[NFC][POWER] add Pre-Commit test case for Inefficient std::bit_floor(x)

[diggerlin]

add a pre-commit test case for Inefficient asm of std::bit_floor(x) for powerpc.

[llvmbot]

@llvm/pr-subscribers-backend-powerpc

Author: zhijian lin (diggerlin)

Changes

add a pre-commit test case for Inefficient asm of std::bit_floor(x) for powerpc.

---

Full diff: https://github.com/llvm/llvm-project/pull/183363.diff

1 Files Affected:

• (added) llvm/test/CodeGen/PowerPC/bit_floor.ll (+75)

diff --git a/llvm/test/CodeGen/PowerPC/bit_floor.ll b/llvm/test/CodeGen/PowerPC/bit_floor.ll
new file mode 100644
index 0000000000000..9d316924c2d3e
--- /dev/null
+++ b/llvm/test/CodeGen/PowerPC/bit_floor.ll
@@ -0,0 +1,75 @@
+; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py UTC_ARGS: --version 6
+; RUN: llc -mtriple=powerpc-ibm-aix-xcoff -ppc-asm-full-reg-names < %s | FileCheck --check-prefix=AIX32 %s
+; RUN: llc -mtriple=powerpc64-ibm-aix-xcoff -ppc-asm-full-reg-names < %s | FileCheck --check-prefixes=COMM64,AIX64  %s
+; RUN: llc -mtriple=powerpc64le-unknown-linux-gnu -ppc-asm-full-reg-names < %s | FileCheck --check-prefixes=COMM64,LNX64 %s
+
+define noundef range(i32 0, -2147483647) i32 @_Z10bitfloor32j(i32 noundef %x) local_unnamed_addr {
+; AIX32-LABEL: _Z10bitfloor32j:
+; AIX32:       # %bb.0: # %entry
+; AIX32-NEXT:    cntlzw r4, r3
+; AIX32-NEXT:    lis r5, -32768
+; AIX32-NEXT:    cmplwi r3, 0
+; AIX32-NEXT:    srw r4, r5, r4
+; AIX32-NEXT:    iseleq r3, 0, r4
+; AIX32-NEXT:    blr
+;
+; COMM64-LABEL: _Z10bitfloor32j:
+; COMM64:       # %bb.0: # %entry
+; COMM64-NEXT:    cntlzw r4, r3
+; AIX64-NEXT:     lis r5, -32768
+; AIX64-NEXT:     cmplwi r3, 0
+; LNX64-NEXT:     cmplwi r3, 0
+; LNX64-NEXT:     lis r5, -32768
+; COMM64-NEXT:    srw r4, r5, r4
+; COMM64-NEXT:    iseleq r3, 0, r4
+; COMM64-NEXT:    blr
+entry:
+  %cmp.i = icmp eq i32 %x, 0
+  %0 = tail call range(i32 0, 33) i32 @llvm.ctlz.i32(i32 %x, i1 true)
+  %shl.i = lshr exact i32 -2147483648, %0
+  %cond.i = select i1 %cmp.i, i32 0, i32 %shl.i
+  ret i32 %cond.i
+}
+
+declare i32 @llvm.ctlz.i32(i32, i1 immarg)
+
+define noundef range(i64 0, -9223372036854775807) i64 @_Z10bitfloor64y(i64 noundef %x) local_unnamed_addr {
+; AIX32-LABEL: _Z10bitfloor64y:
+; AIX32:       # %bb.0: # %entry
+; AIX32-NEXT:    cntlzw r6, r4
+; AIX32-NEXT:    cntlzw r5, r3
+; AIX32-NEXT:    lis r7, -32768
+; AIX32-NEXT:    cmplwi r3, 0
+; AIX32-NEXT:    addi r6, r6, 32
+; AIX32-NEXT:    iseleq r5, r6, r5
+; AIX32-NEXT:    or. r3, r4, r3
+; AIX32-NEXT:    addi r6, r5, -32
+; AIX32-NEXT:    subfic r8, r5, 32
+; AIX32-NEXT:    srw r5, r7, r5
+; AIX32-NEXT:    srw r6, r7, r6
+; AIX32-NEXT:    slw r8, r7, r8
+; AIX32-NEXT:    or r3, r8, r6
+; AIX32-NEXT:    iseleq r4, 0, r3
+; AIX32-NEXT:    iseleq r3, 0, r5
+; AIX32-NEXT:    blr
+;
+; COMM64-LABEL: _Z10bitfloor64y:
+; COMM64:       # %bb.0: # %entry
+; COMM64-NEXT:    li r5, 1
+; COMM64-NEXT:    cntlzd r4, r3
+; COMM64-NEXT:    cmpldi r3, 0
+; COMM64-NEXT:    rldic r5, r5, 63, 0
+; COMM64-NEXT:    srd r4, r5, r4
+; COMM64-NEXT:    iseleq r3, 0, r4
+; COMM64-NEXT:    blr
+
+entry:
+  %cmp.i = icmp eq i64 %x, 0
+  %0 = tail call range(i64 0, 65) i64 @llvm.ctlz.i64(i64 %x, i1 true)
+  %shl.i = lshr exact i64 -9223372036854775808, %0
+  %cond.i = select i1 %cmp.i, i64 0, i64 %shl.i
+  ret i64 %cond.i
+}
+
+declare i64 @llvm.ctlz.i64(i64, i1 immarg)
+

[github-actions]

🐧 Linux x64 Test Results

• 191160 tests passed
• 4966 tests skipped

✅ The build succeeded and all tests passed.

[lei137]

nit: I understand this make things smaller and seems neater --check-prefixes=COMM64,AIX64 but I think in general it's best to use the tool to auto generate the checks if possible. Easier to maintain.

[llvm-ci]

LLVM Buildbot has detected a new failure on builder mlir-rocm-mi200 running on mi200-buildbot while building llvm at step 2 "checkout".

Full details are available at: https://lab.llvm.org/buildbot/#/builders/177/builds/30118

Here is the relevant piece of the build log for the reference

Step 2 (checkout) failure: update (failure)
git version 2.34.1
fatal: unable to access 'https://github.com/llvm/llvm-project.git/': Could not resolve host: github.com
fatal: unable to access 'https://github.com/llvm/llvm-project.git/': Could not resolve host: github.com

[llvm-ci]

LLVM Buildbot has detected a new failure on builder amdgpu-offload-rhel-9-cmake-build-only running on rocm-docker-rhel-9 while building llvm at step 2 "update-annotated-scripts".

Full details are available at: https://lab.llvm.org/buildbot/#/builders/205/builds/37561

Here is the relevant piece of the build log for the reference

Step 2 (update-annotated-scripts) failure: update (failure)
git version 2.43.5
fatal: unable to access 'https://github.com/llvm/llvm-zorg.git/': Could not resolve host: github.com
fatal: unable to access 'https://github.com/llvm/llvm-zorg.git/': Could not resolve host: github.com