Skip to content

[AMDGPU][CodeGen][True16][Test] add test files for gfx12 vop1 instructions in true16/fake16 format #106093

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 1 commit into from
Aug 27, 2024
Merged

[AMDGPU][CodeGen][True16][Test] add test files for gfx12 vop1 instructions in true16/fake16 format #106093

merged 1 commit into from
Aug 27, 2024

Conversation

broxigarchen
Copy link
Contributor

@broxigarchen broxigarchen commented Aug 26, 2024 •
edited
Loading

This is a NFC change to add tests for true16/fake16 flow. We need to have two sets of asm/disasm tests for true16 and fake16 flow and this patch is adding the missing one. The naming convension is that true16 filename is the default one while the fake16 filename has "fake16" attached to it.

This patch

  1. add true16 and fake16 version for vop1 test files
  2. rename a test file to keep a consistant naming pattern

The true16 test file will be updated when more true16 commands are supported in the up coming patches

@broxigarchen broxigarchen mentioned this pull request Aug 26, 2024
Merged
@broxigarchen broxigarchen marked this pull request as ready for review August 26, 2024 17:25
@llvmbot llvmbot added backend:AMDGPU mc Machine (object) code labels Aug 26, 2024
@llvmbot
Copy link
Member

llvmbot commented Aug 26, 2024 •
edited
Loading

@llvm/pr-subscribers-backend-amdgpu

@llvm/pr-subscribers-mc

Author: Brox Chen (broxigarchen)

Changes

This is a NFC change to add tests for true16/fake16 flow. This patch

  1. add true16 and fake16 version for vop1 test files
  2. rename a test file to keep a consistant naming pattern

The true16 test file will be updated when more true16 commands are supported in the up coming patches

Patch is 267.11 KiB, truncated to 20.00 KiB below, full version: https://github.com/llvm/llvm-project/pull/106093.diff

8 Files Affected:

  • (renamed) llvm/test/MC/AMDGPU/gfx12_asm_vop1_err-fake16.s ()
  • (renamed) llvm/test/MC/AMDGPU/gfx12_asm_vop1_err.s ()
  • (added) llvm/test/MC/AMDGPU/gfx12_asm_vop3_from_vop1-fake16.s (+3646)
  • (modified) llvm/test/MC/AMDGPU/gfx12_asm_vop3_from_vop1.s (+1-1)
  • (added) llvm/test/MC/AMDGPU/gfx12_asm_vop3_from_vop1_dpp16-fake16.s (+2731)
  • (modified) llvm/test/MC/AMDGPU/gfx12_asm_vop3_from_vop1_dpp16.s (+1-1)
  • (added) llvm/test/MC/AMDGPU/gfx12_asm_vop3_from_vop1_dpp8-fake16.s (+733)
  • (modified) llvm/test/MC/AMDGPU/gfx12_asm_vop3_from_vop1_dpp8.s (+1-1)
diff --git a/llvm/test/MC/AMDGPU/gfx12_asm_vop1_t16_err-fake16.s b/llvm/test/MC/AMDGPU/gfx12_asm_vop1_err-fake16.s
similarity index 100%
rename from llvm/test/MC/AMDGPU/gfx12_asm_vop1_t16_err-fake16.s
rename to llvm/test/MC/AMDGPU/gfx12_asm_vop1_err-fake16.s
diff --git a/llvm/test/MC/AMDGPU/gfx12_asm_vop1_t16_err.s b/llvm/test/MC/AMDGPU/gfx12_asm_vop1_err.s
similarity index 100%
rename from llvm/test/MC/AMDGPU/gfx12_asm_vop1_t16_err.s
rename to llvm/test/MC/AMDGPU/gfx12_asm_vop1_err.s
diff --git a/llvm/test/MC/AMDGPU/gfx12_asm_vop3_from_vop1-fake16.s b/llvm/test/MC/AMDGPU/gfx12_asm_vop3_from_vop1-fake16.s
new file mode 100644
index 00000000000000..438ebe20e88b4e
--- /dev/null
+++ b/llvm/test/MC/AMDGPU/gfx12_asm_vop3_from_vop1-fake16.s
@@ -0,0 +1,3646 @@
+// RUN: llvm-mc -triple=amdgcn -mcpu=gfx1200 -mattr=+wavefrontsize32,-real-true16 -show-encoding %s | FileCheck --check-prefix=GFX12 %s
+
+v_bfrev_b32_e64 v5, v1
+// GFX12: encoding: [0x05,0x00,0xb8,0xd5,0x01,0x01,0x00,0x00]
+
+v_bfrev_b32_e64 v5, v255
+// GFX12: encoding: [0x05,0x00,0xb8,0xd5,0xff,0x01,0x00,0x00]
+
+v_bfrev_b32_e64 v5, s1
+// GFX12: encoding: [0x05,0x00,0xb8,0xd5,0x01,0x00,0x00,0x00]
+
+v_bfrev_b32_e64 v5, s105
+// GFX12: encoding: [0x05,0x00,0xb8,0xd5,0x69,0x00,0x00,0x00]
+
+v_bfrev_b32_e64 v5, vcc_lo
+// GFX12: encoding: [0x05,0x00,0xb8,0xd5,0x6a,0x00,0x00,0x00]
+
+v_bfrev_b32_e64 v5, vcc_hi
+// GFX12: encoding: [0x05,0x00,0xb8,0xd5,0x6b,0x00,0x00,0x00]
+
+v_bfrev_b32_e64 v5, ttmp15
+// GFX12: encoding: [0x05,0x00,0xb8,0xd5,0x7b,0x00,0x00,0x00]
+
+v_bfrev_b32_e64 v5, m0
+// GFX12: encoding: [0x05,0x00,0xb8,0xd5,0x7d,0x00,0x00,0x00]
+
+v_bfrev_b32_e64 v5, exec_lo
+// GFX12: encoding: [0x05,0x00,0xb8,0xd5,0x7e,0x00,0x00,0x00]
+
+v_bfrev_b32_e64 v5, exec_hi
+// GFX12: encoding: [0x05,0x00,0xb8,0xd5,0x7f,0x00,0x00,0x00]
+
+v_bfrev_b32_e64 v5, null
+// GFX12: encoding: [0x05,0x00,0xb8,0xd5,0x7c,0x00,0x00,0x00]
+
+v_bfrev_b32_e64 v5, -1
+// GFX12: encoding: [0x05,0x00,0xb8,0xd5,0xc1,0x00,0x00,0x00]
+
+v_bfrev_b32_e64 v5, 0.5
+// GFX12: encoding: [0x05,0x00,0xb8,0xd5,0xf0,0x00,0x00,0x00]
+
+v_bfrev_b32_e64 v5, src_scc
+// GFX12: encoding: [0x05,0x00,0xb8,0xd5,0xfd,0x00,0x00,0x00]
+
+v_bfrev_b32_e64 v255, 0xaf123456
+// GFX12: encoding: [0xff,0x00,0xb8,0xd5,0xff,0x00,0x00,0x00,0x56,0x34,0x12,0xaf]
+
+v_ceil_f16_e64 v5, v1
+// GFX12: encoding: [0x05,0x00,0xdc,0xd5,0x01,0x01,0x00,0x00]
+
+v_ceil_f16_e64 v5, v255
+// GFX12: encoding: [0x05,0x00,0xdc,0xd5,0xff,0x01,0x00,0x00]
+
+v_ceil_f16_e64 v5, s1
+// GFX12: encoding: [0x05,0x00,0xdc,0xd5,0x01,0x00,0x00,0x00]
+
+v_ceil_f16_e64 v5, s105
+// GFX12: encoding: [0x05,0x00,0xdc,0xd5,0x69,0x00,0x00,0x00]
+
+v_ceil_f16_e64 v5, vcc_lo
+// GFX12: encoding: [0x05,0x00,0xdc,0xd5,0x6a,0x00,0x00,0x00]
+
+v_ceil_f16_e64 v5, vcc_hi
+// GFX12: encoding: [0x05,0x00,0xdc,0xd5,0x6b,0x00,0x00,0x00]
+
+v_ceil_f16_e64 v5, ttmp15
+// GFX12: encoding: [0x05,0x00,0xdc,0xd5,0x7b,0x00,0x00,0x00]
+
+v_ceil_f16_e64 v5, m0
+// GFX12: encoding: [0x05,0x00,0xdc,0xd5,0x7d,0x00,0x00,0x00]
+
+v_ceil_f16_e64 v5, exec_lo
+// GFX12: encoding: [0x05,0x00,0xdc,0xd5,0x7e,0x00,0x00,0x00]
+
+v_ceil_f16_e64 v5, exec_hi
+// GFX12: encoding: [0x05,0x00,0xdc,0xd5,0x7f,0x00,0x00,0x00]
+
+v_ceil_f16_e64 v5, null
+// GFX12: encoding: [0x05,0x00,0xdc,0xd5,0x7c,0x00,0x00,0x00]
+
+v_ceil_f16_e64 v5, -1
+// GFX12: encoding: [0x05,0x00,0xdc,0xd5,0xc1,0x00,0x00,0x00]
+
+v_ceil_f16_e64 v5, 0.5 mul:2
+// GFX12: encoding: [0x05,0x00,0xdc,0xd5,0xf0,0x00,0x00,0x08]
+
+v_ceil_f16_e64 v5, src_scc mul:4
+// GFX12: encoding: [0x05,0x00,0xdc,0xd5,0xfd,0x00,0x00,0x10]
+
+v_ceil_f16_e64 v255, -|0xfe0b| clamp div:2
+// GFX12: encoding: [0xff,0x81,0xdc,0xd5,0xff,0x00,0x00,0x38,0x0b,0xfe,0x00,0x00]
+
+v_ceil_f32_e64 v5, v1
+// GFX12: encoding: [0x05,0x00,0xa2,0xd5,0x01,0x01,0x00,0x00]
+
+v_ceil_f32_e64 v5, v255
+// GFX12: encoding: [0x05,0x00,0xa2,0xd5,0xff,0x01,0x00,0x00]
+
+v_ceil_f32_e64 v5, s1
+// GFX12: encoding: [0x05,0x00,0xa2,0xd5,0x01,0x00,0x00,0x00]
+
+v_ceil_f32_e64 v5, s105
+// GFX12: encoding: [0x05,0x00,0xa2,0xd5,0x69,0x00,0x00,0x00]
+
+v_ceil_f32_e64 v5, vcc_lo
+// GFX12: encoding: [0x05,0x00,0xa2,0xd5,0x6a,0x00,0x00,0x00]
+
+v_ceil_f32_e64 v5, vcc_hi
+// GFX12: encoding: [0x05,0x00,0xa2,0xd5,0x6b,0x00,0x00,0x00]
+
+v_ceil_f32_e64 v5, ttmp15
+// GFX12: encoding: [0x05,0x00,0xa2,0xd5,0x7b,0x00,0x00,0x00]
+
+v_ceil_f32_e64 v5, m0
+// GFX12: encoding: [0x05,0x00,0xa2,0xd5,0x7d,0x00,0x00,0x00]
+
+v_ceil_f32_e64 v5, exec_lo
+// GFX12: encoding: [0x05,0x00,0xa2,0xd5,0x7e,0x00,0x00,0x00]
+
+v_ceil_f32_e64 v5, exec_hi
+// GFX12: encoding: [0x05,0x00,0xa2,0xd5,0x7f,0x00,0x00,0x00]
+
+v_ceil_f32_e64 v5, null
+// GFX12: encoding: [0x05,0x00,0xa2,0xd5,0x7c,0x00,0x00,0x00]
+
+v_ceil_f32_e64 v5, -1
+// GFX12: encoding: [0x05,0x00,0xa2,0xd5,0xc1,0x00,0x00,0x00]
+
+v_ceil_f32_e64 v5, 0.5 mul:2
+// GFX12: encoding: [0x05,0x00,0xa2,0xd5,0xf0,0x00,0x00,0x08]
+
+v_ceil_f32_e64 v5, src_scc mul:4
+// GFX12: encoding: [0x05,0x00,0xa2,0xd5,0xfd,0x00,0x00,0x10]
+
+v_ceil_f32_e64 v255, -|0xaf123456| clamp div:2
+// GFX12: encoding: [0xff,0x81,0xa2,0xd5,0xff,0x00,0x00,0x38,0x56,0x34,0x12,0xaf]
+
+v_ceil_f64_e64 v[5:6], v[1:2]
+// GFX12: encoding: [0x05,0x00,0x98,0xd5,0x01,0x01,0x00,0x00]
+
+v_ceil_f64_e64 v[5:6], v[254:255]
+// GFX12: encoding: [0x05,0x00,0x98,0xd5,0xfe,0x01,0x00,0x00]
+
+v_ceil_f64_e64 v[5:6], s[2:3]
+// GFX12: encoding: [0x05,0x00,0x98,0xd5,0x02,0x00,0x00,0x00]
+
+v_ceil_f64_e64 v[5:6], s[104:105]
+// GFX12: encoding: [0x05,0x00,0x98,0xd5,0x68,0x00,0x00,0x00]
+
+v_ceil_f64_e64 v[5:6], vcc
+// GFX12: encoding: [0x05,0x00,0x98,0xd5,0x6a,0x00,0x00,0x00]
+
+v_ceil_f64_e64 v[5:6], ttmp[14:15]
+// GFX12: encoding: [0x05,0x00,0x98,0xd5,0x7a,0x00,0x00,0x00]
+
+v_ceil_f64_e64 v[5:6], exec
+// GFX12: encoding: [0x05,0x00,0x98,0xd5,0x7e,0x00,0x00,0x00]
+
+v_ceil_f64_e64 v[5:6], null
+// GFX12: encoding: [0x05,0x00,0x98,0xd5,0x7c,0x00,0x00,0x00]
+
+v_ceil_f64_e64 v[5:6], -1
+// GFX12: encoding: [0x05,0x00,0x98,0xd5,0xc1,0x00,0x00,0x00]
+
+v_ceil_f64_e64 v[5:6], 0.5 mul:2
+// GFX12: encoding: [0x05,0x00,0x98,0xd5,0xf0,0x00,0x00,0x08]
+
+v_ceil_f64_e64 v[5:6], -|src_scc| mul:4
+// GFX12: encoding: [0x05,0x01,0x98,0xd5,0xfd,0x00,0x00,0x30]
+
+v_ceil_f64_e64 v[254:255], 0xaf123456 clamp div:2
+// GFX12: encoding: [0xfe,0x80,0x98,0xd5,0xff,0x00,0x00,0x18,0x56,0x34,0x12,0xaf]
+
+v_cls_i32_e64 v5, v1
+// GFX12: encoding: [0x05,0x00,0xbb,0xd5,0x01,0x01,0x00,0x00]
+
+v_cls_i32_e64 v5, v255
+// GFX12: encoding: [0x05,0x00,0xbb,0xd5,0xff,0x01,0x00,0x00]
+
+v_cls_i32_e64 v5, s1
+// GFX12: encoding: [0x05,0x00,0xbb,0xd5,0x01,0x00,0x00,0x00]
+
+v_cls_i32_e64 v5, s105
+// GFX12: encoding: [0x05,0x00,0xbb,0xd5,0x69,0x00,0x00,0x00]
+
+v_cls_i32_e64 v5, vcc_lo
+// GFX12: encoding: [0x05,0x00,0xbb,0xd5,0x6a,0x00,0x00,0x00]
+
+v_cls_i32_e64 v5, vcc_hi
+// GFX12: encoding: [0x05,0x00,0xbb,0xd5,0x6b,0x00,0x00,0x00]
+
+v_cls_i32_e64 v5, ttmp15
+// GFX12: encoding: [0x05,0x00,0xbb,0xd5,0x7b,0x00,0x00,0x00]
+
+v_cls_i32_e64 v5, m0
+// GFX12: encoding: [0x05,0x00,0xbb,0xd5,0x7d,0x00,0x00,0x00]
+
+v_cls_i32_e64 v5, exec_lo
+// GFX12: encoding: [0x05,0x00,0xbb,0xd5,0x7e,0x00,0x00,0x00]
+
+v_cls_i32_e64 v5, exec_hi
+// GFX12: encoding: [0x05,0x00,0xbb,0xd5,0x7f,0x00,0x00,0x00]
+
+v_cls_i32_e64 v5, null
+// GFX12: encoding: [0x05,0x00,0xbb,0xd5,0x7c,0x00,0x00,0x00]
+
+v_cls_i32_e64 v5, -1
+// GFX12: encoding: [0x05,0x00,0xbb,0xd5,0xc1,0x00,0x00,0x00]
+
+v_cls_i32_e64 v5, 0.5
+// GFX12: encoding: [0x05,0x00,0xbb,0xd5,0xf0,0x00,0x00,0x00]
+
+v_cls_i32_e64 v5, src_scc
+// GFX12: encoding: [0x05,0x00,0xbb,0xd5,0xfd,0x00,0x00,0x00]
+
+v_cls_i32_e64 v255, 0xaf123456
+// GFX12: encoding: [0xff,0x00,0xbb,0xd5,0xff,0x00,0x00,0x00,0x56,0x34,0x12,0xaf]
+
+v_clz_i32_u32_e64 v5, v1
+// GFX12: encoding: [0x05,0x00,0xb9,0xd5,0x01,0x01,0x00,0x00]
+
+v_clz_i32_u32_e64 v5, v255
+// GFX12: encoding: [0x05,0x00,0xb9,0xd5,0xff,0x01,0x00,0x00]
+
+v_clz_i32_u32_e64 v5, s1
+// GFX12: encoding: [0x05,0x00,0xb9,0xd5,0x01,0x00,0x00,0x00]
+
+v_clz_i32_u32_e64 v5, s105
+// GFX12: encoding: [0x05,0x00,0xb9,0xd5,0x69,0x00,0x00,0x00]
+
+v_clz_i32_u32_e64 v5, vcc_lo
+// GFX12: encoding: [0x05,0x00,0xb9,0xd5,0x6a,0x00,0x00,0x00]
+
+v_clz_i32_u32_e64 v5, vcc_hi
+// GFX12: encoding: [0x05,0x00,0xb9,0xd5,0x6b,0x00,0x00,0x00]
+
+v_clz_i32_u32_e64 v5, ttmp15
+// GFX12: encoding: [0x05,0x00,0xb9,0xd5,0x7b,0x00,0x00,0x00]
+
+v_clz_i32_u32_e64 v5, m0
+// GFX12: encoding: [0x05,0x00,0xb9,0xd5,0x7d,0x00,0x00,0x00]
+
+v_clz_i32_u32_e64 v5, exec_lo
+// GFX12: encoding: [0x05,0x00,0xb9,0xd5,0x7e,0x00,0x00,0x00]
+
+v_clz_i32_u32_e64 v5, exec_hi
+// GFX12: encoding: [0x05,0x00,0xb9,0xd5,0x7f,0x00,0x00,0x00]
+
+v_clz_i32_u32_e64 v5, null
+// GFX12: encoding: [0x05,0x00,0xb9,0xd5,0x7c,0x00,0x00,0x00]
+
+v_clz_i32_u32_e64 v5, -1
+// GFX12: encoding: [0x05,0x00,0xb9,0xd5,0xc1,0x00,0x00,0x00]
+
+v_clz_i32_u32_e64 v5, 0.5
+// GFX12: encoding: [0x05,0x00,0xb9,0xd5,0xf0,0x00,0x00,0x00]
+
+v_clz_i32_u32_e64 v5, src_scc
+// GFX12: encoding: [0x05,0x00,0xb9,0xd5,0xfd,0x00,0x00,0x00]
+
+v_clz_i32_u32_e64 v255, 0xaf123456
+// GFX12: encoding: [0xff,0x00,0xb9,0xd5,0xff,0x00,0x00,0x00,0x56,0x34,0x12,0xaf]
+
+v_cos_f16_e64 v5, v1
+// GFX12: encoding: [0x05,0x00,0xe1,0xd5,0x01,0x01,0x00,0x00]
+
+v_cos_f16_e64 v5, v255
+// GFX12: encoding: [0x05,0x00,0xe1,0xd5,0xff,0x01,0x00,0x00]
+
+v_cos_f16_e64 v5, s1
+// GFX12: encoding: [0x05,0x00,0xe1,0xd5,0x01,0x00,0x00,0x00]
+
+v_cos_f16_e64 v5, s105
+// GFX12: encoding: [0x05,0x00,0xe1,0xd5,0x69,0x00,0x00,0x00]
+
+v_cos_f16_e64 v5, vcc_lo
+// GFX12: encoding: [0x05,0x00,0xe1,0xd5,0x6a,0x00,0x00,0x00]
+
+v_cos_f16_e64 v5, vcc_hi
+// GFX12: encoding: [0x05,0x00,0xe1,0xd5,0x6b,0x00,0x00,0x00]
+
+v_cos_f16_e64 v5, ttmp15
+// GFX12: encoding: [0x05,0x00,0xe1,0xd5,0x7b,0x00,0x00,0x00]
+
+v_cos_f16_e64 v5, m0
+// GFX12: encoding: [0x05,0x00,0xe1,0xd5,0x7d,0x00,0x00,0x00]
+
+v_cos_f16_e64 v5, exec_lo
+// GFX12: encoding: [0x05,0x00,0xe1,0xd5,0x7e,0x00,0x00,0x00]
+
+v_cos_f16_e64 v5, exec_hi
+// GFX12: encoding: [0x05,0x00,0xe1,0xd5,0x7f,0x00,0x00,0x00]
+
+v_cos_f16_e64 v5, null
+// GFX12: encoding: [0x05,0x00,0xe1,0xd5,0x7c,0x00,0x00,0x00]
+
+v_cos_f16_e64 v5, -1
+// GFX12: encoding: [0x05,0x00,0xe1,0xd5,0xc1,0x00,0x00,0x00]
+
+v_cos_f16_e64 v5, 0.5 mul:2
+// GFX12: encoding: [0x05,0x00,0xe1,0xd5,0xf0,0x00,0x00,0x08]
+
+v_cos_f16_e64 v5, src_scc mul:4
+// GFX12: encoding: [0x05,0x00,0xe1,0xd5,0xfd,0x00,0x00,0x10]
+
+v_cos_f16_e64 v255, -|0xfe0b| clamp div:2
+// GFX12: encoding: [0xff,0x81,0xe1,0xd5,0xff,0x00,0x00,0x38,0x0b,0xfe,0x00,0x00]
+
+v_cos_f32_e64 v5, v1
+// GFX12: encoding: [0x05,0x00,0xb6,0xd5,0x01,0x01,0x00,0x00]
+
+v_cos_f32_e64 v5, v255
+// GFX12: encoding: [0x05,0x00,0xb6,0xd5,0xff,0x01,0x00,0x00]
+
+v_cos_f32_e64 v5, s1
+// GFX12: encoding: [0x05,0x00,0xb6,0xd5,0x01,0x00,0x00,0x00]
+
+v_cos_f32_e64 v5, s105
+// GFX12: encoding: [0x05,0x00,0xb6,0xd5,0x69,0x00,0x00,0x00]
+
+v_cos_f32_e64 v5, vcc_lo
+// GFX12: encoding: [0x05,0x00,0xb6,0xd5,0x6a,0x00,0x00,0x00]
+
+v_cos_f32_e64 v5, vcc_hi
+// GFX12: encoding: [0x05,0x00,0xb6,0xd5,0x6b,0x00,0x00,0x00]
+
+v_cos_f32_e64 v5, ttmp15
+// GFX12: encoding: [0x05,0x00,0xb6,0xd5,0x7b,0x00,0x00,0x00]
+
+v_cos_f32_e64 v5, m0
+// GFX12: encoding: [0x05,0x00,0xb6,0xd5,0x7d,0x00,0x00,0x00]
+
+v_cos_f32_e64 v5, exec_lo
+// GFX12: encoding: [0x05,0x00,0xb6,0xd5,0x7e,0x00,0x00,0x00]
+
+v_cos_f32_e64 v5, exec_hi
+// GFX12: encoding: [0x05,0x00,0xb6,0xd5,0x7f,0x00,0x00,0x00]
+
+v_cos_f32_e64 v5, null
+// GFX12: encoding: [0x05,0x00,0xb6,0xd5,0x7c,0x00,0x00,0x00]
+
+v_cos_f32_e64 v5, -1
+// GFX12: encoding: [0x05,0x00,0xb6,0xd5,0xc1,0x00,0x00,0x00]
+
+v_cos_f32_e64 v5, 0.5 mul:2
+// GFX12: encoding: [0x05,0x00,0xb6,0xd5,0xf0,0x00,0x00,0x08]
+
+v_cos_f32_e64 v5, src_scc mul:4
+// GFX12: encoding: [0x05,0x00,0xb6,0xd5,0xfd,0x00,0x00,0x10]
+
+v_cos_f32_e64 v255, -|0xaf123456| clamp div:2
+// GFX12: encoding: [0xff,0x81,0xb6,0xd5,0xff,0x00,0x00,0x38,0x56,0x34,0x12,0xaf]
+
+v_ctz_i32_b32_e64 v5, v1
+// GFX12: encoding: [0x05,0x00,0xba,0xd5,0x01,0x01,0x00,0x00]
+
+v_ctz_i32_b32_e64 v5, v255
+// GFX12: encoding: [0x05,0x00,0xba,0xd5,0xff,0x01,0x00,0x00]
+
+v_ctz_i32_b32_e64 v5, s1
+// GFX12: encoding: [0x05,0x00,0xba,0xd5,0x01,0x00,0x00,0x00]
+
+v_ctz_i32_b32_e64 v5, s105
+// GFX12: encoding: [0x05,0x00,0xba,0xd5,0x69,0x00,0x00,0x00]
+
+v_ctz_i32_b32_e64 v5, vcc_lo
+// GFX12: encoding: [0x05,0x00,0xba,0xd5,0x6a,0x00,0x00,0x00]
+
+v_ctz_i32_b32_e64 v5, vcc_hi
+// GFX12: encoding: [0x05,0x00,0xba,0xd5,0x6b,0x00,0x00,0x00]
+
+v_ctz_i32_b32_e64 v5, ttmp15
+// GFX12: encoding: [0x05,0x00,0xba,0xd5,0x7b,0x00,0x00,0x00]
+
+v_ctz_i32_b32_e64 v5, m0
+// GFX12: encoding: [0x05,0x00,0xba,0xd5,0x7d,0x00,0x00,0x00]
+
+v_ctz_i32_b32_e64 v5, exec_lo
+// GFX12: encoding: [0x05,0x00,0xba,0xd5,0x7e,0x00,0x00,0x00]
+
+v_ctz_i32_b32_e64 v5, exec_hi
+// GFX12: encoding: [0x05,0x00,0xba,0xd5,0x7f,0x00,0x00,0x00]
+
+v_ctz_i32_b32_e64 v5, null
+// GFX12: encoding: [0x05,0x00,0xba,0xd5,0x7c,0x00,0x00,0x00]
+
+v_ctz_i32_b32_e64 v5, -1
+// GFX12: encoding: [0x05,0x00,0xba,0xd5,0xc1,0x00,0x00,0x00]
+
+v_ctz_i32_b32_e64 v5, 0.5
+// GFX12: encoding: [0x05,0x00,0xba,0xd5,0xf0,0x00,0x00,0x00]
+
+v_ctz_i32_b32_e64 v5, src_scc
+// GFX12: encoding: [0x05,0x00,0xba,0xd5,0xfd,0x00,0x00,0x00]
+
+v_ctz_i32_b32_e64 v255, 0xaf123456
+// GFX12: encoding: [0xff,0x00,0xba,0xd5,0xff,0x00,0x00,0x00,0x56,0x34,0x12,0xaf]
+
+v_cvt_f32_bf8_e64 v1, s3
+// GFX12: encoding: [0x01,0x00,0xed,0xd5,0x03,0x00,0x00,0x00]
+
+v_cvt_f32_bf8_e64 v1, s3 byte_sel:1
+// GFX12: encoding: [0x01,0x10,0xed,0xd5,0x03,0x00,0x00,0x00]
+
+v_cvt_f32_bf8_e64 v1, s3 byte_sel:2
+// GFX12: encoding: [0x01,0x08,0xed,0xd5,0x03,0x00,0x00,0x00]
+
+v_cvt_f32_bf8_e64 v1, s3 byte_sel:3
+// GFX12: encoding: [0x01,0x18,0xed,0xd5,0x03,0x00,0x00,0x00]
+
+v_cvt_f32_bf8_e64 v1, 3
+// GFX12: encoding: [0x01,0x00,0xed,0xd5,0x83,0x00,0x00,0x00]
+
+v_cvt_f32_bf8_e64 v1, 3 byte_sel:1
+// GFX12: encoding: [0x01,0x10,0xed,0xd5,0x83,0x00,0x00,0x00]
+
+v_cvt_f32_bf8_e64 v1, 3 byte_sel:2
+// GFX12: encoding: [0x01,0x08,0xed,0xd5,0x83,0x00,0x00,0x00]
+
+v_cvt_f32_bf8_e64 v1, 3 byte_sel:3
+// GFX12: encoding: [0x01,0x18,0xed,0xd5,0x83,0x00,0x00,0x00]
+
+v_cvt_f32_bf8_e64 v1, v3
+// GFX12: encoding: [0x01,0x00,0xed,0xd5,0x03,0x01,0x00,0x00]
+
+v_cvt_f32_bf8_e64 v1, v3 byte_sel:1
+// GFX12: encoding: [0x01,0x10,0xed,0xd5,0x03,0x01,0x00,0x00]
+
+v_cvt_f32_bf8_e64 v1, v3 byte_sel:2
+// GFX12: encoding: [0x01,0x08,0xed,0xd5,0x03,0x01,0x00,0x00]
+
+v_cvt_f32_bf8_e64 v1, v3 byte_sel:3
+// GFX12: encoding: [0x01,0x18,0xed,0xd5,0x03,0x01,0x00,0x00]
+
+v_cvt_f32_fp8_e64 v1, s3
+// GFX12: encoding: [0x01,0x00,0xec,0xd5,0x03,0x00,0x00,0x00]
+
+v_cvt_f32_fp8_e64 v1, s3 byte_sel:1
+// GFX12: encoding: [0x01,0x10,0xec,0xd5,0x03,0x00,0x00,0x00]
+
+v_cvt_f32_fp8_e64 v1, s3 byte_sel:2
+// GFX12: encoding: [0x01,0x08,0xec,0xd5,0x03,0x00,0x00,0x00]
+
+v_cvt_f32_fp8_e64 v1, s3 byte_sel:3
+// GFX12: encoding: [0x01,0x18,0xec,0xd5,0x03,0x00,0x00,0x00]
+
+v_cvt_f32_fp8_e64 v1, 3
+// GFX12: encoding: [0x01,0x00,0xec,0xd5,0x83,0x00,0x00,0x00]
+
+v_cvt_f32_fp8_e64 v1, 3 byte_sel:1
+// GFX12: encoding: [0x01,0x10,0xec,0xd5,0x83,0x00,0x00,0x00]
+
+v_cvt_f32_fp8_e64 v1, 3 byte_sel:2
+// GFX12: encoding: [0x01,0x08,0xec,0xd5,0x83,0x00,0x00,0x00]
+
+v_cvt_f32_fp8_e64 v1, 3 byte_sel:3
+// GFX12: encoding: [0x01,0x18,0xec,0xd5,0x83,0x00,0x00,0x00]
+
+v_cvt_f32_fp8_e64 v1, v3
+// GFX12: encoding: [0x01,0x00,0xec,0xd5,0x03,0x01,0x00,0x00]
+
+v_cvt_f32_fp8_e64 v1, v3 byte_sel:1
+// GFX12: encoding: [0x01,0x10,0xec,0xd5,0x03,0x01,0x00,0x00]
+
+v_cvt_f32_fp8_e64 v1, v3 byte_sel:2
+// GFX12: encoding: [0x01,0x08,0xec,0xd5,0x03,0x01,0x00,0x00]
+
+v_cvt_f32_fp8_e64 v1, v3 byte_sel:3
+// GFX12: encoding: [0x01,0x18,0xec,0xd5,0x03,0x01,0x00,0x00]
+
+v_cvt_pk_f32_bf8_e64 v[2:3], s3
+// GFX12: encoding: [0x02,0x00,0xef,0xd5,0x03,0x00,0x00,0x00]
+
+v_cvt_pk_f32_bf8_e64 v[2:3], s3 op_sel:[1,0]
+// GFX12: encoding: [0x02,0x08,0xef,0xd5,0x03,0x00,0x00,0x00]
+
+v_cvt_pk_f32_bf8_e64 v[2:3], 3
+// GFX12: encoding: [0x02,0x00,0xef,0xd5,0x83,0x00,0x00,0x00]
+
+v_cvt_pk_f32_bf8_e64 v[2:3], 3 op_sel:[1,0]
+// GFX12: encoding: [0x02,0x08,0xef,0xd5,0x83,0x00,0x00,0x00]
+
+v_cvt_pk_f32_bf8_e64 v[2:3], v3
+// GFX12: encoding: [0x02,0x00,0xef,0xd5,0x03,0x01,0x00,0x00]
+
+v_cvt_pk_f32_bf8_e64 v[2:3], v3 op_sel:[1,0]
+// GFX12: encoding: [0x02,0x08,0xef,0xd5,0x03,0x01,0x00,0x00]
+
+v_cvt_pk_f32_fp8_e64 v[2:3], s3
+// GFX12: encoding: [0x02,0x00,0xee,0xd5,0x03,0x00,0x00,0x00]
+
+v_cvt_pk_f32_fp8_e64 v[2:3], s3 op_sel:[1,0]
+// GFX12: encoding: [0x02,0x08,0xee,0xd5,0x03,0x00,0x00,0x00]
+
+v_cvt_pk_f32_fp8_e64 v[2:3], 3
+// GFX12: encoding: [0x02,0x00,0xee,0xd5,0x83,0x00,0x00,0x00]
+
+v_cvt_pk_f32_fp8_e64 v[2:3], 3 op_sel:[1,0]
+// GFX12: encoding: [0x02,0x08,0xee,0xd5,0x83,0x00,0x00,0x00]
+
+v_cvt_pk_f32_fp8_e64 v[2:3], v3
+// GFX12: encoding: [0x02,0x00,0xee,0xd5,0x03,0x01,0x00,0x00]
+
+v_cvt_pk_f32_fp8_e64 v[2:3], v3 op_sel:[1,0]
+// GFX12: encoding: [0x02,0x08,0xee,0xd5,0x03,0x01,0x00,0x00]
+
+v_cvt_pk_f32_bf8_e64 v[3:4], s3
+// GFX12: encoding: [0x03,0x00,0xef,0xd5,0x03,0x00,0x00,0x00]
+
+v_cvt_pk_f32_bf8_e64 v[3:4], s3 op_sel:[1,0]
+// GFX12: encoding: [0x03,0x08,0xef,0xd5,0x03,0x00,0x00,0x00]
+
+v_cvt_pk_f32_bf8_e64 v[3:4], 3 op_sel:[1,0]
+// GFX12: encoding: [0x03,0x08,0xef,0xd5,0x83,0x00,0x00,0x00]
+
+v_cvt_pk_f32_bf8_e64 v[3:4], v3
+// GFX12: encoding: [0x03,0x00,0xef,0xd5,0x03,0x01,0x00,0x00]
+
+v_cvt_pk_f32_bf8_e64 v[3:4], v3 op_sel:[1,0]
+// GFX12: encoding: [0x03,0x08,0xef,0xd5,0x03,0x01,0x00,0x00]
+
+v_cvt_pk_f32_fp8_e64 v[3:4], s3
+// GFX12: encoding: [0x03,0x00,0xee,0xd5,0x03,0x00,0x00,0x00]
+
+v_cvt_pk_f32_fp8_e64 v[3:4], 3
+// GFX12: encoding: [0x03,0x00,0xee,0xd5,0x83,0x00,0x00,0x00]
+
+v_cvt_pk_f32_fp8_e64 v[3:4], 3 op_sel:[1,0]
+// GFX12: encoding: [0x03,0x08,0xee,0xd5,0x83,0x00,0x00,0x00]
+
+v_cvt_pk_f32_fp8_e64 v[3:4], v3
+// GFX12: encoding: [0x03,0x00,0xee,0xd5,0x03,0x01,0x00,0x00]
+
+v_cvt_pk_f32_fp8_e64 v[3:4], v3 op_sel:[1,0]
+// GFX12: encoding: [0x03,0x08,0xee,0xd5,0x03,0x01,0x00,0x00]
+
+v_cvt_f16_f32_e64 v5, v1
+// GFX12: encoding: [0x05,0x00,0x8a,0xd5,0x01,0x01,0x00,0x00]
+
+v_cvt_f16_f32_e64 v5, v255
+// GFX12: encoding: [0x05,0x00,0x8a,0xd5,0xff,0x01,0x00,0x00]
+
+v_cvt_f16_f32_e64 v5, s1
+// GFX12: encoding: [0x05,0x00,0x8a,0xd5,0x01,0x00,0x00,0x00]
+
+v_cvt_f16_f32_e64 v5, s105
+// GFX12: encoding: [0x05,0x00,0x8a,0xd5,0x69,0x00,0x00,0x00]
+
+v_cvt_f16_f32_e64 v5, vcc_lo
+// GFX12: encoding: [0x05,0x00,0x8a,0xd5,0x6a,0x00,0x00,0x00]
+
+v_cvt_f16_f32_e64 v5, vcc_hi
+// GFX12: encoding: [0x05,0x00,0x8a,0xd5,0x6b,0x00,0x00,0x00]
+
+v_cvt_f16_f32_e64 v5, ttmp15
+// GFX12: encoding: [0x05,0x00,0x8a,0xd5,0x7b,0x00,0x00,0x00]
+
+v_cvt_f16_f32_e64 v5, m0
+// GFX12: encoding: [0x05,0x00,0x8a,0xd5,0x7d,0x00,0x00,0x00]
+
+v_cvt_f16_f32_e64 v5, exec_lo
+// GFX12: encoding: [0x05,0x00,0x8a,0xd5,0x7e,0x00,0x00,0x00]
+
+v_cvt_f16_f32_e64 v5, exec_hi
+// GFX12: encoding: [0x05,0x00,0x8a,0xd5,0x7f,0x00,0x00,0x00]
+
+v_cvt_f16_f32_e64 v5, null
+// GFX12: encoding: [0x05,0x00,0x8a,0xd5,0x7c,0x00,0x00,0x00]
+
+v_cvt_f16_f32_e64 v5, -1
+// GFX12: encoding: [0x05,0x00,0x8a,0xd5,0xc1,0x00,0x00,0x00]
+
+v_cvt_f16_f32_e64 v5, 0.5 mul:2
+// GFX12: encoding: [0x05,0x00,0x8a,0xd5,0xf0,0x00,0x00,0x08]
+
+v_cvt_f16_f32_e64 v5, src_scc mul:4
+// GFX12: encoding: [0x05,0x00,0x8a,0xd5,0xfd,0x00,0x00,0x10]
+
+v_cvt_f16_f32_e64 v255, -|0xaf123456| clamp div:2
+// GFX12: encoding: [0xff,0x81,0x8a,0xd5,0xff,0x00,0x00,0x38,0x56,0x34,0x12,0xaf]
+
+v_cvt_f16_i16_e64 v5, v1
+// GFX12: encoding: [0x05,0x00,0xd1,0xd5,0x01,0x01,0x00,0x00]
+
+v_cvt_f16_i16_e64 v5, v255
+// GFX12: encoding: [0x05,0x00,0xd1,0xd5,0xff,0x01,0x00,0x00]
+
+v_cvt_f16_i16_e64 v5, s1
+// GFX12: encoding: [0x05,0x00,0xd1,0xd5,0x01,0x00,0x00,0x00]
+
+v_cvt_f16_i16_e64 v5, s105
+// GFX12: encoding: [0x05,0x00,0xd1,0xd5,0x69,0x00,0x00,0x00]
+
+v_cvt_f16_i16_e64 v5, vcc_lo
+// GFX12: encoding: [0x05,0x00,0xd1,0xd5,0x6a,0x00,0x00,0x00]
+
+v_cvt_f16_i16_e64 v5, vcc_hi
+// GFX12: encoding: [0x05,0x00,0xd1,0xd5,0x6b,0x00,0x00,0x00]
+
+v_cvt_f16_i16_e64 v5, ttmp15
+// GFX12: encoding: [0x05,0x00,0xd1,0xd5,0x7b,0x00,0x00,0x00]
+
+v_cvt_f16_i16_e64 v5, m0
+// GFX12: encoding: [0x05,0x00,0xd1,0xd5,0x7d,0x00,0x00,0x00]
+
+v_cvt_f16_i16_e64 v5, exec_lo
+// GFX12: encoding: [0x05,0x00,0xd1,0xd5,0x7e,0x00,0x00,0x00]
+
+v_cvt_f16_i16_e64 v5, exec_hi
+// GFX12: encoding: [0x05,0x00,0xd1,0xd5,0x7f,0x00,0x00,0x00]
+
+v_cvt_f16_i16_e64 v5, null
+// GFX12: encoding: [0x05,0x00,0xd1,0xd5,0x7c,0x00,...
[truncated]

@broxigarchen broxigarchen requested a review from kosarev August 26, 2024 22:27
@broxigarchen broxigarchen force-pushed the main-merge-true16-mc-dup-gfx12-testfile branch from 2ca8c83 to 4e5a8bc Compare August 27, 2024 17:08
Sisyph
Sisyph approved these changes Aug 27, 2024
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 d9cf98f into llvm:main Aug 27, 2024
5 of 7 checks passed
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Reviewers

@Sisyph Sisyph Sisyph approved these changes

@rampitec rampitec Awaiting requested review from rampitec

@kosarev kosarev Awaiting requested review from kosarev

Assignees
No one assigned
Labels
backend:AMDGPU mc Machine (object) code
Projects
None yet
Milestone
No milestone
Development

Successfully merging this pull request may close these issues.
3 participants
@broxigarchen @llvmbot @Sisyph