[llvm] [X86][NFC] Added POWI function testcases (PR #134276)
via llvm-commits
llvm-commits at lists.llvm.org
Thu Apr 3 10:15:02 PDT 2025
https://github.com/JaydeepChauhan14 updated https://github.com/llvm/llvm-project/pull/134276
>From 8f9018bb0b6c1288270742b967bfaf4018146c5b Mon Sep 17 00:00:00 2001
From: Chauhan Jaydeep Ashwinbhai <chauhan.jaydeep.ashwinbhai at intel.com>
Date: Thu, 3 Apr 2025 10:04:05 -0700
Subject: [PATCH 1/2] [X86][NFC] Added POWI function testcases
---
llvm/test/CodeGen/X86/powi-const.ll | 268 ++++++++++++++++++
llvm/test/CodeGen/X86/powi.ll | 411 +++++++++++-----------------
2 files changed, 427 insertions(+), 252 deletions(-)
create mode 100644 llvm/test/CodeGen/X86/powi-const.ll
diff --git a/llvm/test/CodeGen/X86/powi-const.ll b/llvm/test/CodeGen/X86/powi-const.ll
new file mode 100644
index 0000000000000..7dc6564e62a85
--- /dev/null
+++ b/llvm/test/CodeGen/X86/powi-const.ll
@@ -0,0 +1,268 @@
+; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
+; RUN: llc < %s -mtriple=i686-unknown-unknown | FileCheck %s --check-prefix=X86-X87
+; RUN: llc < %s -mtriple=i686-unknown-unknown -mattr=+sse2 | FileCheck %s --check-prefix=X86-SSE
+; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+sse2 | FileCheck %s --check-prefix=X64
+
+; Ideally this would compile to 5 multiplies.
+
+define double @pow_wrapper(double %a) nounwind readonly ssp noredzone {
+; X86-X87-LABEL: pow_wrapper:
+; X86-X87: # %bb.0:
+; X86-X87-NEXT: fldl {{[0-9]+}}(%esp)
+; X86-X87-NEXT: fld %st(0)
+; X86-X87-NEXT: fmul %st(1), %st
+; X86-X87-NEXT: fmul %st, %st(1)
+; X86-X87-NEXT: fmul %st, %st(0)
+; X86-X87-NEXT: fmul %st, %st(1)
+; X86-X87-NEXT: fmul %st, %st(0)
+; X86-X87-NEXT: fmulp %st, %st(1)
+; X86-X87-NEXT: retl
+;
+; X86-SSE-LABEL: pow_wrapper:
+; X86-SSE: # %bb.0:
+; X86-SSE-NEXT: pushl %ebp
+; X86-SSE-NEXT: movl %esp, %ebp
+; X86-SSE-NEXT: andl $-8, %esp
+; X86-SSE-NEXT: subl $8, %esp
+; X86-SSE-NEXT: movsd {{.*#+}} xmm0 = mem[0],zero
+; X86-SSE-NEXT: movapd %xmm0, %xmm1
+; X86-SSE-NEXT: mulsd %xmm0, %xmm1
+; X86-SSE-NEXT: mulsd %xmm1, %xmm0
+; X86-SSE-NEXT: mulsd %xmm1, %xmm1
+; X86-SSE-NEXT: mulsd %xmm1, %xmm0
+; X86-SSE-NEXT: mulsd %xmm1, %xmm1
+; X86-SSE-NEXT: mulsd %xmm0, %xmm1
+; X86-SSE-NEXT: movsd %xmm1, (%esp)
+; X86-SSE-NEXT: fldl (%esp)
+; X86-SSE-NEXT: movl %ebp, %esp
+; X86-SSE-NEXT: popl %ebp
+; X86-SSE-NEXT: retl
+;
+; X64-LABEL: pow_wrapper:
+; X64: # %bb.0:
+; X64-NEXT: movapd %xmm0, %xmm1
+; X64-NEXT: mulsd %xmm0, %xmm1
+; X64-NEXT: mulsd %xmm1, %xmm0
+; X64-NEXT: mulsd %xmm1, %xmm1
+; X64-NEXT: mulsd %xmm1, %xmm0
+; X64-NEXT: mulsd %xmm1, %xmm1
+; X64-NEXT: mulsd %xmm1, %xmm0
+; X64-NEXT: retq
+ %ret = tail call double @llvm.powi.f64.i32(double %a, i32 15) nounwind ; <double> [#uses=1]
+ ret double %ret
+}
+
+define double @pow_wrapper_optsize(double %a) optsize {
+; X86-X87-LABEL: pow_wrapper_optsize:
+; X86-X87: # %bb.0:
+; X86-X87-NEXT: subl $12, %esp
+; X86-X87-NEXT: .cfi_def_cfa_offset 16
+; X86-X87-NEXT: fldl {{[0-9]+}}(%esp)
+; X86-X87-NEXT: fstpl (%esp)
+; X86-X87-NEXT: movl $15, {{[0-9]+}}(%esp)
+; X86-X87-NEXT: calll __powidf2
+; X86-X87-NEXT: addl $12, %esp
+; X86-X87-NEXT: .cfi_def_cfa_offset 4
+; X86-X87-NEXT: retl
+;
+; X86-SSE-LABEL: pow_wrapper_optsize:
+; X86-SSE: # %bb.0:
+; X86-SSE-NEXT: subl $12, %esp
+; X86-SSE-NEXT: .cfi_def_cfa_offset 16
+; X86-SSE-NEXT: movsd {{.*#+}} xmm0 = mem[0],zero
+; X86-SSE-NEXT: movsd %xmm0, (%esp)
+; X86-SSE-NEXT: movl $15, {{[0-9]+}}(%esp)
+; X86-SSE-NEXT: calll __powidf2
+; X86-SSE-NEXT: addl $12, %esp
+; X86-SSE-NEXT: .cfi_def_cfa_offset 4
+; X86-SSE-NEXT: retl
+;
+; X64-LABEL: pow_wrapper_optsize:
+; X64: # %bb.0:
+; X64-NEXT: movl $15, %edi
+; X64-NEXT: jmp __powidf2 at PLT # TAILCALL
+ %ret = tail call double @llvm.powi.f64.i32(double %a, i32 15) nounwind ; <double> [#uses=1]
+ ret double %ret
+}
+
+define double @pow_wrapper_pgso(double %a) !prof !14 {
+; X86-X87-LABEL: pow_wrapper_pgso:
+; X86-X87: # %bb.0:
+; X86-X87-NEXT: subl $12, %esp
+; X86-X87-NEXT: .cfi_def_cfa_offset 16
+; X86-X87-NEXT: fldl {{[0-9]+}}(%esp)
+; X86-X87-NEXT: fstpl (%esp)
+; X86-X87-NEXT: movl $15, {{[0-9]+}}(%esp)
+; X86-X87-NEXT: calll __powidf2
+; X86-X87-NEXT: addl $12, %esp
+; X86-X87-NEXT: .cfi_def_cfa_offset 4
+; X86-X87-NEXT: retl
+;
+; X86-SSE-LABEL: pow_wrapper_pgso:
+; X86-SSE: # %bb.0:
+; X86-SSE-NEXT: subl $12, %esp
+; X86-SSE-NEXT: .cfi_def_cfa_offset 16
+; X86-SSE-NEXT: movsd {{.*#+}} xmm0 = mem[0],zero
+; X86-SSE-NEXT: movsd %xmm0, (%esp)
+; X86-SSE-NEXT: movl $15, {{[0-9]+}}(%esp)
+; X86-SSE-NEXT: calll __powidf2
+; X86-SSE-NEXT: addl $12, %esp
+; X86-SSE-NEXT: .cfi_def_cfa_offset 4
+; X86-SSE-NEXT: retl
+;
+; X64-LABEL: pow_wrapper_pgso:
+; X64: # %bb.0:
+; X64-NEXT: movl $15, %edi
+; X64-NEXT: jmp __powidf2 at PLT # TAILCALL
+ %ret = tail call double @llvm.powi.f64.i32(double %a, i32 15) nounwind ; <double> [#uses=1]
+ ret double %ret
+}
+
+define double @pow_wrapper_minsize(double %a) minsize {
+; X86-X87-LABEL: pow_wrapper_minsize:
+; X86-X87: # %bb.0:
+; X86-X87-NEXT: subl $12, %esp
+; X86-X87-NEXT: .cfi_def_cfa_offset 16
+; X86-X87-NEXT: fldl {{[0-9]+}}(%esp)
+; X86-X87-NEXT: fstpl (%esp)
+; X86-X87-NEXT: movl $15, {{[0-9]+}}(%esp)
+; X86-X87-NEXT: calll __powidf2
+; X86-X87-NEXT: addl $12, %esp
+; X86-X87-NEXT: .cfi_def_cfa_offset 4
+; X86-X87-NEXT: retl
+;
+; X86-SSE-LABEL: pow_wrapper_minsize:
+; X86-SSE: # %bb.0:
+; X86-SSE-NEXT: subl $12, %esp
+; X86-SSE-NEXT: .cfi_def_cfa_offset 16
+; X86-SSE-NEXT: movsd {{.*#+}} xmm0 = mem[0],zero
+; X86-SSE-NEXT: movsd %xmm0, (%esp)
+; X86-SSE-NEXT: movl $15, {{[0-9]+}}(%esp)
+; X86-SSE-NEXT: calll __powidf2
+; X86-SSE-NEXT: addl $12, %esp
+; X86-SSE-NEXT: .cfi_def_cfa_offset 4
+; X86-SSE-NEXT: retl
+;
+; X64-LABEL: pow_wrapper_minsize:
+; X64: # %bb.0:
+; X64-NEXT: pushq $15
+; X64-NEXT: .cfi_adjust_cfa_offset 8
+; X64-NEXT: popq %rdi
+; X64-NEXT: .cfi_adjust_cfa_offset -8
+; X64-NEXT: jmp __powidf2 at PLT # TAILCALL
+ %ret = tail call double @llvm.powi.f64.i32(double %a, i32 15) nounwind ; <double> [#uses=1]
+ ret double %ret
+}
+
+define <2 x float> @powi_v2f32(<2 x float> %a) minsize {
+; X86-X87-LABEL: powi_v2f32:
+; X86-X87: # %bb.0:
+; X86-X87-NEXT: pushl %esi
+; X86-X87-NEXT: .cfi_def_cfa_offset 8
+; X86-X87-NEXT: subl $16, %esp
+; X86-X87-NEXT: .cfi_def_cfa_offset 24
+; X86-X87-NEXT: .cfi_offset %esi, -8
+; X86-X87-NEXT: flds {{[0-9]+}}(%esp)
+; X86-X87-NEXT: fstps {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Folded Spill
+; X86-X87-NEXT: flds {{[0-9]+}}(%esp)
+; X86-X87-NEXT: pushl $15
+; X86-X87-NEXT: .cfi_adjust_cfa_offset 4
+; X86-X87-NEXT: popl %esi
+; X86-X87-NEXT: .cfi_adjust_cfa_offset -4
+; X86-X87-NEXT: movl %esi, {{[0-9]+}}(%esp)
+; X86-X87-NEXT: fstps (%esp)
+; X86-X87-NEXT: calll __powisf2
+; X86-X87-NEXT: fstps {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Folded Spill
+; X86-X87-NEXT: movl %esi, {{[0-9]+}}(%esp)
+; X86-X87-NEXT: flds {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Folded Reload
+; X86-X87-NEXT: fstps (%esp)
+; X86-X87-NEXT: calll __powisf2
+; X86-X87-NEXT: flds {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Folded Reload
+; X86-X87-NEXT: fxch %st(1)
+; X86-X87-NEXT: addl $16, %esp
+; X86-X87-NEXT: .cfi_def_cfa_offset 8
+; X86-X87-NEXT: popl %esi
+; X86-X87-NEXT: .cfi_def_cfa_offset 4
+; X86-X87-NEXT: retl
+;
+; X86-SSE-LABEL: powi_v2f32:
+; X86-SSE: # %bb.0:
+; X86-SSE-NEXT: pushl %esi
+; X86-SSE-NEXT: .cfi_def_cfa_offset 8
+; X86-SSE-NEXT: subl $32, %esp
+; X86-SSE-NEXT: .cfi_def_cfa_offset 40
+; X86-SSE-NEXT: .cfi_offset %esi, -8
+; X86-SSE-NEXT: movups %xmm0, {{[-0-9]+}}(%e{{[sb]}}p) # 16-byte Spill
+; X86-SSE-NEXT: pushl $15
+; X86-SSE-NEXT: .cfi_adjust_cfa_offset 4
+; X86-SSE-NEXT: popl %esi
+; X86-SSE-NEXT: .cfi_adjust_cfa_offset -4
+; X86-SSE-NEXT: movl %esi, {{[0-9]+}}(%esp)
+; X86-SSE-NEXT: movss %xmm0, (%esp)
+; X86-SSE-NEXT: calll __powisf2
+; X86-SSE-NEXT: movl %esi, {{[0-9]+}}(%esp)
+; X86-SSE-NEXT: movups {{[-0-9]+}}(%e{{[sb]}}p), %xmm0 # 16-byte Reload
+; X86-SSE-NEXT: shufps {{.*#+}} xmm0 = xmm0[1,1,1,1]
+; X86-SSE-NEXT: movss %xmm0, (%esp)
+; X86-SSE-NEXT: fstps {{[0-9]+}}(%esp)
+; X86-SSE-NEXT: calll __powisf2
+; X86-SSE-NEXT: fstps {{[0-9]+}}(%esp)
+; X86-SSE-NEXT: movss {{.*#+}} xmm0 = mem[0],zero,zero,zero
+; X86-SSE-NEXT: movss {{.*#+}} xmm1 = mem[0],zero,zero,zero
+; X86-SSE-NEXT: unpcklps {{.*#+}} xmm0 = xmm0[0],xmm1[0],xmm0[1],xmm1[1]
+; X86-SSE-NEXT: addl $32, %esp
+; X86-SSE-NEXT: .cfi_def_cfa_offset 8
+; X86-SSE-NEXT: popl %esi
+; X86-SSE-NEXT: .cfi_def_cfa_offset 4
+; X86-SSE-NEXT: retl
+;
+; X64-LABEL: powi_v2f32:
+; X64: # %bb.0:
+; X64-NEXT: pushq %rbx
+; X64-NEXT: .cfi_def_cfa_offset 16
+; X64-NEXT: subq $32, %rsp
+; X64-NEXT: .cfi_def_cfa_offset 48
+; X64-NEXT: .cfi_offset %rbx, -16
+; X64-NEXT: movaps %xmm0, (%rsp) # 16-byte Spill
+; X64-NEXT: pushq $15
+; X64-NEXT: .cfi_adjust_cfa_offset 8
+; X64-NEXT: popq %rbx
+; X64-NEXT: .cfi_adjust_cfa_offset -8
+; X64-NEXT: movl %ebx, %edi
+; X64-NEXT: callq __powisf2 at PLT
+; X64-NEXT: movaps %xmm0, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
+; X64-NEXT: movaps (%rsp), %xmm0 # 16-byte Reload
+; X64-NEXT: shufps {{.*#+}} xmm0 = xmm0[1,1,1,1]
+; X64-NEXT: movl %ebx, %edi
+; X64-NEXT: callq __powisf2 at PLT
+; X64-NEXT: movaps {{[-0-9]+}}(%r{{[sb]}}p), %xmm1 # 16-byte Reload
+; X64-NEXT: unpcklps {{.*#+}} xmm1 = xmm1[0],xmm0[0],xmm1[1],xmm0[1]
+; X64-NEXT: movaps %xmm1, %xmm0
+; X64-NEXT: addq $32, %rsp
+; X64-NEXT: .cfi_def_cfa_offset 16
+; X64-NEXT: popq %rbx
+; X64-NEXT: .cfi_def_cfa_offset 8
+; X64-NEXT: retq
+ %ret = tail call < 2 x float> @llvm.powi.v2f32.i32(<2 x float> %a, i32 15) nounwind ;
+ ret <2 x float> %ret
+}
+
+declare double @llvm.powi.f64.i32(double, i32) nounwind readonly
+declare < 2 x float> @llvm.powi.v2f32.i32(<2 x float>, i32) nounwind readonly
+
+!llvm.module.flags = !{!0}
+!0 = !{i32 1, !"ProfileSummary", !1}
+!1 = !{!2, !3, !4, !5, !6, !7, !8, !9}
+!2 = !{!"ProfileFormat", !"InstrProf"}
+!3 = !{!"TotalCount", i64 10000}
+!4 = !{!"MaxCount", i64 10}
+!5 = !{!"MaxInternalCount", i64 1}
+!6 = !{!"MaxFunctionCount", i64 1000}
+!7 = !{!"NumCounts", i64 3}
+!8 = !{!"NumFunctions", i64 3}
+!9 = !{!"DetailedSummary", !10}
+!10 = !{!11, !12, !13}
+!11 = !{i32 10000, i64 100, i32 1}
+!12 = !{i32 999000, i64 100, i32 1}
+!13 = !{i32 999999, i64 1, i32 2}
+!14 = !{!"function_entry_count", i64 0}
diff --git a/llvm/test/CodeGen/X86/powi.ll b/llvm/test/CodeGen/X86/powi.ll
index 7dc6564e62a85..6a1a9e3ad7118 100644
--- a/llvm/test/CodeGen/X86/powi.ll
+++ b/llvm/test/CodeGen/X86/powi.ll
@@ -1,268 +1,175 @@
-; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
-; RUN: llc < %s -mtriple=i686-unknown-unknown | FileCheck %s --check-prefix=X86-X87
-; RUN: llc < %s -mtriple=i686-unknown-unknown -mattr=+sse2 | FileCheck %s --check-prefix=X86-SSE
-; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+sse2 | FileCheck %s --check-prefix=X64
+; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py UTC_ARGS: --version 2
+; RUN: llc < %s -mtriple=i686-linux-gnu -fast-isel | FileCheck %s --check-prefixes=FAST-X86
+; RUN: llc < %s -mtriple=i686-linux-gnu -global-isel=0 -fast-isel=0 | FileCheck %s --check-prefixes=SDAG-X86
+; RUN: llc < %s -mtriple=i686-linux-gnu -global-isel -global-isel-abort=2 | FileCheck %s --check-prefixes=GISEL-X86
+; RUN: llc < %s -mtriple=x86_64-unknown-linux-gnu -fast-isel | FileCheck %s --check-prefixes=FAST-X64
+; RUN: llc < %s -mtriple=x86_64-unknown-linux-gnu -global-isel=0 -fast-isel=0 | FileCheck %s --check-prefixes=SDAG-X64
+; RUN: llc < %s -mtriple=x86_64-unknown-linux-gnu -global-isel -global-isel-abort=2 | FileCheck %s --check-prefixes=GISEL-X64
-; Ideally this would compile to 5 multiplies.
-
-define double @pow_wrapper(double %a) nounwind readonly ssp noredzone {
-; X86-X87-LABEL: pow_wrapper:
-; X86-X87: # %bb.0:
-; X86-X87-NEXT: fldl {{[0-9]+}}(%esp)
-; X86-X87-NEXT: fld %st(0)
-; X86-X87-NEXT: fmul %st(1), %st
-; X86-X87-NEXT: fmul %st, %st(1)
-; X86-X87-NEXT: fmul %st, %st(0)
-; X86-X87-NEXT: fmul %st, %st(1)
-; X86-X87-NEXT: fmul %st, %st(0)
-; X86-X87-NEXT: fmulp %st, %st(1)
-; X86-X87-NEXT: retl
+define float @test_powi_f32_i32(float %Val, i32 %x) nounwind {
+; FAST-X86-LABEL: test_powi_f32_i32:
+; FAST-X86: # %bb.0:
+; FAST-X86-NEXT: subl $12, %esp
+; FAST-X86-NEXT: movl {{[0-9]+}}(%esp), %eax
+; FAST-X86-NEXT: flds {{[0-9]+}}(%esp)
+; FAST-X86-NEXT: movl %eax, {{[0-9]+}}(%esp)
+; FAST-X86-NEXT: fstps (%esp)
+; FAST-X86-NEXT: calll __powisf2
+; FAST-X86-NEXT: addl $12, %esp
+; FAST-X86-NEXT: retl
;
-; X86-SSE-LABEL: pow_wrapper:
-; X86-SSE: # %bb.0:
-; X86-SSE-NEXT: pushl %ebp
-; X86-SSE-NEXT: movl %esp, %ebp
-; X86-SSE-NEXT: andl $-8, %esp
-; X86-SSE-NEXT: subl $8, %esp
-; X86-SSE-NEXT: movsd {{.*#+}} xmm0 = mem[0],zero
-; X86-SSE-NEXT: movapd %xmm0, %xmm1
-; X86-SSE-NEXT: mulsd %xmm0, %xmm1
-; X86-SSE-NEXT: mulsd %xmm1, %xmm0
-; X86-SSE-NEXT: mulsd %xmm1, %xmm1
-; X86-SSE-NEXT: mulsd %xmm1, %xmm0
-; X86-SSE-NEXT: mulsd %xmm1, %xmm1
-; X86-SSE-NEXT: mulsd %xmm0, %xmm1
-; X86-SSE-NEXT: movsd %xmm1, (%esp)
-; X86-SSE-NEXT: fldl (%esp)
-; X86-SSE-NEXT: movl %ebp, %esp
-; X86-SSE-NEXT: popl %ebp
-; X86-SSE-NEXT: retl
+; SDAG-X86-LABEL: test_powi_f32_i32:
+; SDAG-X86: # %bb.0:
+; SDAG-X86-NEXT: subl $12, %esp
+; SDAG-X86-NEXT: flds {{[0-9]+}}(%esp)
+; SDAG-X86-NEXT: movl {{[0-9]+}}(%esp), %eax
+; SDAG-X86-NEXT: movl %eax, {{[0-9]+}}(%esp)
+; SDAG-X86-NEXT: fstps (%esp)
+; SDAG-X86-NEXT: calll __powisf2
+; SDAG-X86-NEXT: addl $12, %esp
+; SDAG-X86-NEXT: retl
;
-; X64-LABEL: pow_wrapper:
-; X64: # %bb.0:
-; X64-NEXT: movapd %xmm0, %xmm1
-; X64-NEXT: mulsd %xmm0, %xmm1
-; X64-NEXT: mulsd %xmm1, %xmm0
-; X64-NEXT: mulsd %xmm1, %xmm1
-; X64-NEXT: mulsd %xmm1, %xmm0
-; X64-NEXT: mulsd %xmm1, %xmm1
-; X64-NEXT: mulsd %xmm1, %xmm0
-; X64-NEXT: retq
- %ret = tail call double @llvm.powi.f64.i32(double %a, i32 15) nounwind ; <double> [#uses=1]
- ret double %ret
-}
-
-define double @pow_wrapper_optsize(double %a) optsize {
-; X86-X87-LABEL: pow_wrapper_optsize:
-; X86-X87: # %bb.0:
-; X86-X87-NEXT: subl $12, %esp
-; X86-X87-NEXT: .cfi_def_cfa_offset 16
-; X86-X87-NEXT: fldl {{[0-9]+}}(%esp)
-; X86-X87-NEXT: fstpl (%esp)
-; X86-X87-NEXT: movl $15, {{[0-9]+}}(%esp)
-; X86-X87-NEXT: calll __powidf2
-; X86-X87-NEXT: addl $12, %esp
-; X86-X87-NEXT: .cfi_def_cfa_offset 4
-; X86-X87-NEXT: retl
+; GISEL-X86-LABEL: test_powi_f32_i32:
+; GISEL-X86: # %bb.0:
+; GISEL-X86-NEXT: subl $12, %esp
+; GISEL-X86-NEXT: flds {{[0-9]+}}(%esp)
+; GISEL-X86-NEXT: movl {{[0-9]+}}(%esp), %eax
+; GISEL-X86-NEXT: movl %eax, {{[0-9]+}}(%esp)
+; GISEL-X86-NEXT: fstps (%esp)
+; GISEL-X86-NEXT: calll __powisf2
+; GISEL-X86-NEXT: addl $12, %esp
+; GISEL-X86-NEXT: retl
;
-; X86-SSE-LABEL: pow_wrapper_optsize:
-; X86-SSE: # %bb.0:
-; X86-SSE-NEXT: subl $12, %esp
-; X86-SSE-NEXT: .cfi_def_cfa_offset 16
-; X86-SSE-NEXT: movsd {{.*#+}} xmm0 = mem[0],zero
-; X86-SSE-NEXT: movsd %xmm0, (%esp)
-; X86-SSE-NEXT: movl $15, {{[0-9]+}}(%esp)
-; X86-SSE-NEXT: calll __powidf2
-; X86-SSE-NEXT: addl $12, %esp
-; X86-SSE-NEXT: .cfi_def_cfa_offset 4
-; X86-SSE-NEXT: retl
+; FAST-X64-LABEL: test_powi_f32_i32:
+; FAST-X64: # %bb.0:
+; FAST-X64-NEXT: pushq %rax
+; FAST-X64-NEXT: callq __powisf2 at PLT
+; FAST-X64-NEXT: popq %rax
+; FAST-X64-NEXT: retq
;
-; X64-LABEL: pow_wrapper_optsize:
-; X64: # %bb.0:
-; X64-NEXT: movl $15, %edi
-; X64-NEXT: jmp __powidf2 at PLT # TAILCALL
- %ret = tail call double @llvm.powi.f64.i32(double %a, i32 15) nounwind ; <double> [#uses=1]
- ret double %ret
+; SDAG-X64-LABEL: test_powi_f32_i32:
+; SDAG-X64: # %bb.0:
+; SDAG-X64-NEXT: jmp __powisf2 at PLT # TAILCALL
+;
+; GISEL-X64-LABEL: test_powi_f32_i32:
+; GISEL-X64: # %bb.0:
+; GISEL-X64-NEXT: jmp __powisf2 at PLT # TAILCALL
+ %res = call float @llvm.powi.f32.i32(float %Val, i32 %x)
+ ret float %res
}
-define double @pow_wrapper_pgso(double %a) !prof !14 {
-; X86-X87-LABEL: pow_wrapper_pgso:
-; X86-X87: # %bb.0:
-; X86-X87-NEXT: subl $12, %esp
-; X86-X87-NEXT: .cfi_def_cfa_offset 16
-; X86-X87-NEXT: fldl {{[0-9]+}}(%esp)
-; X86-X87-NEXT: fstpl (%esp)
-; X86-X87-NEXT: movl $15, {{[0-9]+}}(%esp)
-; X86-X87-NEXT: calll __powidf2
-; X86-X87-NEXT: addl $12, %esp
-; X86-X87-NEXT: .cfi_def_cfa_offset 4
-; X86-X87-NEXT: retl
+define double @test_powi_f64_i132(double %Val, i32 %x) nounwind {
+; FAST-X86-LABEL: test_powi_f64_i132:
+; FAST-X86: # %bb.0:
+; FAST-X86-NEXT: subl $12, %esp
+; FAST-X86-NEXT: movl {{[0-9]+}}(%esp), %eax
+; FAST-X86-NEXT: fldl {{[0-9]+}}(%esp)
+; FAST-X86-NEXT: movl %eax, {{[0-9]+}}(%esp)
+; FAST-X86-NEXT: fstpl (%esp)
+; FAST-X86-NEXT: calll __powidf2
+; FAST-X86-NEXT: addl $12, %esp
+; FAST-X86-NEXT: retl
;
-; X86-SSE-LABEL: pow_wrapper_pgso:
-; X86-SSE: # %bb.0:
-; X86-SSE-NEXT: subl $12, %esp
-; X86-SSE-NEXT: .cfi_def_cfa_offset 16
-; X86-SSE-NEXT: movsd {{.*#+}} xmm0 = mem[0],zero
-; X86-SSE-NEXT: movsd %xmm0, (%esp)
-; X86-SSE-NEXT: movl $15, {{[0-9]+}}(%esp)
-; X86-SSE-NEXT: calll __powidf2
-; X86-SSE-NEXT: addl $12, %esp
-; X86-SSE-NEXT: .cfi_def_cfa_offset 4
-; X86-SSE-NEXT: retl
+; SDAG-X86-LABEL: test_powi_f64_i132:
+; SDAG-X86: # %bb.0:
+; SDAG-X86-NEXT: subl $12, %esp
+; SDAG-X86-NEXT: fldl {{[0-9]+}}(%esp)
+; SDAG-X86-NEXT: movl {{[0-9]+}}(%esp), %eax
+; SDAG-X86-NEXT: movl %eax, {{[0-9]+}}(%esp)
+; SDAG-X86-NEXT: fstpl (%esp)
+; SDAG-X86-NEXT: calll __powidf2
+; SDAG-X86-NEXT: addl $12, %esp
+; SDAG-X86-NEXT: retl
;
-; X64-LABEL: pow_wrapper_pgso:
-; X64: # %bb.0:
-; X64-NEXT: movl $15, %edi
-; X64-NEXT: jmp __powidf2 at PLT # TAILCALL
- %ret = tail call double @llvm.powi.f64.i32(double %a, i32 15) nounwind ; <double> [#uses=1]
- ret double %ret
-}
-
-define double @pow_wrapper_minsize(double %a) minsize {
-; X86-X87-LABEL: pow_wrapper_minsize:
-; X86-X87: # %bb.0:
-; X86-X87-NEXT: subl $12, %esp
-; X86-X87-NEXT: .cfi_def_cfa_offset 16
-; X86-X87-NEXT: fldl {{[0-9]+}}(%esp)
-; X86-X87-NEXT: fstpl (%esp)
-; X86-X87-NEXT: movl $15, {{[0-9]+}}(%esp)
-; X86-X87-NEXT: calll __powidf2
-; X86-X87-NEXT: addl $12, %esp
-; X86-X87-NEXT: .cfi_def_cfa_offset 4
-; X86-X87-NEXT: retl
+; GISEL-X86-LABEL: test_powi_f64_i132:
+; GISEL-X86: # %bb.0:
+; GISEL-X86-NEXT: subl $12, %esp
+; GISEL-X86-NEXT: fldl {{[0-9]+}}(%esp)
+; GISEL-X86-NEXT: movl {{[0-9]+}}(%esp), %eax
+; GISEL-X86-NEXT: movl %eax, {{[0-9]+}}(%esp)
+; GISEL-X86-NEXT: fstpl (%esp)
+; GISEL-X86-NEXT: calll __powidf2
+; GISEL-X86-NEXT: addl $12, %esp
+; GISEL-X86-NEXT: retl
;
-; X86-SSE-LABEL: pow_wrapper_minsize:
-; X86-SSE: # %bb.0:
-; X86-SSE-NEXT: subl $12, %esp
-; X86-SSE-NEXT: .cfi_def_cfa_offset 16
-; X86-SSE-NEXT: movsd {{.*#+}} xmm0 = mem[0],zero
-; X86-SSE-NEXT: movsd %xmm0, (%esp)
-; X86-SSE-NEXT: movl $15, {{[0-9]+}}(%esp)
-; X86-SSE-NEXT: calll __powidf2
-; X86-SSE-NEXT: addl $12, %esp
-; X86-SSE-NEXT: .cfi_def_cfa_offset 4
-; X86-SSE-NEXT: retl
+; FAST-X64-LABEL: test_powi_f64_i132:
+; FAST-X64: # %bb.0:
+; FAST-X64-NEXT: pushq %rax
+; FAST-X64-NEXT: callq __powidf2 at PLT
+; FAST-X64-NEXT: popq %rax
+; FAST-X64-NEXT: retq
;
-; X64-LABEL: pow_wrapper_minsize:
-; X64: # %bb.0:
-; X64-NEXT: pushq $15
-; X64-NEXT: .cfi_adjust_cfa_offset 8
-; X64-NEXT: popq %rdi
-; X64-NEXT: .cfi_adjust_cfa_offset -8
-; X64-NEXT: jmp __powidf2 at PLT # TAILCALL
- %ret = tail call double @llvm.powi.f64.i32(double %a, i32 15) nounwind ; <double> [#uses=1]
- ret double %ret
+; SDAG-X64-LABEL: test_powi_f64_i132:
+; SDAG-X64: # %bb.0:
+; SDAG-X64-NEXT: jmp __powidf2 at PLT # TAILCALL
+;
+; GISEL-X64-LABEL: test_powi_f64_i132:
+; GISEL-X64: # %bb.0:
+; GISEL-X64-NEXT: jmp __powidf2 at PLT # TAILCALL
+ %res = call double @llvm.powi.f64.i32(double %Val, i32 %x)
+ ret double %res
}
-define <2 x float> @powi_v2f32(<2 x float> %a) minsize {
-; X86-X87-LABEL: powi_v2f32:
-; X86-X87: # %bb.0:
-; X86-X87-NEXT: pushl %esi
-; X86-X87-NEXT: .cfi_def_cfa_offset 8
-; X86-X87-NEXT: subl $16, %esp
-; X86-X87-NEXT: .cfi_def_cfa_offset 24
-; X86-X87-NEXT: .cfi_offset %esi, -8
-; X86-X87-NEXT: flds {{[0-9]+}}(%esp)
-; X86-X87-NEXT: fstps {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Folded Spill
-; X86-X87-NEXT: flds {{[0-9]+}}(%esp)
-; X86-X87-NEXT: pushl $15
-; X86-X87-NEXT: .cfi_adjust_cfa_offset 4
-; X86-X87-NEXT: popl %esi
-; X86-X87-NEXT: .cfi_adjust_cfa_offset -4
-; X86-X87-NEXT: movl %esi, {{[0-9]+}}(%esp)
-; X86-X87-NEXT: fstps (%esp)
-; X86-X87-NEXT: calll __powisf2
-; X86-X87-NEXT: fstps {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Folded Spill
-; X86-X87-NEXT: movl %esi, {{[0-9]+}}(%esp)
-; X86-X87-NEXT: flds {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Folded Reload
-; X86-X87-NEXT: fstps (%esp)
-; X86-X87-NEXT: calll __powisf2
-; X86-X87-NEXT: flds {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Folded Reload
-; X86-X87-NEXT: fxch %st(1)
-; X86-X87-NEXT: addl $16, %esp
-; X86-X87-NEXT: .cfi_def_cfa_offset 8
-; X86-X87-NEXT: popl %esi
-; X86-X87-NEXT: .cfi_def_cfa_offset 4
-; X86-X87-NEXT: retl
+define x86_fp80 @test_powi_f80_i32(x86_fp80 %Val, i32 %x) nounwind {
+; FAST-X86-LABEL: test_powi_f80_i32:
+; FAST-X86: # %bb.0:
+; FAST-X86-NEXT: subl $28, %esp
+; FAST-X86-NEXT: movl {{[0-9]+}}(%esp), %eax
+; FAST-X86-NEXT: fldt {{[0-9]+}}(%esp)
+; FAST-X86-NEXT: movl %eax, {{[0-9]+}}(%esp)
+; FAST-X86-NEXT: fstpt (%esp)
+; FAST-X86-NEXT: calll __powixf2
+; FAST-X86-NEXT: addl $28, %esp
+; FAST-X86-NEXT: retl
+;
+; SDAG-X86-LABEL: test_powi_f80_i32:
+; SDAG-X86: # %bb.0:
+; SDAG-X86-NEXT: subl $28, %esp
+; SDAG-X86-NEXT: fldt {{[0-9]+}}(%esp)
+; SDAG-X86-NEXT: movl {{[0-9]+}}(%esp), %eax
+; SDAG-X86-NEXT: movl %eax, {{[0-9]+}}(%esp)
+; SDAG-X86-NEXT: fstpt (%esp)
+; SDAG-X86-NEXT: calll __powixf2
+; SDAG-X86-NEXT: addl $28, %esp
+; SDAG-X86-NEXT: retl
+;
+; GISEL-X86-LABEL: test_powi_f80_i32:
+; GISEL-X86: # %bb.0:
+; GISEL-X86-NEXT: subl $28, %esp
+; GISEL-X86-NEXT: fldt {{[0-9]+}}(%esp)
+; GISEL-X86-NEXT: movl {{[0-9]+}}(%esp), %eax
+; GISEL-X86-NEXT: movl %eax, {{[0-9]+}}(%esp)
+; GISEL-X86-NEXT: fstpt (%esp)
+; GISEL-X86-NEXT: calll __powixf2
+; GISEL-X86-NEXT: addl $28, %esp
+; GISEL-X86-NEXT: retl
;
-; X86-SSE-LABEL: powi_v2f32:
-; X86-SSE: # %bb.0:
-; X86-SSE-NEXT: pushl %esi
-; X86-SSE-NEXT: .cfi_def_cfa_offset 8
-; X86-SSE-NEXT: subl $32, %esp
-; X86-SSE-NEXT: .cfi_def_cfa_offset 40
-; X86-SSE-NEXT: .cfi_offset %esi, -8
-; X86-SSE-NEXT: movups %xmm0, {{[-0-9]+}}(%e{{[sb]}}p) # 16-byte Spill
-; X86-SSE-NEXT: pushl $15
-; X86-SSE-NEXT: .cfi_adjust_cfa_offset 4
-; X86-SSE-NEXT: popl %esi
-; X86-SSE-NEXT: .cfi_adjust_cfa_offset -4
-; X86-SSE-NEXT: movl %esi, {{[0-9]+}}(%esp)
-; X86-SSE-NEXT: movss %xmm0, (%esp)
-; X86-SSE-NEXT: calll __powisf2
-; X86-SSE-NEXT: movl %esi, {{[0-9]+}}(%esp)
-; X86-SSE-NEXT: movups {{[-0-9]+}}(%e{{[sb]}}p), %xmm0 # 16-byte Reload
-; X86-SSE-NEXT: shufps {{.*#+}} xmm0 = xmm0[1,1,1,1]
-; X86-SSE-NEXT: movss %xmm0, (%esp)
-; X86-SSE-NEXT: fstps {{[0-9]+}}(%esp)
-; X86-SSE-NEXT: calll __powisf2
-; X86-SSE-NEXT: fstps {{[0-9]+}}(%esp)
-; X86-SSE-NEXT: movss {{.*#+}} xmm0 = mem[0],zero,zero,zero
-; X86-SSE-NEXT: movss {{.*#+}} xmm1 = mem[0],zero,zero,zero
-; X86-SSE-NEXT: unpcklps {{.*#+}} xmm0 = xmm0[0],xmm1[0],xmm0[1],xmm1[1]
-; X86-SSE-NEXT: addl $32, %esp
-; X86-SSE-NEXT: .cfi_def_cfa_offset 8
-; X86-SSE-NEXT: popl %esi
-; X86-SSE-NEXT: .cfi_def_cfa_offset 4
-; X86-SSE-NEXT: retl
+; FAST-X64-LABEL: test_powi_f80_i32:
+; FAST-X64: # %bb.0:
+; FAST-X64-NEXT: subq $24, %rsp
+; FAST-X64-NEXT: fldt {{[0-9]+}}(%rsp)
+; FAST-X64-NEXT: fstpt (%rsp)
+; FAST-X64-NEXT: callq __powixf2 at PLT
+; FAST-X64-NEXT: addq $24, %rsp
+; FAST-X64-NEXT: retq
;
-; X64-LABEL: powi_v2f32:
-; X64: # %bb.0:
-; X64-NEXT: pushq %rbx
-; X64-NEXT: .cfi_def_cfa_offset 16
-; X64-NEXT: subq $32, %rsp
-; X64-NEXT: .cfi_def_cfa_offset 48
-; X64-NEXT: .cfi_offset %rbx, -16
-; X64-NEXT: movaps %xmm0, (%rsp) # 16-byte Spill
-; X64-NEXT: pushq $15
-; X64-NEXT: .cfi_adjust_cfa_offset 8
-; X64-NEXT: popq %rbx
-; X64-NEXT: .cfi_adjust_cfa_offset -8
-; X64-NEXT: movl %ebx, %edi
-; X64-NEXT: callq __powisf2 at PLT
-; X64-NEXT: movaps %xmm0, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
-; X64-NEXT: movaps (%rsp), %xmm0 # 16-byte Reload
-; X64-NEXT: shufps {{.*#+}} xmm0 = xmm0[1,1,1,1]
-; X64-NEXT: movl %ebx, %edi
-; X64-NEXT: callq __powisf2 at PLT
-; X64-NEXT: movaps {{[-0-9]+}}(%r{{[sb]}}p), %xmm1 # 16-byte Reload
-; X64-NEXT: unpcklps {{.*#+}} xmm1 = xmm1[0],xmm0[0],xmm1[1],xmm0[1]
-; X64-NEXT: movaps %xmm1, %xmm0
-; X64-NEXT: addq $32, %rsp
-; X64-NEXT: .cfi_def_cfa_offset 16
-; X64-NEXT: popq %rbx
-; X64-NEXT: .cfi_def_cfa_offset 8
-; X64-NEXT: retq
- %ret = tail call < 2 x float> @llvm.powi.v2f32.i32(<2 x float> %a, i32 15) nounwind ;
- ret <2 x float> %ret
+; SDAG-X64-LABEL: test_powi_f80_i32:
+; SDAG-X64: # %bb.0:
+; SDAG-X64-NEXT: subq $24, %rsp
+; SDAG-X64-NEXT: fldt {{[0-9]+}}(%rsp)
+; SDAG-X64-NEXT: fstpt (%rsp)
+; SDAG-X64-NEXT: callq __powixf2 at PLT
+; SDAG-X64-NEXT: addq $24, %rsp
+; SDAG-X64-NEXT: retq
+;
+; GISEL-X64-LABEL: test_powi_f80_i32:
+; GISEL-X64: # %bb.0:
+; GISEL-X64-NEXT: subq $24, %rsp
+; GISEL-X64-NEXT: fldt {{[0-9]+}}(%rsp)
+; GISEL-X64-NEXT: fstpt (%rsp)
+; GISEL-X64-NEXT: callq __powixf2 at PLT
+; GISEL-X64-NEXT: addq $24, %rsp
+; GISEL-X64-NEXT: retq
+ %res = call x86_fp80 @llvm.powi.f80.i32(x86_fp80 %Val, i32 %x)
+ ret x86_fp80 %res
}
-
-declare double @llvm.powi.f64.i32(double, i32) nounwind readonly
-declare < 2 x float> @llvm.powi.v2f32.i32(<2 x float>, i32) nounwind readonly
-
-!llvm.module.flags = !{!0}
-!0 = !{i32 1, !"ProfileSummary", !1}
-!1 = !{!2, !3, !4, !5, !6, !7, !8, !9}
-!2 = !{!"ProfileFormat", !"InstrProf"}
-!3 = !{!"TotalCount", i64 10000}
-!4 = !{!"MaxCount", i64 10}
-!5 = !{!"MaxInternalCount", i64 1}
-!6 = !{!"MaxFunctionCount", i64 1000}
-!7 = !{!"NumCounts", i64 3}
-!8 = !{!"NumFunctions", i64 3}
-!9 = !{!"DetailedSummary", !10}
-!10 = !{!11, !12, !13}
-!11 = !{i32 10000, i64 100, i32 1}
-!12 = !{i32 999000, i64 100, i32 1}
-!13 = !{i32 999999, i64 1, i32 2}
-!14 = !{!"function_entry_count", i64 0}
>From 358c37b047034e2d21574d3d014b202de5b342f2 Mon Sep 17 00:00:00 2001
From: Chauhan Jaydeep Ashwinbhai <chauhan.jaydeep.ashwinbhai at intel.com>
Date: Thu, 3 Apr 2025 10:14:50 -0700
Subject: [PATCH 2/2] Testcase name updated
---
llvm/test/CodeGen/X86/powi.ll | 14 +++++++-------
1 file changed, 7 insertions(+), 7 deletions(-)
diff --git a/llvm/test/CodeGen/X86/powi.ll b/llvm/test/CodeGen/X86/powi.ll
index 6a1a9e3ad7118..e4c691bfbd2e5 100644
--- a/llvm/test/CodeGen/X86/powi.ll
+++ b/llvm/test/CodeGen/X86/powi.ll
@@ -58,8 +58,8 @@ define float @test_powi_f32_i32(float %Val, i32 %x) nounwind {
ret float %res
}
-define double @test_powi_f64_i132(double %Val, i32 %x) nounwind {
-; FAST-X86-LABEL: test_powi_f64_i132:
+define double @test_powi_f64_i32(double %Val, i32 %x) nounwind {
+; FAST-X86-LABEL: test_powi_f64_i32:
; FAST-X86: # %bb.0:
; FAST-X86-NEXT: subl $12, %esp
; FAST-X86-NEXT: movl {{[0-9]+}}(%esp), %eax
@@ -70,7 +70,7 @@ define double @test_powi_f64_i132(double %Val, i32 %x) nounwind {
; FAST-X86-NEXT: addl $12, %esp
; FAST-X86-NEXT: retl
;
-; SDAG-X86-LABEL: test_powi_f64_i132:
+; SDAG-X86-LABEL: test_powi_f64_i32:
; SDAG-X86: # %bb.0:
; SDAG-X86-NEXT: subl $12, %esp
; SDAG-X86-NEXT: fldl {{[0-9]+}}(%esp)
@@ -81,7 +81,7 @@ define double @test_powi_f64_i132(double %Val, i32 %x) nounwind {
; SDAG-X86-NEXT: addl $12, %esp
; SDAG-X86-NEXT: retl
;
-; GISEL-X86-LABEL: test_powi_f64_i132:
+; GISEL-X86-LABEL: test_powi_f64_i32:
; GISEL-X86: # %bb.0:
; GISEL-X86-NEXT: subl $12, %esp
; GISEL-X86-NEXT: fldl {{[0-9]+}}(%esp)
@@ -92,18 +92,18 @@ define double @test_powi_f64_i132(double %Val, i32 %x) nounwind {
; GISEL-X86-NEXT: addl $12, %esp
; GISEL-X86-NEXT: retl
;
-; FAST-X64-LABEL: test_powi_f64_i132:
+; FAST-X64-LABEL: test_powi_f64_i32:
; FAST-X64: # %bb.0:
; FAST-X64-NEXT: pushq %rax
; FAST-X64-NEXT: callq __powidf2 at PLT
; FAST-X64-NEXT: popq %rax
; FAST-X64-NEXT: retq
;
-; SDAG-X64-LABEL: test_powi_f64_i132:
+; SDAG-X64-LABEL: test_powi_f64_i32:
; SDAG-X64: # %bb.0:
; SDAG-X64-NEXT: jmp __powidf2 at PLT # TAILCALL
;
-; GISEL-X64-LABEL: test_powi_f64_i132:
+; GISEL-X64-LABEL: test_powi_f64_i32:
; GISEL-X64: # %bb.0:
; GISEL-X64-NEXT: jmp __powidf2 at PLT # TAILCALL
%res = call double @llvm.powi.f64.i32(double %Val, i32 %x)
More information about the llvm-commits
mailing list