Skip to content

AMDGPU: Add global-isel checks and rename fptrunc.v2f16.fpmath.ll #136609

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.

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 1 commit into from
Apr 21, 2025

Conversation

changpeng
Copy link
Contributor

Also remove the checks with -enable-unsafe-fp-math (already in fptrunc.f16.ll)

  Also remove the checks with -enable-unsafe-fp-math (already in fptrunc.f16.ll)
@llvmbot
Copy link
Member

llvmbot commented Apr 21, 2025

@llvm/pr-subscribers-backend-amdgpu

Author: Changpeng Fang (changpeng)

Changes

Also remove the checks with -enable-unsafe-fp-math (already in fptrunc.f16.ll)


Patch is 21.28 KiB, truncated to 20.00 KiB below, full version: https://github.com/llvm/llvm-project/pull/136609.diff

2 Files Affected:

  • (removed) llvm/test/CodeGen/AMDGPU/fptrunc.v2f16.fpmath.ll (-213)
  • (added) llvm/test/CodeGen/AMDGPU/fptrunc.v2f16.no.fast.path.ll (+193)
diff --git a/llvm/test/CodeGen/AMDGPU/fptrunc.v2f16.fpmath.ll b/llvm/test/CodeGen/AMDGPU/fptrunc.v2f16.fpmath.ll
deleted file mode 100644
index 9e9b9629931b9..0000000000000
--- a/llvm/test/CodeGen/AMDGPU/fptrunc.v2f16.fpmath.ll
+++ /dev/null
@@ -1,213 +0,0 @@
-; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py UTC_ARGS: --version 5
-; RUN: llc -mtriple=amdgcn -mcpu=gfx950 -global-isel=0 < %s | FileCheck -check-prefixes=GFX950,GFX950-SAFE-SDAG %s
-; RUN: llc -mtriple=amdgcn -mcpu=gfx950 -global-isel=0 < %s | FileCheck -check-prefixes=GFX950,GFX950-SAFE-GISEL %s
-; RUN: llc -mtriple=amdgcn -mcpu=gfx950 -global-isel=0 -enable-unsafe-fp-math < %s | FileCheck -check-prefixes=GFX950,GFX950-UNSAFE %s
-; RUN: llc -mtriple=amdgcn -mcpu=gfx950 -global-isel=0 -enable-unsafe-fp-math < %s | FileCheck -check-prefixes=GFX950,GFX950-UNSAFE %s
-
-define <2 x half> @v_test_cvt_v2f32_v2f16(<2 x float> %src) {
-; GFX950-LABEL: v_test_cvt_v2f32_v2f16:
-; GFX950:       ; %bb.0:
-; GFX950-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
-; GFX950-NEXT:    v_cvt_pk_f16_f32 v0, v0, v1
-; GFX950-NEXT:    s_setpc_b64 s[30:31]
-  %res = fptrunc <2 x float> %src to <2 x half>
-  ret <2 x half> %res
-}
-
-define <2 x half> @v_test_cvt_v2f64_v2f16(<2 x double> %src) {
-; GFX950-SAFE-SDAG-LABEL: v_test_cvt_v2f64_v2f16:
-; GFX950-SAFE-SDAG:       ; %bb.0:
-; GFX950-SAFE-SDAG-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
-; GFX950-SAFE-SDAG-NEXT:    s_movk_i32 s0, 0x1ff
-; GFX950-SAFE-SDAG-NEXT:    v_and_or_b32 v0, v1, s0, v0
-; GFX950-SAFE-SDAG-NEXT:    v_cmp_ne_u32_e32 vcc, 0, v0
-; GFX950-SAFE-SDAG-NEXT:    v_lshrrev_b32_e32 v4, 8, v1
-; GFX950-SAFE-SDAG-NEXT:    s_movk_i32 s1, 0xffe
-; GFX950-SAFE-SDAG-NEXT:    v_cndmask_b32_e64 v0, 0, 1, vcc
-; GFX950-SAFE-SDAG-NEXT:    v_bfe_u32 v5, v1, 20, 11
-; GFX950-SAFE-SDAG-NEXT:    v_and_or_b32 v0, v4, s1, v0
-; GFX950-SAFE-SDAG-NEXT:    v_sub_u32_e32 v6, 0x3f1, v5
-; GFX950-SAFE-SDAG-NEXT:    v_or_b32_e32 v4, 0x1000, v0
-; GFX950-SAFE-SDAG-NEXT:    v_med3_i32 v6, v6, 0, 13
-; GFX950-SAFE-SDAG-NEXT:    v_lshrrev_b32_e32 v7, v6, v4
-; GFX950-SAFE-SDAG-NEXT:    v_lshlrev_b32_e32 v6, v6, v7
-; GFX950-SAFE-SDAG-NEXT:    v_cmp_ne_u32_e32 vcc, v6, v4
-; GFX950-SAFE-SDAG-NEXT:    v_add_u32_e32 v5, 0xfffffc10, v5
-; GFX950-SAFE-SDAG-NEXT:    v_lshl_or_b32 v6, v5, 12, v0
-; GFX950-SAFE-SDAG-NEXT:    v_cndmask_b32_e64 v4, 0, 1, vcc
-; GFX950-SAFE-SDAG-NEXT:    v_or_b32_e32 v4, v7, v4
-; GFX950-SAFE-SDAG-NEXT:    v_cmp_gt_i32_e32 vcc, 1, v5
-; GFX950-SAFE-SDAG-NEXT:    s_movk_i32 s2, 0x40f
-; GFX950-SAFE-SDAG-NEXT:    v_lshrrev_b32_e32 v1, 16, v1
-; GFX950-SAFE-SDAG-NEXT:    v_cndmask_b32_e32 v4, v6, v4, vcc
-; GFX950-SAFE-SDAG-NEXT:    v_and_b32_e32 v6, 7, v4
-; GFX950-SAFE-SDAG-NEXT:    v_cmp_lt_i32_e32 vcc, 5, v6
-; GFX950-SAFE-SDAG-NEXT:    v_lshrrev_b32_e32 v4, 2, v4
-; GFX950-SAFE-SDAG-NEXT:    s_mov_b32 s3, 0x8000
-; GFX950-SAFE-SDAG-NEXT:    v_cndmask_b32_e64 v7, 0, 1, vcc
-; GFX950-SAFE-SDAG-NEXT:    v_cmp_eq_u32_e32 vcc, 3, v6
-; GFX950-SAFE-SDAG-NEXT:    s_nop 1
-; GFX950-SAFE-SDAG-NEXT:    v_cndmask_b32_e64 v6, 0, 1, vcc
-; GFX950-SAFE-SDAG-NEXT:    v_or_b32_e32 v6, v6, v7
-; GFX950-SAFE-SDAG-NEXT:    v_add_u32_e32 v4, v4, v6
-; GFX950-SAFE-SDAG-NEXT:    v_mov_b32_e32 v6, 0x7c00
-; GFX950-SAFE-SDAG-NEXT:    v_cmp_gt_i32_e32 vcc, 31, v5
-; GFX950-SAFE-SDAG-NEXT:    v_mov_b32_e32 v7, 0x7e00
-; GFX950-SAFE-SDAG-NEXT:    s_nop 0
-; GFX950-SAFE-SDAG-NEXT:    v_cndmask_b32_e32 v4, v6, v4, vcc
-; GFX950-SAFE-SDAG-NEXT:    v_cmp_ne_u32_e32 vcc, 0, v0
-; GFX950-SAFE-SDAG-NEXT:    s_nop 1
-; GFX950-SAFE-SDAG-NEXT:    v_cndmask_b32_e32 v0, v6, v7, vcc
-; GFX950-SAFE-SDAG-NEXT:    v_cmp_eq_u32_e32 vcc, s2, v5
-; GFX950-SAFE-SDAG-NEXT:    s_nop 1
-; GFX950-SAFE-SDAG-NEXT:    v_cndmask_b32_e32 v0, v4, v0, vcc
-; GFX950-SAFE-SDAG-NEXT:    v_and_or_b32 v0, v1, s3, v0
-; GFX950-SAFE-SDAG-NEXT:    v_and_or_b32 v1, v3, s0, v2
-; GFX950-SAFE-SDAG-NEXT:    v_cmp_ne_u32_e32 vcc, 0, v1
-; GFX950-SAFE-SDAG-NEXT:    v_lshrrev_b32_e32 v2, 8, v3
-; GFX950-SAFE-SDAG-NEXT:    v_bfe_u32 v4, v3, 20, 11
-; GFX950-SAFE-SDAG-NEXT:    v_cndmask_b32_e64 v1, 0, 1, vcc
-; GFX950-SAFE-SDAG-NEXT:    v_and_or_b32 v1, v2, s1, v1
-; GFX950-SAFE-SDAG-NEXT:    v_sub_u32_e32 v5, 0x3f1, v4
-; GFX950-SAFE-SDAG-NEXT:    v_or_b32_e32 v2, 0x1000, v1
-; GFX950-SAFE-SDAG-NEXT:    v_med3_i32 v5, v5, 0, 13
-; GFX950-SAFE-SDAG-NEXT:    v_lshrrev_b32_e32 v8, v5, v2
-; GFX950-SAFE-SDAG-NEXT:    v_lshlrev_b32_e32 v5, v5, v8
-; GFX950-SAFE-SDAG-NEXT:    v_cmp_ne_u32_e32 vcc, v5, v2
-; GFX950-SAFE-SDAG-NEXT:    v_add_u32_e32 v4, 0xfffffc10, v4
-; GFX950-SAFE-SDAG-NEXT:    v_lshl_or_b32 v5, v4, 12, v1
-; GFX950-SAFE-SDAG-NEXT:    v_cndmask_b32_e64 v2, 0, 1, vcc
-; GFX950-SAFE-SDAG-NEXT:    v_or_b32_e32 v2, v8, v2
-; GFX950-SAFE-SDAG-NEXT:    v_cmp_gt_i32_e32 vcc, 1, v4
-; GFX950-SAFE-SDAG-NEXT:    s_mov_b32 s0, 0x5040100
-; GFX950-SAFE-SDAG-NEXT:    s_nop 0
-; GFX950-SAFE-SDAG-NEXT:    v_cndmask_b32_e32 v2, v5, v2, vcc
-; GFX950-SAFE-SDAG-NEXT:    v_and_b32_e32 v5, 7, v2
-; GFX950-SAFE-SDAG-NEXT:    v_cmp_lt_i32_e32 vcc, 5, v5
-; GFX950-SAFE-SDAG-NEXT:    v_lshrrev_b32_e32 v2, 2, v2
-; GFX950-SAFE-SDAG-NEXT:    s_nop 0
-; GFX950-SAFE-SDAG-NEXT:    v_cndmask_b32_e64 v8, 0, 1, vcc
-; GFX950-SAFE-SDAG-NEXT:    v_cmp_eq_u32_e32 vcc, 3, v5
-; GFX950-SAFE-SDAG-NEXT:    s_nop 1
-; GFX950-SAFE-SDAG-NEXT:    v_cndmask_b32_e64 v5, 0, 1, vcc
-; GFX950-SAFE-SDAG-NEXT:    v_or_b32_e32 v5, v5, v8
-; GFX950-SAFE-SDAG-NEXT:    v_add_u32_e32 v2, v2, v5
-; GFX950-SAFE-SDAG-NEXT:    v_cmp_gt_i32_e32 vcc, 31, v4
-; GFX950-SAFE-SDAG-NEXT:    s_nop 1
-; GFX950-SAFE-SDAG-NEXT:    v_cndmask_b32_e32 v2, v6, v2, vcc
-; GFX950-SAFE-SDAG-NEXT:    v_cmp_ne_u32_e32 vcc, 0, v1
-; GFX950-SAFE-SDAG-NEXT:    s_nop 1
-; GFX950-SAFE-SDAG-NEXT:    v_cndmask_b32_e32 v1, v6, v7, vcc
-; GFX950-SAFE-SDAG-NEXT:    v_cmp_eq_u32_e32 vcc, s2, v4
-; GFX950-SAFE-SDAG-NEXT:    s_nop 1
-; GFX950-SAFE-SDAG-NEXT:    v_cndmask_b32_e32 v1, v2, v1, vcc
-; GFX950-SAFE-SDAG-NEXT:    v_lshrrev_b32_e32 v2, 16, v3
-; GFX950-SAFE-SDAG-NEXT:    v_and_or_b32 v1, v2, s3, v1
-; GFX950-SAFE-SDAG-NEXT:    v_perm_b32 v0, v1, v0, s0
-; GFX950-SAFE-SDAG-NEXT:    s_setpc_b64 s[30:31]
-;
-; GFX950-SAFE-GISEL-LABEL: v_test_cvt_v2f64_v2f16:
-; GFX950-SAFE-GISEL:       ; %bb.0:
-; GFX950-SAFE-GISEL-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
-; GFX950-SAFE-GISEL-NEXT:    s_movk_i32 s0, 0x1ff
-; GFX950-SAFE-GISEL-NEXT:    v_and_or_b32 v0, v1, s0, v0
-; GFX950-SAFE-GISEL-NEXT:    v_cmp_ne_u32_e32 vcc, 0, v0
-; GFX950-SAFE-GISEL-NEXT:    v_lshrrev_b32_e32 v4, 8, v1
-; GFX950-SAFE-GISEL-NEXT:    s_movk_i32 s1, 0xffe
-; GFX950-SAFE-GISEL-NEXT:    v_cndmask_b32_e64 v0, 0, 1, vcc
-; GFX950-SAFE-GISEL-NEXT:    v_bfe_u32 v5, v1, 20, 11
-; GFX950-SAFE-GISEL-NEXT:    v_and_or_b32 v0, v4, s1, v0
-; GFX950-SAFE-GISEL-NEXT:    v_sub_u32_e32 v6, 0x3f1, v5
-; GFX950-SAFE-GISEL-NEXT:    v_or_b32_e32 v4, 0x1000, v0
-; GFX950-SAFE-GISEL-NEXT:    v_med3_i32 v6, v6, 0, 13
-; GFX950-SAFE-GISEL-NEXT:    v_lshrrev_b32_e32 v7, v6, v4
-; GFX950-SAFE-GISEL-NEXT:    v_lshlrev_b32_e32 v6, v6, v7
-; GFX950-SAFE-GISEL-NEXT:    v_cmp_ne_u32_e32 vcc, v6, v4
-; GFX950-SAFE-GISEL-NEXT:    v_add_u32_e32 v5, 0xfffffc10, v5
-; GFX950-SAFE-GISEL-NEXT:    v_lshl_or_b32 v6, v5, 12, v0
-; GFX950-SAFE-GISEL-NEXT:    v_cndmask_b32_e64 v4, 0, 1, vcc
-; GFX950-SAFE-GISEL-NEXT:    v_or_b32_e32 v4, v7, v4
-; GFX950-SAFE-GISEL-NEXT:    v_cmp_gt_i32_e32 vcc, 1, v5
-; GFX950-SAFE-GISEL-NEXT:    s_movk_i32 s2, 0x40f
-; GFX950-SAFE-GISEL-NEXT:    v_lshrrev_b32_e32 v1, 16, v1
-; GFX950-SAFE-GISEL-NEXT:    v_cndmask_b32_e32 v4, v6, v4, vcc
-; GFX950-SAFE-GISEL-NEXT:    v_and_b32_e32 v6, 7, v4
-; GFX950-SAFE-GISEL-NEXT:    v_cmp_lt_i32_e32 vcc, 5, v6
-; GFX950-SAFE-GISEL-NEXT:    v_lshrrev_b32_e32 v4, 2, v4
-; GFX950-SAFE-GISEL-NEXT:    s_mov_b32 s3, 0x8000
-; GFX950-SAFE-GISEL-NEXT:    v_cndmask_b32_e64 v7, 0, 1, vcc
-; GFX950-SAFE-GISEL-NEXT:    v_cmp_eq_u32_e32 vcc, 3, v6
-; GFX950-SAFE-GISEL-NEXT:    s_nop 1
-; GFX950-SAFE-GISEL-NEXT:    v_cndmask_b32_e64 v6, 0, 1, vcc
-; GFX950-SAFE-GISEL-NEXT:    v_or_b32_e32 v6, v6, v7
-; GFX950-SAFE-GISEL-NEXT:    v_add_u32_e32 v4, v4, v6
-; GFX950-SAFE-GISEL-NEXT:    v_mov_b32_e32 v6, 0x7c00
-; GFX950-SAFE-GISEL-NEXT:    v_cmp_gt_i32_e32 vcc, 31, v5
-; GFX950-SAFE-GISEL-NEXT:    v_mov_b32_e32 v7, 0x7e00
-; GFX950-SAFE-GISEL-NEXT:    s_nop 0
-; GFX950-SAFE-GISEL-NEXT:    v_cndmask_b32_e32 v4, v6, v4, vcc
-; GFX950-SAFE-GISEL-NEXT:    v_cmp_ne_u32_e32 vcc, 0, v0
-; GFX950-SAFE-GISEL-NEXT:    s_nop 1
-; GFX950-SAFE-GISEL-NEXT:    v_cndmask_b32_e32 v0, v6, v7, vcc
-; GFX950-SAFE-GISEL-NEXT:    v_cmp_eq_u32_e32 vcc, s2, v5
-; GFX950-SAFE-GISEL-NEXT:    s_nop 1
-; GFX950-SAFE-GISEL-NEXT:    v_cndmask_b32_e32 v0, v4, v0, vcc
-; GFX950-SAFE-GISEL-NEXT:    v_and_or_b32 v0, v1, s3, v0
-; GFX950-SAFE-GISEL-NEXT:    v_and_or_b32 v1, v3, s0, v2
-; GFX950-SAFE-GISEL-NEXT:    v_cmp_ne_u32_e32 vcc, 0, v1
-; GFX950-SAFE-GISEL-NEXT:    v_lshrrev_b32_e32 v2, 8, v3
-; GFX950-SAFE-GISEL-NEXT:    v_bfe_u32 v4, v3, 20, 11
-; GFX950-SAFE-GISEL-NEXT:    v_cndmask_b32_e64 v1, 0, 1, vcc
-; GFX950-SAFE-GISEL-NEXT:    v_and_or_b32 v1, v2, s1, v1
-; GFX950-SAFE-GISEL-NEXT:    v_sub_u32_e32 v5, 0x3f1, v4
-; GFX950-SAFE-GISEL-NEXT:    v_or_b32_e32 v2, 0x1000, v1
-; GFX950-SAFE-GISEL-NEXT:    v_med3_i32 v5, v5, 0, 13
-; GFX950-SAFE-GISEL-NEXT:    v_lshrrev_b32_e32 v8, v5, v2
-; GFX950-SAFE-GISEL-NEXT:    v_lshlrev_b32_e32 v5, v5, v8
-; GFX950-SAFE-GISEL-NEXT:    v_cmp_ne_u32_e32 vcc, v5, v2
-; GFX950-SAFE-GISEL-NEXT:    v_add_u32_e32 v4, 0xfffffc10, v4
-; GFX950-SAFE-GISEL-NEXT:    v_lshl_or_b32 v5, v4, 12, v1
-; GFX950-SAFE-GISEL-NEXT:    v_cndmask_b32_e64 v2, 0, 1, vcc
-; GFX950-SAFE-GISEL-NEXT:    v_or_b32_e32 v2, v8, v2
-; GFX950-SAFE-GISEL-NEXT:    v_cmp_gt_i32_e32 vcc, 1, v4
-; GFX950-SAFE-GISEL-NEXT:    s_mov_b32 s0, 0x5040100
-; GFX950-SAFE-GISEL-NEXT:    s_nop 0
-; GFX950-SAFE-GISEL-NEXT:    v_cndmask_b32_e32 v2, v5, v2, vcc
-; GFX950-SAFE-GISEL-NEXT:    v_and_b32_e32 v5, 7, v2
-; GFX950-SAFE-GISEL-NEXT:    v_cmp_lt_i32_e32 vcc, 5, v5
-; GFX950-SAFE-GISEL-NEXT:    v_lshrrev_b32_e32 v2, 2, v2
-; GFX950-SAFE-GISEL-NEXT:    s_nop 0
-; GFX950-SAFE-GISEL-NEXT:    v_cndmask_b32_e64 v8, 0, 1, vcc
-; GFX950-SAFE-GISEL-NEXT:    v_cmp_eq_u32_e32 vcc, 3, v5
-; GFX950-SAFE-GISEL-NEXT:    s_nop 1
-; GFX950-SAFE-GISEL-NEXT:    v_cndmask_b32_e64 v5, 0, 1, vcc
-; GFX950-SAFE-GISEL-NEXT:    v_or_b32_e32 v5, v5, v8
-; GFX950-SAFE-GISEL-NEXT:    v_add_u32_e32 v2, v2, v5
-; GFX950-SAFE-GISEL-NEXT:    v_cmp_gt_i32_e32 vcc, 31, v4
-; GFX950-SAFE-GISEL-NEXT:    s_nop 1
-; GFX950-SAFE-GISEL-NEXT:    v_cndmask_b32_e32 v2, v6, v2, vcc
-; GFX950-SAFE-GISEL-NEXT:    v_cmp_ne_u32_e32 vcc, 0, v1
-; GFX950-SAFE-GISEL-NEXT:    s_nop 1
-; GFX950-SAFE-GISEL-NEXT:    v_cndmask_b32_e32 v1, v6, v7, vcc
-; GFX950-SAFE-GISEL-NEXT:    v_cmp_eq_u32_e32 vcc, s2, v4
-; GFX950-SAFE-GISEL-NEXT:    s_nop 1
-; GFX950-SAFE-GISEL-NEXT:    v_cndmask_b32_e32 v1, v2, v1, vcc
-; GFX950-SAFE-GISEL-NEXT:    v_lshrrev_b32_e32 v2, 16, v3
-; GFX950-SAFE-GISEL-NEXT:    v_and_or_b32 v1, v2, s3, v1
-; GFX950-SAFE-GISEL-NEXT:    v_perm_b32 v0, v1, v0, s0
-; GFX950-SAFE-GISEL-NEXT:    s_setpc_b64 s[30:31]
-;
-; GFX950-UNSAFE-LABEL: v_test_cvt_v2f64_v2f16:
-; GFX950-UNSAFE:       ; %bb.0:
-; GFX950-UNSAFE-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
-; GFX950-UNSAFE-NEXT:    v_cvt_f32_f64_e32 v0, v[0:1]
-; GFX950-UNSAFE-NEXT:    v_cvt_f32_f64_e32 v1, v[2:3]
-; GFX950-UNSAFE-NEXT:    v_cvt_f16_f32_e32 v0, v0
-; GFX950-UNSAFE-NEXT:    v_cvt_f16_f32_e32 v1, v1
-; GFX950-UNSAFE-NEXT:    s_mov_b32 s0, 0x5040100
-; GFX950-UNSAFE-NEXT:    v_perm_b32 v0, v1, v0, s0
-; GFX950-UNSAFE-NEXT:    s_setpc_b64 s[30:31]
-  %res = fptrunc <2 x double> %src to <2 x half>
-  ret <2 x half> %res
-}
-
diff --git a/llvm/test/CodeGen/AMDGPU/fptrunc.v2f16.no.fast.path.ll b/llvm/test/CodeGen/AMDGPU/fptrunc.v2f16.no.fast.path.ll
new file mode 100644
index 0000000000000..5a5e39489d888
--- /dev/null
+++ b/llvm/test/CodeGen/AMDGPU/fptrunc.v2f16.no.fast.path.ll
@@ -0,0 +1,193 @@
+; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py UTC_ARGS: --version 5
+; RUN: llc -mtriple=amdgcn -mcpu=gfx950 -global-isel=0 < %s | FileCheck -check-prefixes=GFX950,GFX950-SDAG %s
+; RUN: llc -mtriple=amdgcn -mcpu=gfx950 -global-isel=1 < %s | FileCheck -check-prefixes=GFX950,GFX950-GISEL %s
+
+define <2 x half> @v_test_cvt_v2f32_v2f16(<2 x float> %src) {
+; GFX950-LABEL: v_test_cvt_v2f32_v2f16:
+; GFX950:       ; %bb.0:
+; GFX950-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
+; GFX950-NEXT:    v_cvt_pk_f16_f32 v0, v0, v1
+; GFX950-NEXT:    s_setpc_b64 s[30:31]
+  %res = fptrunc <2 x float> %src to <2 x half>
+  ret <2 x half> %res
+}
+
+define <2 x half> @v_test_cvt_v2f64_v2f16(<2 x double> %src) {
+; GFX950-SDAG-LABEL: v_test_cvt_v2f64_v2f16:
+; GFX950-SDAG:       ; %bb.0:
+; GFX950-SDAG-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
+; GFX950-SDAG-NEXT:    s_movk_i32 s0, 0x1ff
+; GFX950-SDAG-NEXT:    v_and_or_b32 v0, v1, s0, v0
+; GFX950-SDAG-NEXT:    v_cmp_ne_u32_e32 vcc, 0, v0
+; GFX950-SDAG-NEXT:    v_lshrrev_b32_e32 v4, 8, v1
+; GFX950-SDAG-NEXT:    s_movk_i32 s1, 0xffe
+; GFX950-SDAG-NEXT:    v_cndmask_b32_e64 v0, 0, 1, vcc
+; GFX950-SDAG-NEXT:    v_bfe_u32 v5, v1, 20, 11
+; GFX950-SDAG-NEXT:    v_and_or_b32 v0, v4, s1, v0
+; GFX950-SDAG-NEXT:    v_sub_u32_e32 v6, 0x3f1, v5
+; GFX950-SDAG-NEXT:    v_or_b32_e32 v4, 0x1000, v0
+; GFX950-SDAG-NEXT:    v_med3_i32 v6, v6, 0, 13
+; GFX950-SDAG-NEXT:    v_lshrrev_b32_e32 v7, v6, v4
+; GFX950-SDAG-NEXT:    v_lshlrev_b32_e32 v6, v6, v7
+; GFX950-SDAG-NEXT:    v_cmp_ne_u32_e32 vcc, v6, v4
+; GFX950-SDAG-NEXT:    v_add_u32_e32 v5, 0xfffffc10, v5
+; GFX950-SDAG-NEXT:    v_lshl_or_b32 v6, v5, 12, v0
+; GFX950-SDAG-NEXT:    v_cndmask_b32_e64 v4, 0, 1, vcc
+; GFX950-SDAG-NEXT:    v_or_b32_e32 v4, v7, v4
+; GFX950-SDAG-NEXT:    v_cmp_gt_i32_e32 vcc, 1, v5
+; GFX950-SDAG-NEXT:    s_movk_i32 s2, 0x40f
+; GFX950-SDAG-NEXT:    v_lshrrev_b32_e32 v1, 16, v1
+; GFX950-SDAG-NEXT:    v_cndmask_b32_e32 v4, v6, v4, vcc
+; GFX950-SDAG-NEXT:    v_and_b32_e32 v6, 7, v4
+; GFX950-SDAG-NEXT:    v_cmp_lt_i32_e32 vcc, 5, v6
+; GFX950-SDAG-NEXT:    v_lshrrev_b32_e32 v4, 2, v4
+; GFX950-SDAG-NEXT:    s_mov_b32 s3, 0x8000
+; GFX950-SDAG-NEXT:    v_cndmask_b32_e64 v7, 0, 1, vcc
+; GFX950-SDAG-NEXT:    v_cmp_eq_u32_e32 vcc, 3, v6
+; GFX950-SDAG-NEXT:    s_nop 1
+; GFX950-SDAG-NEXT:    v_cndmask_b32_e64 v6, 0, 1, vcc
+; GFX950-SDAG-NEXT:    v_or_b32_e32 v6, v6, v7
+; GFX950-SDAG-NEXT:    v_add_u32_e32 v4, v4, v6
+; GFX950-SDAG-NEXT:    v_mov_b32_e32 v6, 0x7c00
+; GFX950-SDAG-NEXT:    v_cmp_gt_i32_e32 vcc, 31, v5
+; GFX950-SDAG-NEXT:    v_mov_b32_e32 v7, 0x7e00
+; GFX950-SDAG-NEXT:    s_nop 0
+; GFX950-SDAG-NEXT:    v_cndmask_b32_e32 v4, v6, v4, vcc
+; GFX950-SDAG-NEXT:    v_cmp_ne_u32_e32 vcc, 0, v0
+; GFX950-SDAG-NEXT:    s_nop 1
+; GFX950-SDAG-NEXT:    v_cndmask_b32_e32 v0, v6, v7, vcc
+; GFX950-SDAG-NEXT:    v_cmp_eq_u32_e32 vcc, s2, v5
+; GFX950-SDAG-NEXT:    s_nop 1
+; GFX950-SDAG-NEXT:    v_cndmask_b32_e32 v0, v4, v0, vcc
+; GFX950-SDAG-NEXT:    v_and_or_b32 v0, v1, s3, v0
+; GFX950-SDAG-NEXT:    v_and_or_b32 v1, v3, s0, v2
+; GFX950-SDAG-NEXT:    v_cmp_ne_u32_e32 vcc, 0, v1
+; GFX950-SDAG-NEXT:    v_lshrrev_b32_e32 v2, 8, v3
+; GFX950-SDAG-NEXT:    v_bfe_u32 v4, v3, 20, 11
+; GFX950-SDAG-NEXT:    v_cndmask_b32_e64 v1, 0, 1, vcc
+; GFX950-SDAG-NEXT:    v_and_or_b32 v1, v2, s1, v1
+; GFX950-SDAG-NEXT:    v_sub_u32_e32 v5, 0x3f1, v4
+; GFX950-SDAG-NEXT:    v_or_b32_e32 v2, 0x1000, v1
+; GFX950-SDAG-NEXT:    v_med3_i32 v5, v5, 0, 13
+; GFX950-SDAG-NEXT:    v_lshrrev_b32_e32 v8, v5, v2
+; GFX950-SDAG-NEXT:    v_lshlrev_b32_e32 v5, v5, v8
+; GFX950-SDAG-NEXT:    v_cmp_ne_u32_e32 vcc, v5, v2
+; GFX950-SDAG-NEXT:    v_add_u32_e32 v4, 0xfffffc10, v4
+; GFX950-SDAG-NEXT:    v_lshl_or_b32 v5, v4, 12, v1
+; GFX950-SDAG-NEXT:    v_cndmask_b32_e64 v2, 0, 1, vcc
+; GFX950-SDAG-NEXT:    v_or_b32_e32 v2, v8, v2
+; GFX950-SDAG-NEXT:    v_cmp_gt_i32_e32 vcc, 1, v4
+; GFX950-SDAG-NEXT:    s_mov_b32 s0, 0x5040100
+; GFX950-SDAG-NEXT:    s_nop 0
+; GFX950-SDAG-NEXT:    v_cndmask_b32_e32 v2, v5, v2, vcc
+; GFX950-SDAG-NEXT:    v_and_b32_e32 v5, 7, v2
+; GFX950-SDAG-NEXT:    v_cmp_lt_i32_e32 vcc, 5, v5
+; GFX950-SDAG-NEXT:    v_lshrrev_b32_e32 v2, 2, v2
+; GFX950-SDAG-NEXT:    s_nop 0
+; GFX950-SDAG-NEXT:    v_cndmask_b32_e64 v8, 0, 1, vcc
+; GFX950-SDAG-NEXT:    v_cmp_eq_u32_e32 vcc, 3, v5
+; GFX950-SDAG-NEXT:    s_nop 1
+; GFX950-SDAG-NEXT:    v_cndmask_b32_e64 v5, 0, 1, vcc
+; GFX950-SDAG-NEXT:    v_or_b32_e32 v5, v5, v8
+; GFX950-SDAG-NEXT:    v_add_u32_e32 v2, v2, v5
+; GFX950-SDAG-NEXT:    v_cmp_gt_i32_e32 vcc, 31, v4
+; GFX950-SDAG-NEXT:    s_nop 1
+; GFX950-SDAG-NEXT:    v_cndmask_b32_e32 v2, v6, v2, vcc
+; GFX950-SDAG-NEXT:    v_cmp_ne_u32_e32 vcc, 0, v1
+; GFX950-SDAG-NEXT:    s_nop 1
+; GFX950-SDAG-NEXT:    v_cndmask_b32_e32 v1, v6, v7, vcc
+; GFX950-SDAG-NEXT:    v_cmp_eq_u32_e32 vcc, s2, v4
+; GFX950-SDAG-NEXT:    s_nop 1
+; GFX950-SDAG-NEXT:    v_cndmask_b32_e32 v1, v2, v1, vcc
+; GFX950-SDAG-NEXT:    v_lshrrev_b32_e32 v2, 16, v3
+; GFX950-SDAG-NEXT:    v_and_or_b32 v1, v2, s3, v1
+; GFX950-SDAG-NEXT:    v_perm_b32 v0, v1, v0, s0
+; GFX950-SDAG-NEXT:    s_setpc_b64 s[30:31]
+;
+; GFX950-GISEL-LABEL: v_test_cvt_v2f64_v2f16:
+; GFX950-GISEL:       ; %bb.0:
+; GFX950-GISEL-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
+; GFX950-GISEL-NEXT:    v_mov_b32_e32 v7, 0x1ff
+; GFX950-GISEL-NEXT:    v_and_or_b32 v0, v1, v7, v0
+; GFX950-GISEL-NEXT:    v_bfe_u32 v4, v1, 20, 11
+; GFX950-GISEL-NEXT:    v_cmp_ne_u32_e32 vcc, 0, v0
+; GFX950-GISEL-NEXT:    v_add_u32_e32 v4, 0xfffffc10, v4
+; GFX950-GISEL-NEXT:    v_lshrrev_b32_e32 v5, 8, v1
+; GFX950-GISEL-NEXT:    v_mov_b32_e32 v6, 0xffe
+; GFX950-GISEL-NEXT:    v_cndmask_b32_e64 v0, 0, 1, vcc
+; GFX950-GISEL-NEXT:    v_and_or_b32 v0, v5, v6, v0
+; GFX950-GISEL-NEXT:    v_sub_u32_e32 v10, 1, v4
+; GFX950-GISEL-NEXT:    v_cmp_ne_u32_e32 vcc, 0, v0
+; GFX950-GISEL-NEXT:    v_lshl_or_b32 v9, v4, 12, v0
+; GFX950-GISEL-NEXT:    v_med3_i32 v10, v10, 0, 13
+; GFX950-GISEL-NEXT:    v_or_b32_e32 v0, 0x1000, v0
+; GFX950-GISEL-NEXT:    v_lshrrev_b32_e32 v11, v10, v0
+; GFX950-GISEL-NEXT:    v_lshlrev_b32_e32 v10, v10, v11
+; GFX950-GISEL-NEXT:    v_cndmask_b32_e64 v5, 0, 1, vcc
+; GFX950-GISEL-NEXT:    v_cmp_ne_u32_e32 vcc, v10, v0
+; GFX950-GISEL-NEXT:    v_mov_b32_e32 v8, 0x7c00
+; GFX950-GISEL-NEXT:    v_lshl_or_b32 v5, v5, 9, v8
+; GFX950-GISEL-NEXT:    v_cndmask_b32_e64 v0, 0, 1, vcc
+; GFX950-GISEL-NEXT:    v_or_b32_e32 v0, v11, v0
+; GFX950-GISEL-NEXT:    v_cmp_gt_i32_e32 vcc, 1, v4
+; GFX950-GISEL-NEXT:    v_lshrrev_b32_e32 v1, 16, v1
+; GFX950-GISEL-NEXT:    v_and_or_b32 v2, v3, v7, v2
+; GFX950-GISEL-NEXT:    v_cndmask_b32_e32 v0, v9, v0, vcc
+; GFX950-GISEL-NEXT:    v_and_b32_e32 v9, 7, v0
+; GFX950-GISEL-NEXT:    v_cmp_eq_u32_e32 vcc, 3, v9
+; GFX950-GISEL-NEXT:    v_cmp_lt_i32_e64 s[0:1], 5, v9
+; GFX950-GISEL-NEXT:    s_or_b64 s[0:1], vcc, s[0:1]
+; GFX950-GISEL-NEXT:    v_lshrrev_b32_e32 v0, 2, v0
+; GFX950-GISEL-NEXT:    v_cndmask_b32_e64 v9, 0, 1, s[0:1]
+; GFX950-GISEL-NEXT:    v_add_u32_e32 v0, v0, v9
+; GFX950-GISEL-NEXT:    v_cmp_lt_i32_e32 vcc, 30, v4
+; GFX950-GISEL-NEXT:    v_mov_b32_e32 v9, 0x40f
+; GFX950-GISEL-NEXT:    s_nop 0
+; GFX950-GISEL-NEXT:    v_cndmask_b32_e32 v0, v0, v8, vcc
+; GFX950-GISEL-NEXT:    v_cmp_eq_u32_e32 vcc, v4, v9
+; GFX950-GISEL-NEXT:    v_mov_b32_e32 v4, 0x8000
+; GFX950-GISEL-NEXT:    s_nop 0
+; GFX950-GISEL-NEXT:    v_cndmask_b32_e32 v0, v0, v5, vcc
+; GFX950-GISEL-NEXT:    v_and_or_b32 v0, v1, v4, v0
+; GFX950-GISEL-NEXT:    v_bfe_u32 v1, v3, 20, 11
+; GFX950-GISEL-NEXT:    v_cmp_ne_u32_e32 vcc, 0, v2
+; GFX950-GISEL-NEXT:    v_add_u32_e32 v1, 0xfffffc10, v1
+; GFX950-GISEL-NEXT:    v_lshrrev_b32_e32 v5, 8, v3
+; GFX950-GISEL-NEXT:    v_cndmask_b32_e64 v2, 0, 1, vcc
+; GFX950-GISEL-NEXT:    v_and_or_b32 v2, v5, v6, v2
+; GFX950-GISEL-NEXT:    v_sub_u32_e32 v7, 1, v1
+; GFX950-GISEL-NEXT:    v_cmp_ne_u32_e32 vcc, 0, v2
+; GFX950-GISEL-NEXT:    v_lshl_or_b32 v6, v1, 12, v2
+; GFX950-GISEL-NEXT:    v_med3_i32 v7, v7, 0, 13
+; GFX950-GISEL-NEXT:    v_or_b32_e32 v2, 0x1000, v2
+; GFX950-GISEL-NEXT:    v_lshrrev_b32_e32 v10, v7, v2
+; GFX950-GISEL-NEXT:    v_lshlrev_b32_e32 v7, v7, v10
+; GFX950-GISEL-NEXT:    v_cndmask_b32_e64 v5, 0, 1, vcc
+; GFX950-GISEL-NEXT:    v_cmp_ne_u32_e32 vcc, v7...
[truncated]

@changpeng changpeng merged commit a945f59 into llvm:main Apr 21, 2025
13 checks passed
@changpeng changpeng deleted the f161 branch April 21, 2025 21:38
IanWood1 pushed a commit to IanWood1/llvm-project that referenced this pull request May 6, 2025
…vm#136609)

Also remove the checks with -enable-unsafe-fp-math (already in fptrunc.f16.ll)
IanWood1 pushed a commit to IanWood1/llvm-project that referenced this pull request May 6, 2025
…vm#136609)

Also remove the checks with -enable-unsafe-fp-math (already in fptrunc.f16.ll)
IanWood1 pushed a commit to IanWood1/llvm-project that referenced this pull request May 6, 2025
…vm#136609)

Also remove the checks with -enable-unsafe-fp-math (already in fptrunc.f16.ll)
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Projects
None yet
Development

Successfully merging this pull request may close these issues.

2 participants