xf1,0x09,0x04]
+
+v_dot2_bf16_bf16 v2, v0, 0.15915494, v2
+// CHECK: v_dot2_bf16_bf16 v2, v0, 0.15915494, v2 ; encoding: [0x02,0x00,0x67,0xd6,0x00,0xf1,0x09,0x04]
+
+v_dot2_bf16_bf16 v2, v0, 0x3e22, v2
+// CHECK: v_dot2_bf16_bf16 v2, v0, 0.15915494, v2 ; encoding: [0x02,0x00,0x67,0xd6,0x00,0xf1,0x09,0x04]
+
+v_dot2_bf16_bf16 v2, v0, v2, 0.15915494
+// CHECK: v_dot2_bf16_bf16 v2, v0, v2, 0.15915494 ; encoding: [0x02,0x00,0x67,0xd6,0x00,0x05,0xe2,0x03]
+
+v_dot2_f32_bf16 v2, v1, 0, v2
+// CHECK: v_dot2_f32_bf16 v2, v1, 0, v2           ; encoding: [0x02,0x40,0x1a,0xcc,0x01,0x01,0x09,0x1c]
+
+v_dot2_f32_bf16 v2, v1, 0.5, v2
+// CHECK: v_dot2_f32_bf16 v2, v1, 0.5, v2         ; encoding: [0x02,0x40,0x1a,0xcc,0x01,0xe1,0x09,0x1c]
+
+v_dot2_f32_bf16 v2, v1, -0.5, v2
+// CHECK: v_dot2_f32_bf16 v2, v1, -0.5, v2        ; encoding: [0x02,0x40,0x1a,0xcc,0x01,0xe3,0x09,0x1c]
+
+v_dot2_f32_bf16 v2, v1, 1.0, v2
+// CHECK: v_dot2_f32_bf16 v2, v1, 1.0, v2         ; encoding: [0x02,0x40,0x1a,0xcc,0x01,0xe5,0x09,0x1c]
+
+v_dot2_f32_bf16 v2, v1, -1.0, v2
+// CHECK: v_dot2_f32_bf16 v2, v1, -1.0, v2        ; encoding: [0x02,0x40,0x1a,0xcc,0x01,0xe7,0x09,0x1c]
+
+v_dot2_f32_bf16 v2, v1, 2.0, v2
+// CHECK: v_dot2_f32_bf16 v2, v1, 2.0, v2         ; encoding: [0x02,0x40,0x1a,0xcc,0x01,0xe9,0x09,0x1c]
+
+v_dot2_f32_bf16 v2, v1, -2.0, v2
+// CHECK: v_dot2_f32_bf16 v2, v1, -2.0, v2        ; encoding: [0x02,0x40,0x1a,0xcc,0x01,0xeb,0x09,0x1c]
+
+v_dot2_f32_bf16 v2, v1, 4.0, v2
+// CHECK: v_dot2_f32_bf16 v2, v1, 4.0, v2         ; encoding: [0x02,0x40,0x1a,0xcc,0x01,0xed,0x09,0x1c]
+
+v_dot2_f32_bf16 v2, v1, -4.0, v2
+// CHECK: v_dot2_f32_bf16 v2, v1, -4.0, v2        ; encoding: [0x02,0x40,0x1a,0xcc,0x01,0xef,0x09,0x1c]
+
+v_dot2_f32_bf16 v2, v1, 0.15915494, v2
+// CHECK: v_dot2_f32_bf16 v2, v1, 0.15915494, v2  ; encoding: [0x02,0x40,0x1a,0xcc,0x01,0xf1,0x09,0x1c]
+
+v_dot2_f32_bf16 v2, v1, 0x3e22, v2
+// CHECK: v_dot2_f32_bf16 v2, v1, 0.15915494, v2  ; encoding: [0x02,0x40,0x1a,0xcc,0x01,0xf1,0x09,0x1c]
+
+v_dot2_f32_bf16 v2, 0.5, v1, v2
+// CHECK: v_dot2_f32_bf16 v2, 0.5, v1, v2         ; encoding: [0x02,0x40,0x1a,0xcc,0xf0,0x02,0x0a,0x1c]
+
+v_dot2_f32_bf16 v2, -0.5, v1, v2
+// CHECK: v_dot2_f32_bf16 v2, -0.5, v1, v2        ; encoding: [0x02,0x40,0x1a,0xcc,0xf1,0x02,0x0a,0x1c]
+
+v_dot2_f32_bf16 v2, 1.0, v1, v2
+// CHECK: v_dot2_f32_bf16 v2, 1.0, v1, v2         ; encoding: [0x02,0x40,0x1a,0xcc,0xf2,0x02,0x0a,0x1c]
+
+v_dot2_f32_bf16 v2, -1.0, v1, v2
+// CHECK: v_dot2_f32_bf16 v2, -1.0, v1, v2        ; encoding: [0x02,0x40,0x1a,0xcc,0xf3,0x02,0x0a,0x1c]
+
+v_dot2_f32_bf16 v2, 2.0, v1, v2
+// CHECK: v_dot2_f32_bf16 v2, 2.0, v1, v2         ; encoding: [0x02,0x40,0x1a,0xcc,0xf4,0x02,0x0a,0x1c]
+
+v_dot2_f32_bf16 v2, -2.0, v1, v2
+// CHECK: v_dot2_f32_bf16 v2, -2.0, v1, v2        ; encoding: [0x02,0x40,0x1a,0xcc,0xf5,0x02,0x0a,0x1c]
+
+v_dot2_f32_bf16 v2, 4.0, v1, v2
+// CHECK: v_dot2_f32_bf16 v2, 4.0, v1, v2         ; encoding: [0x02,0x40,0x1a,0xcc,0xf6,0x02,0x0a,0x1c]
+
+v_dot2_f32_bf16 v2, -4.0, v1, v2
+// CHECK: v_dot2_f32_bf16 v2, -4.0, v1, v2        ; encoding: [0x02,0x40,0x1a,0xcc,0xf7,0x02,0x0a,0x1c]
+
+v_dot2_f32_bf16 v2, 100.0, v1, v2
+// CHECK: v_dot2_f32_bf16 v2, 0x42c8, v1, v2      ; encoding: [0x02,0x40,0x1a,0xcc,0xff,0x02,0x0a,0x1c,0xc8,0x42,0x00,0x00]
+
+v_dot2_f32_bf16 v2, v1, 100.0, v2
+// CHECK: v_dot2_f32_bf16 v2, v1, 0x42c8, v2      ; encoding: [0x02,0x40,0x1a,0xcc,0x01,0xff,0x09,0x1c,0xc8,0x42,0x00,0x00]
diff --git a/llvm/test/MC/AMDGPU/bf16_imm.s b/llvm/test/MC/AMDGPU/bf16_imm.s
index 7cf18103adfe5..d79649073aa89 100644
--- a/llvm/test/MC/AMDGPU/bf16_imm.s
+++ b/llvm/test/MC/AMDGPU/bf16_imm.s
@@ -1,54 +1,54 @@
 // NOTE: Assertions have been autogenerated by utils/update_mc_test_checks.py UTC_ARGS: --unique --version 5
-// RUN: llvm-mc -triple=amdgcn -mcpu=gfx1100 -show-encoding %s | FileCheck %s
-// RUN: llvm-mc -triple=amdgcn -mcpu=gfx1200 -show-encoding %s | FileCheck %s
+// RUN: llvm-mc -triple=amdgcn -mcpu=gfx1100 -mattr=+real-true16 -show-encoding %s | FileCheck %s
+// RUN: llvm-mc -triple=amdgcn -mcpu=gfx1200 -mattr=+real-true16 -show-encoding %s | FileCheck %s
 
-v_dot2_bf16_bf16 v5, v1, v2, 100.0
-// CHECK: v_dot2_bf16_bf16 v5, v1, v2, 0x42c8     ; encoding: [0x05,0x00,0x67,0xd6,0x01,0x05,0xfe,0x03,0xc8,0x42,0x00,0x00]
+v_dot2_bf16_bf16 v5.l, v1, v2, 100.0
+// CHECK: v_dot2_bf16_bf16 v5.l, v1, v2, 0x42c8   ; encoding: [0x05,0x00,0x67,0xd6,0x01,0x05,0xfe,0x03,0xc8,0x42,0x00,0x00]
 
-v_dot2_bf16_bf16 v2, v0, 1.0, v2
-// CHECK: v_dot2_bf16_bf16 v2, v0, 1.0, v2        ; encoding: [0x02,0x00,0x67,0xd6,0x00,0xe5,0x09,0x04]
+v_dot2_bf16_bf16 v2.l, v0, 1.0, v2.l
+// CHECK: v_dot2_bf16_bf16 v2.l, v0, 1.0, v2.l    ; encoding: [0x02,0x00,0x67,0xd6,0x00,0xe5,0x09,0x04]
 
-v_dot2_bf16_bf16 v2, 1.0, v0, v2
-// CHECK: v_dot2_bf16_bf16 v2, 1.0, v0, v2        ; encoding: [0x02,0x00,0x67,0xd6,0xf2,0x00,0x0a,0x04]
+v_dot2_bf16_bf16 v2.l, 1.0, v0, v2.l
+// CHECK: v_dot2_bf16_bf16 v2.l, 1.0, v0, v2.l    ; encoding: [0x02,0x00,0x67,0xd6,0xf2,0x00,0x0a,0x04]
 
-v_dot2_bf16_bf16 v5, v1, v2, 1.0
-// CHECK: v_dot2_bf16_bf16 v5, v1, v2, 1.0        ; encoding: [0x05,0x00,0x67,0xd6,0x01,0x05,0xca,0x03]
+v_dot2_bf16_bf16 v5.l, v1, v2, 1.0
+// CHECK: v_dot2_bf16_bf16 v5.l, v1, v2, 1.0      ; encoding: [0x05,0x00,0x67,0xd6,0x01,0x05,0xca,0x03]
 
-v_dot2_bf16_bf16 v2, v0, -1.0, v2
-// CHECK: v_dot2_bf16_bf16 v2, v0, -1.0, v2       ; encoding: [0x02,0x00,0x67,0xd6,0x00,0xe7,0x09,0x04]
+v_dot2_bf16_bf16 v2.l, v0, -1.0, v2.l
+// CHECK: v_dot2_bf16_bf16 v2.l, v0, -1.0, v2.l   ; encoding: [0x02,0x00,0x67,0xd6,0x00,0xe7,0x09,0x04]
 
-v_dot2_bf16_bf16 v2, v0, 0.5, v2
-// CHECK: v_dot2_bf16_bf16 v2, v0, 0.5, v2        ; encoding: [0x02,0x00,0x67,0xd6,0x00,0xe1,0x09,0x04]
+v_dot2_bf16_bf16 v2.l, v0, 0.5, v2.l
+// CHECK: v_dot2_bf16_bf16 v2.l, v0, 0.5, v2.l    ; encoding: [0x02,0x00,0x67,0xd6,0x00,0xe1,0x09,0x04]
 
-v_dot2_bf16_bf16 v2, v0, -0.5, v2
-// CHECK: v_dot2_bf16_bf16 v2, v0, -0.5, v2       ; encoding: [0x02,0x00,0x67,0xd6,0x00,0xe3,0x09,0x04]
+v_dot2_bf16_bf16 v2.l, v0, -0.5, v2.l
+// CHECK: v_dot2_bf16_bf16 v2.l, v0, -0.5, v2.l   ; encoding: [0x02,0x00,0x67,0xd6,0x00,0xe3,0x09,0x04]
 
-v_dot2_bf16_bf16 v2, v0, 2.0, v2
-// CHECK: v_dot2_bf16_bf16 v2, v0, 2.0, v2        ; encoding: [0x02,0x00,0x67,0xd6,0x00,0xe9,0x09,0x04]
+v_dot2_bf16_bf16 v2.l, v0, 2.0, v2.l
+// CHECK: v_dot2_bf16_bf16 v2.l, v0, 2.0, v2.l    ; encoding: [0x02,0x00,0x67,0xd6,0x00,0xe9,0x09,0x04]
 
-v_dot2_bf16_bf16 v2, v0, -2.0, v2
-// CHECK: v_dot2_bf16_bf16 v2, v0, -2.0, v2       ; encoding: [0x02,0x00,0x67,0xd6,0x00,0xeb,0x09,0x04]
+v_dot2_bf16_bf16 v2.l, v0, -2.0, v2.l
+// CHECK: v_dot2_bf16_bf16 v2.l, v0, -2.0, v2.l   ; encoding: [0x02,0x00,0x67,0xd6,0x00,0xeb,0x09,0x04]
 
-v_dot2_bf16_bf16 v2, v0, 4.0, v2
-// CHECK: v_dot2_bf16_bf16 v2, v0, 4.0, v2        ; encoding: [0x02,0x00,0x67,0xd6,0x00,0xed,0x09,0x04]
+v_dot2_bf16_bf16 v2.l, v0, 4.0, v2.l
+// CHECK: v_dot2_bf16_bf16 v2.l, v0, 4.0, v2.l    ; encoding: [0x02,0x00,0x67,0xd6,0x00,0xed,0x09,0x04]
 
-v_dot2_bf16_bf16 v2, v0, -4.0, v2
-// CHECK: v_dot2_bf16_bf16 v2, v0, -4.0, v2       ; encoding: [0x02,0x00,0x67,0xd6,0x00,0xef,0x09,0x04]
+v_dot2_bf16_bf16 v2.l, v0, -4.0, v2.l
+// CHECK: v_dot2_bf16_bf16 v2.l, v0, -4.0, v2.l   ; encoding: [0x02,0x00,0x67,0xd6,0x00,0xef,0x09,0x04]
 
 // Check 1/(2*pi) rounded value and ideomatic fp32 0.15915494 value
 // which cannot be accurately represented in bf16.
 
-v_dot2_bf16_bf16 v2, v0, 0.158203125, v2
-// CHECK: v_dot2_bf16_bf16 v2, v0, 0.15915494, v2 ; encoding: [0x02,0x00,0x67,0xd6,0x00,0xf1,0x09,0x04]
+v_dot2_bf16_bf16 v2.l, v0, 0.158203125, v2.l
+// CHECK: v_dot2_bf16_bf16 v2.l, v0, 0.15915494, v2.l ; encoding: [0x02,0x00,0x67,0xd6,0x00,0xf1,0x09,0x04]
 
-v_dot2_bf16_bf16 v2, v0, 0.15915494, v2
-// CHECK: v_dot2_bf16_bf16 v2, v0, 0.15915494, v2 ; encoding: [0x02,0x00,0x67,0xd6,0x00,0xf1,0x09,0x04]
+v_dot2_bf16_bf16 v2.l, v0, 0.15915494, v2.l
+// CHECK: v_dot2_bf16_bf16 v2.l, v0, 0.15915494, v2.l ; encoding: [0x02,0x00,0x67,0xd6,0x00,0xf1,0x09,0x04]
 
-v_dot2_bf16_bf16 v2, v0, 0x3e22, v2
-// CHECK: v_dot2_bf16_bf16 v2, v0, 0.15915494, v2 ; encoding: [0x02,0x00,0x67,0xd6,0x00,0xf1,0x09,0x04]
+v_dot2_bf16_bf16 v2.l, v0, 0x3e22, v2.l
+// CHECK: v_dot2_bf16_bf16 v2.l, v0, 0.15915494, v2.l ; encoding: [0x02,0x00,0x67,0xd6,0x00,0xf1,0x09,0x04]
 
-v_dot2_bf16_bf16 v2, v0, v2, 0.15915494
-// CHECK: v_dot2_bf16_bf16 v2, v0, v2, 0.15915494 ; encoding: [0x02,0x00,0x67,0xd6,0x00,0x05,0xe2,0x03]
+v_dot2_bf16_bf16 v2.l, v0, v2, 0.15915494
+// CHECK: v_dot2_bf16_bf16 v2.l, v0, v2, 0.15915494 ; encoding: [0x02,0x00,0x67,0xd6,0x00,0x05,0xe2,0x03]
 
 v_dot2_f32_bf16 v2, v1, 0, v2
 // CHECK: v_dot2_f32_bf16 v2, v1, 0, v2           ; encoding: [0x02,0x40,0x1a,0xcc,0x01,0x01,0x09,0x1c]
diff --git a/llvm/test/MC/AMDGPU/gfx11-promotions-fake16.s b/llvm/test/MC/AMDGPU/gfx11-promotions-fake16.s
new file mode 100644
index 0000000000000..95a52ffe103fa
--- /dev/null
+++ b/llvm/test/MC/AMDGPU/gfx11-promotions-fake16.s
@@ -0,0 +1,353 @@
+// RUN: llvm-mc -triple=amdgcn -show-encoding -mcpu=gfx1100 -mattr=+wavefrontsize32,-real-true16 %s | FileCheck --check-prefix=GFX11 %s
+
+// Check opcode promotions and forced suffices.
+// 1. When a suffix is optional, check that it may be omitted.
+// 2. When a suffix is optional, check that it may be specified w/o any effect.
+// 3. When a suffix is required, check that specifying it enforces opcode promotion.
+// 4. When a suffix is required, check that omitting the suffix results in a different encoding.
+
+//===----------------------------------------------------------------------===//
+// VOP1.
+//===----------------------------------------------------------------------===//
+
+v_mov_b32 v0, v1
+// GFX11: v_mov_b32_e32 v0, v1                    ; encoding: [0x01,0x03,0x00,0x7e]
+
+v_mov_b32_e32 v0, v1
+// GFX11: v_mov_b32_e32 v0, v1                    ; encoding: [0x01,0x03,0x00,0x7e]
+
+//===----------------------------------------------------------------------===//
+// VOP2.
+//===----------------------------------------------------------------------===//
+
+v_add_f16 v5, v1, v2
+// GFX11: v_add_f16_e32 v5, v1, v2                ; encoding: [0x01,0x05,0x0a,0x64]
+
+v_add_f16_e32 v5, v1, v2
+// GFX11: v_add_f16_e32 v5, v1, v2                ; encoding: [0x01,0x05,0x0a,0x64]
+
+//===----------------------------------------------------------------------===//
+// VOPC.
+//===----------------------------------------------------------------------===//
+
+v_cmp_lt_f32 vcc_lo, v1, v2
+// GFX11: v_cmp_lt_f32_e32 vcc_lo, v1, v2         ; encoding: [0x01,0x05,0x22,0x7c]
+
+v_cmp_lt_f32_e32 vcc_lo, v1, v2
+// GFX11: v_cmp_lt_f32_e32 vcc_lo, v1, v2         ; encoding: [0x01,0x05,0x22,0x7c]
+
+//===----------------------------------------------------------------------===//
+// VOPCX.
+//===----------------------------------------------------------------------===//
+
+v_cmpx_class_f16 v1, v2
+// GFX11: v_cmpx_class_f16_e32 v1, v2             ; encoding: [0x01,0x05,0xfa,0x7d]
+
+v_cmpx_class_f16_e32 v1, v2
+// GFX11: v_cmpx_class_f16_e32 v1, v2             ; encoding: [0x01,0x05,0xfa,0x7d]
+
+//===----------------------------------------------------------------------===//
+// VOP1.DPP8.
+//===----------------------------------------------------------------------===//
+
+v_bfrev_b32 v5, v1 dpp8:[7,6,5,4,3,2,1,0]
+// GFX11: v_bfrev_b32_dpp v5, v1 dpp8:[7,6,5,4,3,2,1,0] ; encoding: [0xe9,0x70,0x0a,0x7e,0x01,0x77,0x39,0x05]
+
+v_bfrev_b32_dpp v5, v1 dpp8:[7,6,5,4,3,2,1,0]
+// GFX11: v_bfrev_b32_dpp v5, v1 dpp8:[7,6,5,4,3,2,1,0] ; encoding: [0xe9,0x70,0x0a,0x7e,0x01,0x77,0x39,0x05]
+
+//===----------------------------------------------------------------------===//
+// VOP1.DPP16.
+//===----------------------------------------------------------------------===//
+
+v_bfrev_b32 v5, v1 quad_perm:[3,2,1,0]
+// GFX11: v_bfrev_b32_dpp v5, v1 quad_perm:[3,2,1,0] row_mask:0xf bank_mask:0xf ; encoding: [0xfa,0x70,0x0a,0x7e,0x01,0x1b,0x00,0xff]
+
+v_bfrev_b32_dpp v5, v1 quad_perm:[3,2,1,0]
+// GFX11: v_bfrev_b32_dpp v5, v1 quad_perm:[3,2,1,0] row_mask:0xf bank_mask:0xf ; encoding: [0xfa,0x70,0x0a,0x7e,0x01,0x1b,0x00,0xff]
+
+//===----------------------------------------------------------------------===//
+// VOP2.DPP8.
+//===----------------------------------------------------------------------===//
+
+v_add_f16 v5, v1, v2 dpp8:[7,6,5,4,3,2,1,0]
+// GFX11: v_add_f16_dpp v5, v1, v2 dpp8:[7,6,5,4,3,2,1,0] ; encoding: [0xe9,0x04,0x0a,0x64,0x01,0x77,0x39,0x05]
+
+v_add_f16_dpp v5, v1, v2 dpp8:[7,6,5,4,3,2,1,0]
+// GFX11: v_add_f16_dpp v5, v1, v2 dpp8:[7,6,5,4,3,2,1,0] ; encoding: [0xe9,0x04,0x0a,0x64,0x01,0x77,0x39,0x05]
+
+//===----------------------------------------------------------------------===//
+// VOP2.DPP16.
+//===----------------------------------------------------------------------===//
+
+v_add_f16 v5, v1, v2 quad_perm:[3,2,1,0]
+// GFX11: v_add_f16_dpp v5, v1, v2 quad_perm:[3,2,1,0] row_mask:0xf bank_mask:0xf ; encoding: [0xfa,0x04,0x0a,0x64,0x01,0x1b,0x00,0xff]
+
+v_add_f16_dpp v5, v1, v2 quad_perm:[3,2,1,0]
+// GFX11: v_add_f16_dpp v5, v1, v2 quad_perm:[3,2,1,0] row_mask:0xf bank_mask:0xf ; encoding: [0xfa,0x04,0x0a,0x64,0x01,0x1b,0x00,0xff]
+
+//===----------------------------------------------------------------------===//
+// VOPC.DPP8.
+//===----------------------------------------------------------------------===//
+
+v_cmp_le_u16 v1, v2 dpp8:[7,7,7,3,4,4,6,7] fi:1
+// GFX11: v_cmp_le_u16 vcc_lo, v1, v2 dpp8:[7,7,7,3,4,4,6,7] fi:1 ; encoding: [0xea,0x04,0x76,0x7c,0x01,0xff,0x47,0xfa]
+
+v_cmp_le_u16_dpp v1, v2 dpp8:[7,7,7,3,4,4,6,7] fi:1
+// GFX11: v_cmp_le_u16 vcc_lo, v1, v2 dpp8:[7,7,7,3,4,4,6,7] fi:1 ; encoding: [0xea,0x04,0x76,0x7c,0x01,0xff,0x47,0xfa]
+
+//===----------------------------------------------------------------------===//
+// VOPC.DPP16.
+//===----------------------------------------------------------------------===//
+
+v_cmp_gt_u16 v1, v2 row_shl:0x7 row_mask:0x0 bank_mask:0x0 fi:1
+// GFX11: v_cmp_gt_u16 vcc_lo, v1, v2 row_shl:7 row_mask:0x0 bank_mask:0x0 fi:1 ; encoding: [0xfa,0x04,0x78,0x7c,0x01,0x07,0x05,0x00]
+
+v_cmp_gt_u16_dpp v1, v2 row_shl:0x7 row_mask:0x0 bank_mask:0x0 fi:1
+// GFX11: v_cmp_gt_u16 vcc_lo, v1, v2 row_shl:7 row_mask:0x0 bank_mask:0x0 fi:1 ; encoding: [0xfa,0x04,0x78,0x7c,0x01,0x07,0x05,0x00]
+
+//===----------------------------------------------------------------------===//
+// VOPCX.DPP8.
+//===----------------------------------------------------------------------===//
+
+v_cmpx_class_f16 v1, v2 dpp8:[7,6,5,4,3,2,1,0]
+// GFX11: v_cmpx_class_f16 v1, v2 dpp8:[7,6,5,4,3,2,1,0] ; encoding: [0xe9,0x04,0xfa,0x7d,0x01,0x77,0x39,0x05]
+
+v_cmpx_class_f16_dpp v1, v2 dpp8:[7,6,5,4,3,2,1,0]
+// GFX11: v_cmpx_class_f16 v1, v2 dpp8:[7,6,5,4,3,2,1,0] ; encoding: [0xe9,0x04,0xfa,0x7d,0x01,0x77,0x39,0x05]
+
+//===----------------------------------------------------------------------===//
+// VOPCX.DPP16.
+//===----------------------------------------------------------------------===//
+
+v_cmpx_class_f16 v1, v2 quad_perm:[3,2,1,0]
+// GFX11: v_cmpx_class_f16 v1, v2 quad_perm:[3,2,1,0] row_mask:0xf bank_mask:0xf ; encoding: [0xfa,0x04,0xfa,0x7d,0x01,0x1b,0x00,0xff]
+
+v_cmpx_class_f16_dpp v1, v2 quad_perm:[3,2,1,0]
+// GFX11: v_cmpx_class_f16 v1, v2 quad_perm:[3,2,1,0] row_mask:0xf bank_mask:0xf ; encoding: [0xfa,0x04,0xfa,0x7d,0x01,0x1b,0x00,0xff]
+
+//===----------------------------------------------------------------------===//
+// VOP1 -> VOP3.
+//===----------------------------------------------------------------------===//
+
+v_sin_f32 v5, 0.5 mul:2
+// GFX11: v_sin_f32_e64 v5, 0.5 mul:2             ; encoding: [0x05,0x00,0xb5,0xd5,0xf0,0x00,0x00,0x08]
+
+v_sin_f32_e64 v5, 0.5 mul:2
+// GFX11: v_sin_f32_e64 v5, 0.5 mul:2             ; encoding: [0x05,0x00,0xb5,0xd5,0xf0,0x00,0x00,0x08]
+
+v_sin_f32_e64 v5, v1
+// GFX11: v_sin_f32_e64 v5, v1                    ; encoding: [0x05,0x00,0xb5,0xd5,0x01,0x01,0x00,0x00]
+
+v_sin_f32 v5, v1
+// GFX11: v_sin_f32_e32 v5, v1                    ; encoding: [0x01,0x6b,0x0a,0x7e]
+
+//===----------------------------------------------------------------------===//
+// VOP2 -> VOP3.
+//===----------------------------------------------------------------------===//
+
+v_add_f32 v5, v1, -v2
+// GFX11: v_add_f32_e64 v5, v1, -v2               ; encoding: [0x05,0x00,0x03,0xd5,0x01,0x05,0x02,0x40]
+
+v_add_f32_e64 v5, v1, -v2
+// GFX11: v_add_f32_e64 v5, v1, -v2               ; encoding: [0x05,0x00,0x03,0xd5,0x01,0x05,0x02,0x40]
+
+v_add_f32_e64 v5, v1, v2
+// GFX11: v_add_f32_e64 v5, v1, v2                ; encoding: [0x05,0x00,0x03,0xd5,0x01,0x05,0x02,0x00]
+
+v_add_f32 v5, v1, v2
+// GFX11: v_add_f32_e32 v5, v1, v2                ; encoding: [0x01,0x05,0x0a,0x06]
+
+//===----------------------------------------------------------------------===//
+// VOPC -> VOP3.
+//===----------------------------------------------------------------------===//
+
+v_cmp_f_f32 s10, -v1, v2
+// GFX11: v_cmp_f_f32_e64 s10, -v1, v2            ; encoding: [0x0a,0x00,0x10,0xd4,0x01,0x05,0x02,0x20]
+
+v_cmp_f_f32_e64 s10, -v1, v2
+// GFX11: v_cmp_f_f32_e64 s10, -v1, v2            ; encoding: [0x0a,0x00,0x10,0xd4,0x01,0x05,0x02,0x20]
+
+v_cmp_f_f32_e64 vcc_lo, v1, v2
+// GFX11: v_cmp_f_f32_e64 vcc_lo, v1, v2          ; encoding: [0x6a,0x00,0x10,0xd4,0x01,0x05,0x02,0x00]
+
+v_cmp_f_f32 vcc_lo, v1, v2
+// GFX11: v_cmp_f_f32_e32 vcc_lo, v1, v2          ; encoding: [0x01,0x05,0x20,0x7c]
+
+//===----------------------------------------------------------------------===//
+// VOPCX -> VOP3.
+//===----------------------------------------------------------------------===//
+
+v_cmpx_f_f32 -v1, v2
+// GFX11: v_cmpx_f_f32_e64 -v1, v2                ; encoding: [0x7e,0x00,0x90,0xd4,0x01,0x05,0x02,0x20]
+
+v_cmpx_f_f32_e64 -v1, v2
+// GFX11: v_cmpx_f_f32_e64 -v1, v2                ; encoding: [0x7e,0x00,0x90,0xd4,0x01,0x05,0x02,0x20]
+
+v_cmpx_f_f32_e64 v1, v2
+// GFX11: v_cmpx_f_f32_e64 v1, v2                 ; encoding: [0x7e,0x00,0x90,0xd4,0x01,0x05,0x02,0x00]
+
+v_cmpx_f_f32 v1, v2
+// GFX11: v_cmpx_f_f32_e...
[truncated]

@broxigarchen broxigarchen force-pushed the main-merge-true16-mc-clean-up-4 branch from 5a21bc5 to bc2eeac Compare April 15, 2025 17:15
@broxigarchen broxigarchen force-pushed the main-merge-true16-mc-clean-up-4 branch from bc2eeac to 25dc466 Compare April 15, 2025 17:17
Sisyph
Sisyph reviewed Apr 15, 2025
View reviewed changes
llvm/test/MC/AMDGPU/gfx11_asm_vinterp_err-fake16.s
@@ -0,0 +1,43 @@
// RUN: not llvm-mc -triple=amdgcn -mcpu=gfx1100 -mattr=-real-true16 %s 2>&1 | FileCheck %s -check-prefix=GCN-ERR --implicit-check-not=error: --strict-whitespace
// RUN: not llvm-mc -triple=amdgcn -mcpu=gfx1200 -mattr=-real-true16 %s 2>&1 | FileCheck %s -check-prefix=GCN-ERR --implicit-check-not=error: --strict-whitespace

Copy link
Contributor

Choose a reason for hiding this comment

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

Do you also need to add -mattr=+real-true16 to llvm/test/MC/AMDGPU/gfx11_asm_vinterp_err.s?

Copy link
Contributor Author

Choose a reason for hiding this comment

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

it's in a seperate PR #135823

@broxigarchen broxigarchen force-pushed the main-merge-true16-mc-clean-up-4 branch from 75d7a52 to fce2897 Compare April 15, 2025 19:44
@broxigarchen broxigarchen force-pushed the main-merge-true16-mc-clean-up-4 branch from fce2897 to 6d793ef Compare April 15, 2025 19:45
@broxigarchen broxigarchen requested a review from kosarev April 15, 2025 19:45
Sisyph
Sisyph approved these changes Apr 15, 2025
View reviewed changes
Copy link
Contributor

@Sisyph Sisyph left a comment

Choose a reason for hiding this comment

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

LGTM

@broxigarchen broxigarchen merged commit 181872f into llvm:main Apr 16, 2025
9 of 11 checks passed
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Reviewers

@shiltian shiltian shiltian approved these changes

@Sisyph Sisyph Sisyph approved these changes

@kosarev kosarev Awaiting requested review from kosarev

Assignees

No one assigned

Labels

backend:AMDGPU llvm:mc Machine (object) code

Projects

None yet

Milestone

No milestone

Development

Successfully merging this pull request may close these issues.

5 participants

@broxigarchen @llvmbot @Sisyph @shiltian @kosarev