Skip to content

[X86] Modify tests for constrained rounding functions #116951

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 3 commits into from
Nov 25, 2024
Merged

[X86] Modify tests for constrained rounding functions #116951

merged 3 commits into from
Nov 25, 2024

Conversation

spavloff
Copy link
Collaborator

The existing tests for constrained functions often use constant arguments. If constant evaluation is enhanced, such tests will not check code generation of the tested functions. To avoid it, the tests are modified to use loaded value instead of constants. Now only the tests for rounding functions are changed.

@spavloff
[X86] Modify tests for constrained rounding functions
e7e8312 
The existing tests for constrained functions often use constant
arguments. If constant evaluation is enhanced, such tests will not check
code generation of the tested functions. To avoid it, the tests are
modified to use loaded value instead of constants. Now only the tests
for rounding functions are changed.
@spavloff spavloff requested review from RKSimon and topperc November 20, 2024 10:40
@spavloff spavloff self-assigned this Nov 20, 2024
@llvmbot
Copy link
Member

llvmbot commented Nov 20, 2024

@llvm/pr-subscribers-backend-x86

Author: Serge Pavlov (spavloff)

Changes

The existing tests for constrained functions often use constant arguments. If constant evaluation is enhanced, such tests will not check code generation of the tested functions. To avoid it, the tests are modified to use loaded value instead of constants. Now only the tests for rounding functions are changed.

Patch is 59.53 KiB, truncated to 20.00 KiB below, full version: https://github.com/llvm/llvm-project/pull/116951.diff

1 Files Affected:

  • (modified) llvm/test/CodeGen/X86/vector-constrained-fp-intrinsics.ll (+424-309)
diff --git a/llvm/test/CodeGen/X86/vector-constrained-fp-intrinsics.ll b/llvm/test/CodeGen/X86/vector-constrained-fp-intrinsics.ll
index 21dfdc3c2abe49..579b15bbbb0ff8 100644
--- a/llvm/test/CodeGen/X86/vector-constrained-fp-intrinsics.ll
+++ b/llvm/test/CodeGen/X86/vector-constrained-fp-intrinsics.ll
@@ -3011,12 +3011,12 @@ entry:
   ret <4 x double> %log2
 }
 
-define <1 x float> @constrained_vector_rint_v1f32() #0 {
+define <1 x float> @constrained_vector_rint_v1f32(ptr %a) #0 {
 ; CHECK-LABEL: constrained_vector_rint_v1f32:
 ; CHECK:       # %bb.0: # %entry
 ; CHECK-NEXT:    pushq %rax
 ; CHECK-NEXT:    .cfi_def_cfa_offset 16
-; CHECK-NEXT:    movss {{.*#+}} xmm0 = [4.2E+1,0.0E+0,0.0E+0,0.0E+0]
+; CHECK-NEXT:    movss {{.*#+}} xmm0 = mem[0],zero,zero,zero
 ; CHECK-NEXT:    callq rintf@PLT
 ; CHECK-NEXT:    popq %rax
 ; CHECK-NEXT:    .cfi_def_cfa_offset 8
@@ -3024,98 +3024,111 @@ define <1 x float> @constrained_vector_rint_v1f32() #0 {
 ;
 ; AVX-LABEL: constrained_vector_rint_v1f32:
 ; AVX:       # %bb.0: # %entry
-; AVX-NEXT:    vmovss {{.*#+}} xmm0 = [4.2E+1,0.0E+0,0.0E+0,0.0E+0]
+; AVX-NEXT:    vmovss {{.*#+}} xmm0 = mem[0],zero,zero,zero
 ; AVX-NEXT:    vroundss $4, %xmm0, %xmm0, %xmm0
 ; AVX-NEXT:    retq
 entry:
+  %b = load <1 x float>, ptr %a
   %rint = call <1 x float> @llvm.experimental.constrained.rint.v1f32(
-                             <1 x float> <float 42.0>,
+                             <1 x float> %b,
                              metadata !"round.dynamic",
                              metadata !"fpexcept.strict") #0
   ret <1 x float> %rint
 }
 
-define <2 x double> @constrained_vector_rint_v2f64() #0 {
+define <2 x double> @constrained_vector_rint_v2f64(ptr %a) #0 {
 ; CHECK-LABEL: constrained_vector_rint_v2f64:
 ; CHECK:       # %bb.0: # %entry
-; CHECK-NEXT:    subq $24, %rsp
-; CHECK-NEXT:    .cfi_def_cfa_offset 32
-; CHECK-NEXT:    movsd {{.*#+}} xmm0 = [4.2E+1,0.0E+0]
-; CHECK-NEXT:    callq rint@PLT
+; CHECK-NEXT:    subq $40, %rsp
+; CHECK-NEXT:    .cfi_def_cfa_offset 48
+; CHECK-NEXT:    movaps (%rdi), %xmm0
 ; CHECK-NEXT:    movaps %xmm0, (%rsp) # 16-byte Spill
-; CHECK-NEXT:    movsd {{.*#+}} xmm0 = [4.2100000000000001E+1,0.0E+0]
 ; CHECK-NEXT:    callq rint@PLT
-; CHECK-NEXT:    unpcklpd (%rsp), %xmm0 # 16-byte Folded Reload
-; CHECK-NEXT:    # xmm0 = xmm0[0],mem[0]
-; CHECK-NEXT:    addq $24, %rsp
+; CHECK-NEXT:    movaps %xmm0, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
+; CHECK-NEXT:    movaps (%rsp), %xmm0 # 16-byte Reload
+; CHECK-NEXT:    movhlps {{.*#+}} xmm0 = xmm0[1,1]
+; CHECK-NEXT:    callq rint@PLT
+; CHECK-NEXT:    movaps {{[-0-9]+}}(%r{{[sb]}}p), %xmm1 # 16-byte Reload
+; CHECK-NEXT:    movlhps {{.*#+}} xmm1 = xmm1[0],xmm0[0]
+; CHECK-NEXT:    movaps %xmm1, %xmm0
+; CHECK-NEXT:    addq $40, %rsp
 ; CHECK-NEXT:    .cfi_def_cfa_offset 8
 ; CHECK-NEXT:    retq
 ;
 ; AVX-LABEL: constrained_vector_rint_v2f64:
 ; AVX:       # %bb.0: # %entry
-; AVX-NEXT:    vroundpd $4, {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm0
+; AVX-NEXT:    vroundpd $4, (%rdi), %xmm0
 ; AVX-NEXT:    retq
 entry:
+  %b = load <2 x double>, ptr %a
   %rint = call <2 x double> @llvm.experimental.constrained.rint.v2f64(
-                        <2 x double> <double 42.1, double 42.0>,
+                        <2 x double> %b,
                         metadata !"round.dynamic",
                         metadata !"fpexcept.strict") #0
   ret <2 x double> %rint
 }
 
-define <3 x float> @constrained_vector_rint_v3f32() #0 {
+define <3 x float> @constrained_vector_rint_v3f32(ptr %a) #0 {
 ; CHECK-LABEL: constrained_vector_rint_v3f32:
 ; CHECK:       # %bb.0: # %entry
-; CHECK-NEXT:    subq $40, %rsp
-; CHECK-NEXT:    .cfi_def_cfa_offset 48
-; CHECK-NEXT:    movss {{.*#+}} xmm0 = [4.4E+1,0.0E+0,0.0E+0,0.0E+0]
+; CHECK-NEXT:    subq $56, %rsp
+; CHECK-NEXT:    .cfi_def_cfa_offset 64
+; CHECK-NEXT:    movaps (%rdi), %xmm0
+; CHECK-NEXT:    movaps %xmm0, (%rsp) # 16-byte Spill
+; CHECK-NEXT:    movhlps {{.*#+}} xmm0 = xmm0[1,1]
 ; CHECK-NEXT:    callq rintf@PLT
 ; CHECK-NEXT:    movaps %xmm0, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
-; CHECK-NEXT:    movss {{.*#+}} xmm0 = [4.2E+1,0.0E+0,0.0E+0,0.0E+0]
+; CHECK-NEXT:    movaps (%rsp), %xmm0 # 16-byte Reload
+; CHECK-NEXT:    shufps {{.*#+}} xmm0 = xmm0[1,1,1,1]
 ; CHECK-NEXT:    callq rintf@PLT
-; CHECK-NEXT:    movaps %xmm0, (%rsp) # 16-byte Spill
-; CHECK-NEXT:    movss {{.*#+}} xmm0 = [4.3E+1,0.0E+0,0.0E+0,0.0E+0]
+; CHECK-NEXT:    movaps %xmm0, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
+; CHECK-NEXT:    movaps (%rsp), %xmm0 # 16-byte Reload
 ; CHECK-NEXT:    callq rintf@PLT
-; CHECK-NEXT:    movaps (%rsp), %xmm1 # 16-byte Reload
-; CHECK-NEXT:    unpcklps {{.*#+}} xmm1 = xmm1[0],xmm0[0],xmm1[1],xmm0[1]
-; CHECK-NEXT:    unpcklpd {{[-0-9]+}}(%r{{[sb]}}p), %xmm1 # 16-byte Folded Reload
-; CHECK-NEXT:    # xmm1 = xmm1[0],mem[0]
-; CHECK-NEXT:    movaps %xmm1, %xmm0
-; CHECK-NEXT:    addq $40, %rsp
+; CHECK-NEXT:    unpcklps {{[-0-9]+}}(%r{{[sb]}}p), %xmm0 # 16-byte Folded Reload
+; CHECK-NEXT:    # xmm0 = xmm0[0],mem[0],xmm0[1],mem[1]
+; CHECK-NEXT:    unpcklpd {{[-0-9]+}}(%r{{[sb]}}p), %xmm0 # 16-byte Folded Reload
+; CHECK-NEXT:    # xmm0 = xmm0[0],mem[0]
+; CHECK-NEXT:    addq $56, %rsp
 ; CHECK-NEXT:    .cfi_def_cfa_offset 8
 ; CHECK-NEXT:    retq
 ;
 ; AVX-LABEL: constrained_vector_rint_v3f32:
 ; AVX:       # %bb.0: # %entry
-; AVX-NEXT:    vmovss {{.*#+}} xmm0 = [4.4E+1,0.0E+0,0.0E+0,0.0E+0]
+; AVX-NEXT:    vmovss {{.*#+}} xmm0 = mem[0],zero,zero,zero
 ; AVX-NEXT:    vroundss $4, %xmm0, %xmm0, %xmm0
-; AVX-NEXT:    vmovss {{.*#+}} xmm1 = [4.2E+1,0.0E+0,0.0E+0,0.0E+0]
+; AVX-NEXT:    vmovss {{.*#+}} xmm1 = mem[0],zero,zero,zero
+; AVX-NEXT:    vmovss {{.*#+}} xmm2 = mem[0],zero,zero,zero
 ; AVX-NEXT:    vroundss $4, %xmm1, %xmm1, %xmm1
-; AVX-NEXT:    vmovss {{.*#+}} xmm2 = [4.3E+1,0.0E+0,0.0E+0,0.0E+0]
 ; AVX-NEXT:    vroundss $4, %xmm2, %xmm2, %xmm2
 ; AVX-NEXT:    vinsertps {{.*#+}} xmm1 = xmm1[0],xmm2[0],xmm1[2,3]
 ; AVX-NEXT:    vinsertps {{.*#+}} xmm0 = xmm1[0,1],xmm0[0],xmm1[3]
 ; AVX-NEXT:    retq
  entry:
+  %b = load <3 x float>, ptr %a
   %rint = call <3 x float> @llvm.experimental.constrained.rint.v3f32(
-                              <3 x float> <float 42.0, float 43.0, float 44.0>,
+                              <3 x float> %b,
                               metadata !"round.dynamic",
                               metadata !"fpexcept.strict") #0
   ret <3 x float> %rint
 }
 
-define <3 x double> @constrained_vector_rint_v3f64() #0 {
+define <3 x double> @constrained_vector_rint_v3f64(ptr %a) #0 {
 ; CHECK-LABEL: constrained_vector_rint_v3f64:
 ; CHECK:       # %bb.0: # %entry
-; CHECK-NEXT:    subq $24, %rsp
-; CHECK-NEXT:    .cfi_def_cfa_offset 32
-; CHECK-NEXT:    movsd {{.*#+}} xmm0 = [4.2100000000000001E+1,0.0E+0]
+; CHECK-NEXT:    subq $40, %rsp
+; CHECK-NEXT:    .cfi_def_cfa_offset 48
+; CHECK-NEXT:    movsd {{.*#+}} xmm0 = mem[0],zero
+; CHECK-NEXT:    movsd %xmm0, {{[-0-9]+}}(%r{{[sb]}}p) # 8-byte Spill
+; CHECK-NEXT:    movaps (%rdi), %xmm0
+; CHECK-NEXT:    movaps %xmm0, (%rsp) # 16-byte Spill
+; CHECK-NEXT:    movhlps {{.*#+}} xmm0 = xmm0[1,1]
 ; CHECK-NEXT:    callq rint@PLT
 ; CHECK-NEXT:    movsd %xmm0, {{[-0-9]+}}(%r{{[sb]}}p) # 8-byte Spill
-; CHECK-NEXT:    movsd {{.*#+}} xmm0 = [4.2E+1,0.0E+0]
+; CHECK-NEXT:    movaps (%rsp), %xmm0 # 16-byte Reload
 ; CHECK-NEXT:    callq rint@PLT
 ; CHECK-NEXT:    movsd %xmm0, (%rsp) # 8-byte Spill
-; CHECK-NEXT:    movsd {{.*#+}} xmm0 = [4.2200000000000003E+1,0.0E+0]
+; CHECK-NEXT:    movsd {{[-0-9]+}}(%r{{[sb]}}p), %xmm0 # 8-byte Reload
+; CHECK-NEXT:    # xmm0 = mem[0],zero
 ; CHECK-NEXT:    callq rint@PLT
 ; CHECK-NEXT:    movsd %xmm0, {{[0-9]+}}(%rsp)
 ; CHECK-NEXT:    fldl {{[0-9]+}}(%rsp)
@@ -3124,70 +3137,76 @@ define <3 x double> @constrained_vector_rint_v3f64() #0 {
 ; CHECK-NEXT:    # xmm0 = mem[0],zero
 ; CHECK-NEXT:    movsd {{[-0-9]+}}(%r{{[sb]}}p), %xmm1 # 8-byte Reload
 ; CHECK-NEXT:    # xmm1 = mem[0],zero
-; CHECK-NEXT:    addq $24, %rsp
+; CHECK-NEXT:    addq $40, %rsp
 ; CHECK-NEXT:    .cfi_def_cfa_offset 8
 ; CHECK-NEXT:    retq
 ;
 ; AVX-LABEL: constrained_vector_rint_v3f64:
 ; AVX:       # %bb.0: # %entry
-; AVX-NEXT:    vmovsd {{.*#+}} xmm0 = [4.2200000000000003E+1,0.0E+0]
+; AVX-NEXT:    vmovsd {{.*#+}} xmm0 = mem[0],zero
 ; AVX-NEXT:    vroundsd $4, %xmm0, %xmm0, %xmm0
-; AVX-NEXT:    vroundpd $4, {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm1
+; AVX-NEXT:    vroundpd $4, (%rdi), %xmm1
 ; AVX-NEXT:    vinsertf128 $1, %xmm0, %ymm1, %ymm0
 ; AVX-NEXT:    retq
 entry:
+  %b = load <3 x double>, ptr %a
   %rint = call <3 x double> @llvm.experimental.constrained.rint.v3f64(
-                          <3 x double> <double 42.0, double 42.1, double 42.2>,
+                          <3 x double> %b,
                           metadata !"round.dynamic",
                           metadata !"fpexcept.strict") #0
   ret <3 x double> %rint
 }
 
-define <4 x double> @constrained_vector_rint_v4f64() #0 {
+define <4 x double> @constrained_vector_rint_v4f64(ptr %a) #0 {
 ; CHECK-LABEL: constrained_vector_rint_v4f64:
 ; CHECK:       # %bb.0: # %entry
-; CHECK-NEXT:    subq $40, %rsp
-; CHECK-NEXT:    .cfi_def_cfa_offset 48
-; CHECK-NEXT:    movsd {{.*#+}} xmm0 = [4.2200000000000003E+1,0.0E+0]
+; CHECK-NEXT:    subq $56, %rsp
+; CHECK-NEXT:    .cfi_def_cfa_offset 64
+; CHECK-NEXT:    movaps (%rdi), %xmm1
+; CHECK-NEXT:    movaps %xmm1, (%rsp) # 16-byte Spill
+; CHECK-NEXT:    movaps 16(%rdi), %xmm0
+; CHECK-NEXT:    movaps %xmm0, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
+; CHECK-NEXT:    movaps %xmm1, %xmm0
 ; CHECK-NEXT:    callq rint@PLT
-; CHECK-NEXT:    movaps %xmm0, (%rsp) # 16-byte Spill
-; CHECK-NEXT:    movsd {{.*#+}} xmm0 = [4.2100000000000001E+1,0.0E+0]
+; CHECK-NEXT:    movaps %xmm0, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
+; CHECK-NEXT:    movaps (%rsp), %xmm0 # 16-byte Reload
+; CHECK-NEXT:    movhlps {{.*#+}} xmm0 = xmm0[1,1]
 ; CHECK-NEXT:    callq rint@PLT
-; CHECK-NEXT:    unpcklpd (%rsp), %xmm0 # 16-byte Folded Reload
-; CHECK-NEXT:    # xmm0 = xmm0[0],mem[0]
-; CHECK-NEXT:    movaps %xmm0, (%rsp) # 16-byte Spill
-; CHECK-NEXT:    movsd {{.*#+}} xmm0 = [4.2399999999999999E+1,0.0E+0]
+; CHECK-NEXT:    movaps {{[-0-9]+}}(%r{{[sb]}}p), %xmm1 # 16-byte Reload
+; CHECK-NEXT:    movlhps {{.*#+}} xmm1 = xmm1[0],xmm0[0]
+; CHECK-NEXT:    movaps %xmm1, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
+; CHECK-NEXT:    movaps {{[-0-9]+}}(%r{{[sb]}}p), %xmm0 # 16-byte Reload
 ; CHECK-NEXT:    callq rint@PLT
-; CHECK-NEXT:    movaps %xmm0, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
-; CHECK-NEXT:    movsd {{.*#+}} xmm0 = [4.2299999999999997E+1,0.0E+0]
+; CHECK-NEXT:    movaps %xmm0, (%rsp) # 16-byte Spill
+; CHECK-NEXT:    movaps {{[-0-9]+}}(%r{{[sb]}}p), %xmm0 # 16-byte Reload
+; CHECK-NEXT:    movhlps {{.*#+}} xmm0 = xmm0[1,1]
 ; CHECK-NEXT:    callq rint@PLT
-; CHECK-NEXT:    movaps %xmm0, %xmm1
-; CHECK-NEXT:    unpcklpd {{[-0-9]+}}(%r{{[sb]}}p), %xmm1 # 16-byte Folded Reload
-; CHECK-NEXT:    # xmm1 = xmm1[0],mem[0]
-; CHECK-NEXT:    movaps (%rsp), %xmm0 # 16-byte Reload
-; CHECK-NEXT:    addq $40, %rsp
+; CHECK-NEXT:    movaps (%rsp), %xmm1 # 16-byte Reload
+; CHECK-NEXT:    movlhps {{.*#+}} xmm1 = xmm1[0],xmm0[0]
+; CHECK-NEXT:    movaps {{[-0-9]+}}(%r{{[sb]}}p), %xmm0 # 16-byte Reload
+; CHECK-NEXT:    addq $56, %rsp
 ; CHECK-NEXT:    .cfi_def_cfa_offset 8
 ; CHECK-NEXT:    retq
 ;
 ; AVX-LABEL: constrained_vector_rint_v4f64:
 ; AVX:       # %bb.0: # %entry
-; AVX-NEXT:    vroundpd $4, {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %ymm0
+; AVX-NEXT:    vroundpd $4, (%rdi), %ymm0
 ; AVX-NEXT:    retq
 entry:
+  %b = load <4 x double>, ptr %a
   %rint = call <4 x double> @llvm.experimental.constrained.rint.v4f64(
-                        <4 x double> <double 42.1, double 42.2,
-                                      double 42.3, double 42.4>,
+                        <4 x double> %b,
                         metadata !"round.dynamic",
                         metadata !"fpexcept.strict") #0
   ret <4 x double> %rint
 }
 
-define <1 x float> @constrained_vector_nearbyint_v1f32() #0 {
+define <1 x float> @constrained_vector_nearbyint_v1f32(ptr %a) #0 {
 ; CHECK-LABEL: constrained_vector_nearbyint_v1f32:
 ; CHECK:       # %bb.0: # %entry
 ; CHECK-NEXT:    pushq %rax
 ; CHECK-NEXT:    .cfi_def_cfa_offset 16
-; CHECK-NEXT:    movss {{.*#+}} xmm0 = [4.2E+1,0.0E+0,0.0E+0,0.0E+0]
+; CHECK-NEXT:    movss {{.*#+}} xmm0 = mem[0],zero,zero,zero
 ; CHECK-NEXT:    callq nearbyintf@PLT
 ; CHECK-NEXT:    popq %rax
 ; CHECK-NEXT:    .cfi_def_cfa_offset 8
@@ -3195,98 +3214,111 @@ define <1 x float> @constrained_vector_nearbyint_v1f32() #0 {
 ;
 ; AVX-LABEL: constrained_vector_nearbyint_v1f32:
 ; AVX:       # %bb.0: # %entry
-; AVX-NEXT:    vmovss {{.*#+}} xmm0 = [4.2E+1,0.0E+0,0.0E+0,0.0E+0]
+; AVX-NEXT:    vmovss {{.*#+}} xmm0 = mem[0],zero,zero,zero
 ; AVX-NEXT:    vroundss $12, %xmm0, %xmm0, %xmm0
 ; AVX-NEXT:    retq
 entry:
+  %b = load <1 x float>, ptr %a
   %nearby = call <1 x float> @llvm.experimental.constrained.nearbyint.v1f32(
-                               <1 x float> <float 42.0>,
+                               <1 x float> %b,
                                metadata !"round.dynamic",
                                metadata !"fpexcept.strict") #0
   ret <1 x float> %nearby
 }
 
-define <2 x double> @constrained_vector_nearbyint_v2f64() #0 {
+define <2 x double> @constrained_vector_nearbyint_v2f64(ptr %a) #0 {
 ; CHECK-LABEL: constrained_vector_nearbyint_v2f64:
 ; CHECK:       # %bb.0: # %entry
-; CHECK-NEXT:    subq $24, %rsp
-; CHECK-NEXT:    .cfi_def_cfa_offset 32
-; CHECK-NEXT:    movsd {{.*#+}} xmm0 = [4.2E+1,0.0E+0]
-; CHECK-NEXT:    callq nearbyint@PLT
+; CHECK-NEXT:    subq $40, %rsp
+; CHECK-NEXT:    .cfi_def_cfa_offset 48
+; CHECK-NEXT:    movaps (%rdi), %xmm0
 ; CHECK-NEXT:    movaps %xmm0, (%rsp) # 16-byte Spill
-; CHECK-NEXT:    movsd {{.*#+}} xmm0 = [4.2100000000000001E+1,0.0E+0]
 ; CHECK-NEXT:    callq nearbyint@PLT
-; CHECK-NEXT:    unpcklpd (%rsp), %xmm0 # 16-byte Folded Reload
-; CHECK-NEXT:    # xmm0 = xmm0[0],mem[0]
-; CHECK-NEXT:    addq $24, %rsp
+; CHECK-NEXT:    movaps %xmm0, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
+; CHECK-NEXT:    movaps (%rsp), %xmm0 # 16-byte Reload
+; CHECK-NEXT:    movhlps {{.*#+}} xmm0 = xmm0[1,1]
+; CHECK-NEXT:    callq nearbyint@PLT
+; CHECK-NEXT:    movaps {{[-0-9]+}}(%r{{[sb]}}p), %xmm1 # 16-byte Reload
+; CHECK-NEXT:    movlhps {{.*#+}} xmm1 = xmm1[0],xmm0[0]
+; CHECK-NEXT:    movaps %xmm1, %xmm0
+; CHECK-NEXT:    addq $40, %rsp
 ; CHECK-NEXT:    .cfi_def_cfa_offset 8
 ; CHECK-NEXT:    retq
 ;
 ; AVX-LABEL: constrained_vector_nearbyint_v2f64:
 ; AVX:       # %bb.0: # %entry
-; AVX-NEXT:    vroundpd $12, {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm0
+; AVX-NEXT:    vroundpd $12, (%rdi), %xmm0
 ; AVX-NEXT:    retq
 entry:
+  %b = load <2 x double>, ptr %a
   %nearby = call <2 x double> @llvm.experimental.constrained.nearbyint.v2f64(
-                                <2 x double> <double 42.1, double 42.0>,
+                                <2 x double> %b,
                                 metadata !"round.dynamic",
                                 metadata !"fpexcept.strict") #0
   ret <2 x double> %nearby
 }
 
-define <3 x float> @constrained_vector_nearbyint_v3f32() #0 {
+define <3 x float> @constrained_vector_nearbyint_v3f32(ptr %a) #0 {
 ; CHECK-LABEL: constrained_vector_nearbyint_v3f32:
 ; CHECK:       # %bb.0: # %entry
-; CHECK-NEXT:    subq $40, %rsp
-; CHECK-NEXT:    .cfi_def_cfa_offset 48
-; CHECK-NEXT:    movss {{.*#+}} xmm0 = [4.4E+1,0.0E+0,0.0E+0,0.0E+0]
+; CHECK-NEXT:    subq $56, %rsp
+; CHECK-NEXT:    .cfi_def_cfa_offset 64
+; CHECK-NEXT:    movaps (%rdi), %xmm0
+; CHECK-NEXT:    movaps %xmm0, (%rsp) # 16-byte Spill
+; CHECK-NEXT:    movhlps {{.*#+}} xmm0 = xmm0[1,1]
 ; CHECK-NEXT:    callq nearbyintf@PLT
 ; CHECK-NEXT:    movaps %xmm0, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
-; CHECK-NEXT:    movss {{.*#+}} xmm0 = [4.2E+1,0.0E+0,0.0E+0,0.0E+0]
+; CHECK-NEXT:    movaps (%rsp), %xmm0 # 16-byte Reload
+; CHECK-NEXT:    shufps {{.*#+}} xmm0 = xmm0[1,1,1,1]
 ; CHECK-NEXT:    callq nearbyintf@PLT
-; CHECK-NEXT:    movaps %xmm0, (%rsp) # 16-byte Spill
-; CHECK-NEXT:    movss {{.*#+}} xmm0 = [4.3E+1,0.0E+0,0.0E+0,0.0E+0]
+; CHECK-NEXT:    movaps %xmm0, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
+; CHECK-NEXT:    movaps (%rsp), %xmm0 # 16-byte Reload
 ; CHECK-NEXT:    callq nearbyintf@PLT
-; CHECK-NEXT:    movaps (%rsp), %xmm1 # 16-byte Reload
-; CHECK-NEXT:    unpcklps {{.*#+}} xmm1 = xmm1[0],xmm0[0],xmm1[1],xmm0[1]
-; CHECK-NEXT:    unpcklpd {{[-0-9]+}}(%r{{[sb]}}p), %xmm1 # 16-byte Folded Reload
-; CHECK-NEXT:    # xmm1 = xmm1[0],mem[0]
-; CHECK-NEXT:    movaps %xmm1, %xmm0
-; CHECK-NEXT:    addq $40, %rsp
+; CHECK-NEXT:    unpcklps {{[-0-9]+}}(%r{{[sb]}}p), %xmm0 # 16-byte Folded Reload
+; CHECK-NEXT:    # xmm0 = xmm0[0],mem[0],xmm0[1],mem[1]
+; CHECK-NEXT:    unpcklpd {{[-0-9]+}}(%r{{[sb]}}p), %xmm0 # 16-byte Folded Reload
+; CHECK-NEXT:    # xmm0 = xmm0[0],mem[0]
+; CHECK-NEXT:    addq $56, %rsp
 ; CHECK-NEXT:    .cfi_def_cfa_offset 8
 ; CHECK-NEXT:    retq
 ;
 ; AVX-LABEL: constrained_vector_nearbyint_v3f32:
 ; AVX:       # %bb.0: # %entry
-; AVX-NEXT:    vmovss {{.*#+}} xmm0 = [4.4E+1,0.0E+0,0.0E+0,0.0E+0]
+; AVX-NEXT:    vmovss {{.*#+}} xmm0 = mem[0],zero,zero,zero
 ; AVX-NEXT:    vroundss $12, %xmm0, %xmm0, %xmm0
-; AVX-NEXT:    vmovss {{.*#+}} xmm1 = [4.2E+1,0.0E+0,0.0E+0,0.0E+0]
+; AVX-NEXT:    vmovss {{.*#+}} xmm1 = mem[0],zero,zero,zero
+; AVX-NEXT:    vmovss {{.*#+}} xmm2 = mem[0],zero,zero,zero
 ; AVX-NEXT:    vroundss $12, %xmm1, %xmm1, %xmm1
-; AVX-NEXT:    vmovss {{.*#+}} xmm2 = [4.3E+1,0.0E+0,0.0E+0,0.0E+0]
 ; AVX-NEXT:    vroundss $12, %xmm2, %xmm2, %xmm2
 ; AVX-NEXT:    vinsertps {{.*#+}} xmm1 = xmm1[0],xmm2[0],xmm1[2,3]
 ; AVX-NEXT:    vinsertps {{.*#+}} xmm0 = xmm1[0,1],xmm0[0],xmm1[3]
 ; AVX-NEXT:    retq
 entry:
+  %b = load <3 x float>, ptr %a
   %nearby = call <3 x float> @llvm.experimental.constrained.nearbyint.v3f32(
-                              <3 x float> <float 42.0, float 43.0, float 44.0>,
+                              <3 x float> %b,
                               metadata !"round.dynamic",
                               metadata !"fpexcept.strict") #0
   ret <3 x float> %nearby
 }
 
-define <3 x double> @constrained_vector_nearby_v3f64() #0 {
+define <3 x double> @constrained_vector_nearby_v3f64(ptr %a) #0 {
 ; CHECK-LABEL: constrained_vector_nearby_v3f64:
 ; CHECK:       # %bb.0: # %entry
-; CHECK-NEXT:    subq $24, %rsp
-; CHECK-NEXT:    .cfi_def_cfa_offset 32
-; CHECK-NEXT:    movsd {{.*#+}} xmm0 = [4.2100000000000001E+1,0.0E+0]
+; CHECK-NEXT:    subq $40, %rsp
+; CHECK-NEXT:    .cfi_def_cfa_offset 48
+; CHECK-NEXT:    movsd {{.*#+}} xmm0 = mem[0],zero
+; CHECK-NEXT:    movsd %xmm0, {{[-0-9]+}}(%r{{[sb]}}p) # 8-byte Spill
+; CHECK-NEXT:    movaps (%rdi), %xmm0
+; CHECK-NEXT:    movaps %xmm0, (%rsp) # 16-byte Spill
+; CHECK-NEXT:    movhlps {{.*#+}} xmm0 = xmm0[1,1]
 ; CHECK-NEXT:    callq nearbyint@PLT
 ; CHECK-NEXT:    movsd %xmm0, {{[-0-9]+}}(%r{{[sb]}}p) # 8-byte Spill
-; CHECK-NEXT:    movsd {{.*#+}} xmm0 = [4.2E+1,0.0E+0]
+; CHECK-NEXT:    movaps (%rsp), %xmm0 # 16-byte Reload
 ; CHECK-NEXT:    callq nearbyint@PLT
 ; CHECK-NEXT:    movsd %xmm0, (%rsp) # 8-byte Spill
-; CHECK-NEXT:    movsd {{.*#+}} xmm0 = [4.2200000000000003E+1,0.0E+0]
+; CHECK-NEXT:    movsd {{[-0-9]+}}(%r{{[sb]}}p), %xmm0 # 8-byte Reload
+; CHECK-NEXT:    # xmm0 = mem[0],zero
 ; CHECK-NEXT:    callq nearbyint@PLT
 ; CHECK-NEXT:    movsd %xmm0, {{[0-9]+}}(%rsp)
 ; CHECK-NEXT:    fldl {{[0-9]+}}(%rsp)
@@ -3295,59 +3327,65 @@ define <3 x double> @constrained_vector_nearby_v3f64() #0 {
 ; CHECK-NEXT:    # xmm0 = mem[0],zero
 ; CHECK-NEXT:    movsd {{[-0-9]+}}(%r{{[sb]}}p), %xmm1 # 8-byte Reload
 ; CHECK-NEXT:    # xmm1 = mem[0],zero
-; CHECK-NEXT:    addq $24, %rsp
+; CHECK-NEXT:    addq $40, %rsp
 ; CHECK-NEXT:    .cfi_def_cfa_offset 8
 ; CHECK-NEXT:    retq
 ;
 ; AVX-LABEL: constrained_vector_nearby_v3f64:
 ; AVX:       # %bb.0: # %entry
-; AVX-NEXT:    vmovsd {{.*#+}} xmm0 = [4.2200000000000003E+1,0.0E+0]
+; AVX-NEXT:    vmovsd {{.*#+}} xmm0 = mem[0],zero
 ; AVX-NEXT:    vroundsd $12, %xmm0, %xmm0, %xmm0
-; AVX-NEXT:    vroundpd $12, {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm1
+; AVX-NEXT:    vroundpd $12, (%rdi), %xmm1
 ; AVX-NEXT:    vinsertf128 $1, %xmm0, %ymm1, %ymm0
 ; AVX-NEXT:    retq
 entry:
+  %b = load <3 x double>, p...
[truncated]

@RKSimon
Copy link
Collaborator

RKSimon commented Nov 20, 2024

Are we losing all test coverage of the constant cases?

@spavloff
Copy link
Collaborator Author

spavloff commented Nov 20, 2024 •
edited
Loading

Of all the changed functions only some cannot be evaluated at compile time:

@constrained_vector_rint_v2f64()
@constrained_vector_rint_v3f64()
@constrained_vector_rint_v4f64()
@constrained_vector_nearbyint_v2f64()
@constrained_vector_nearby_v3f64()
@constrained_vector_nearbyint_v4f64()

They are put back to the test file.

RKSimon
RKSimon approved these changes Nov 24, 2024
View reviewed changes
Copy link
Collaborator

@RKSimon RKSimon left a comment

Choose a reason for hiding this comment

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

LGTM with one minor

llvm/test/CodeGen/X86/vector-constrained-fp-intrinsics.ll Outdated
@@ -3011,25 +3011,26 @@ entry:
ret <4 x double> %log2
}

define <1 x float> @constrained_vector_rint_v1f32() #0 {
define <1 x float> @constrained_vector_rint_v1f32(ptr %a) #0 {
Copy link
Collaborator

Choose a reason for hiding this comment

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

change the name to constrained_vector_rint_v1f32_var (etc.) to match the naming convention for constant / non-constant tests

@spavloff spavloff merged commit 809c5ac into llvm:main Nov 25, 2024
6 of 8 checks passed
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Reviewers

@RKSimon RKSimon RKSimon approved these changes

@topperc topperc Awaiting requested review from topperc

Assignees

@spavloff spavloff

Labels
backend:X86
Projects
None yet
Milestone
No milestone
Development

Successfully merging this pull request may close these issues.
3 participants
@spavloff @llvmbot @RKSimon