[llvm] r267373 - [X86] Add a complete set of tests for all operand sizes of cttz/ctlz with and without zero undef being lowered to bsf/bsr.
Craig Topper via llvm-commits
llvm-commits at lists.llvm.org
Sun Apr 24 18:01:15 PDT 2016
Author: ctopper
Date: Sun Apr 24 20:01:15 2016
New Revision: 267373
URL: http://llvm.org/viewvc/llvm-project?rev=267373&view=rev
Log:
[X86] Add a complete set of tests for all operand sizes of cttz/ctlz with and without zero undef being lowered to bsf/bsr.
Modified:
llvm/trunk/test/CodeGen/X86/clz.ll
Modified: llvm/trunk/test/CodeGen/X86/clz.ll
URL: http://llvm.org/viewvc/llvm-project/llvm/trunk/test/CodeGen/X86/clz.ll?rev=267373&r1=267372&r2=267373&view=diff
==============================================================================
--- llvm/trunk/test/CodeGen/X86/clz.ll (original)
+++ llvm/trunk/test/CodeGen/X86/clz.ll Sun Apr 24 20:01:15 2016
@@ -87,6 +87,41 @@ define i64 @ctlz_i64(i64 %x) {
ret i64 %tmp
}
+define i8 @ctlz_i8_zero_test(i8 %n) {
+; Generate a test and branch to handle zero inputs because bsr/bsf are very slow.
+
+; CHECK-LABEL: ctlz_i8_zero_test:
+; CHECK: # BB#0:
+; CHECK-NEXT: movb $8, %al
+; CHECK-NEXT: testb %dil, %dil
+; CHECK-NEXT: je .LBB8_2
+; CHECK-NEXT: # BB#1: # %cond.false
+; CHECK-NEXT: movzbl %dil, %eax
+; CHECK-NEXT: bsrl %eax, %eax
+; CHECK-NEXT: xorl $7, %eax
+; CHECK-NEXT: .LBB8_2: # %cond.end
+; CHECK-NEXT: retq
+ %tmp1 = call i8 @llvm.ctlz.i8(i8 %n, i1 false)
+ ret i8 %tmp1
+}
+
+define i16 @ctlz_i16_zero_test(i16 %n) {
+; Generate a test and branch to handle zero inputs because bsr/bsf are very slow.
+
+; CHECK-LABEL: ctlz_i16_zero_test:
+; CHECK: # BB#0:
+; CHECK-NEXT: movw $16, %ax
+; CHECK-NEXT: testw %di, %di
+; CHECK-NEXT: je .LBB9_2
+; CHECK-NEXT: # BB#1: # %cond.false
+; CHECK-NEXT: bsrw %di, %ax
+; CHECK-NEXT: xorl $15, %eax
+; CHECK-NEXT: .LBB9_2: # %cond.end
+; CHECK-NEXT: retq
+ %tmp1 = call i16 @llvm.ctlz.i16(i16 %n, i1 false)
+ ret i16 %tmp1
+}
+
define i32 @ctlz_i32_zero_test(i32 %n) {
; Generate a test and branch to handle zero inputs because bsr/bsf are very slow.
@@ -94,16 +129,98 @@ define i32 @ctlz_i32_zero_test(i32 %n) {
; CHECK: # BB#0:
; CHECK-NEXT: movl $32, %eax
; CHECK-NEXT: testl %edi, %edi
-; CHECK-NEXT: je .LBB8_2
+; CHECK-NEXT: je .LBB10_2
; CHECK-NEXT: # BB#1: # %cond.false
; CHECK-NEXT: bsrl %edi, %eax
; CHECK-NEXT: xorl $31, %eax
-; CHECK-NEXT: .LBB8_2: # %cond.end
+; CHECK-NEXT: .LBB10_2: # %cond.end
; CHECK-NEXT: retq
%tmp1 = call i32 @llvm.ctlz.i32(i32 %n, i1 false)
ret i32 %tmp1
}
+define i64 @ctlz_i64_zero_test(i64 %n) {
+; Generate a test and branch to handle zero inputs because bsr/bsf are very slow.
+
+; CHECK-LABEL: ctlz_i64_zero_test:
+; CHECK: # BB#0:
+; CHECK-NEXT: movl $64, %eax
+; CHECK-NEXT: testq %rdi, %rdi
+; CHECK-NEXT: je .LBB11_2
+; CHECK-NEXT: # BB#1: # %cond.false
+; CHECK-NEXT: bsrq %rdi, %rax
+; CHECK-NEXT: xorq $63, %rax
+; CHECK-NEXT: .LBB11_2: # %cond.end
+; CHECK-NEXT: retq
+ %tmp1 = call i64 @llvm.ctlz.i64(i64 %n, i1 false)
+ ret i64 %tmp1
+}
+
+define i8 @cttz_i8_zero_test(i8 %n) {
+; Generate a test and branch to handle zero inputs because bsr/bsf are very slow.
+
+; CHECK-LABEL: cttz_i8_zero_test:
+; CHECK: # BB#0:
+; CHECK-NEXT: movb $8, %al
+; CHECK-NEXT: testb %dil, %dil
+; CHECK-NEXT: je .LBB12_2
+; CHECK-NEXT: # BB#1: # %cond.false
+; CHECK-NEXT: movzbl %dil, %eax
+; CHECK-NEXT: bsfl %eax, %eax
+; CHECK-NEXT: .LBB12_2: # %cond.end
+; CHECK-NEXT: retq
+ %tmp1 = call i8 @llvm.cttz.i8(i8 %n, i1 false)
+ ret i8 %tmp1
+}
+
+define i16 @cttz_i16_zero_test(i16 %n) {
+; Generate a test and branch to handle zero inputs because bsr/bsf are very slow.
+
+; CHECK-LABEL: cttz_i16_zero_test:
+; CHECK: # BB#0:
+; CHECK-NEXT: movw $16, %ax
+; CHECK-NEXT: testw %di, %di
+; CHECK-NEXT: je .LBB13_2
+; CHECK-NEXT: # BB#1: # %cond.false
+; CHECK-NEXT: bsfw %di, %ax
+; CHECK-NEXT: .LBB13_2: # %cond.end
+; CHECK-NEXT: retq
+ %tmp1 = call i16 @llvm.cttz.i16(i16 %n, i1 false)
+ ret i16 %tmp1
+}
+
+define i32 @cttz_i32_zero_test(i32 %n) {
+; Generate a test and branch to handle zero inputs because bsr/bsf are very slow.
+
+; CHECK-LABEL: cttz_i32_zero_test:
+; CHECK: # BB#0:
+; CHECK-NEXT: movl $32, %eax
+; CHECK-NEXT: testl %edi, %edi
+; CHECK-NEXT: je .LBB14_2
+; CHECK-NEXT: # BB#1: # %cond.false
+; CHECK-NEXT: bsfl %edi, %eax
+; CHECK-NEXT: .LBB14_2: # %cond.end
+; CHECK-NEXT: retq
+ %tmp1 = call i32 @llvm.cttz.i32(i32 %n, i1 false)
+ ret i32 %tmp1
+}
+
+define i64 @cttz_i64_zero_test(i64 %n) {
+; Generate a test and branch to handle zero inputs because bsr/bsf are very slow.
+
+; CHECK-LABEL: cttz_i64_zero_test:
+; CHECK: # BB#0:
+; CHECK-NEXT: movl $64, %eax
+; CHECK-NEXT: testq %rdi, %rdi
+; CHECK-NEXT: je .LBB15_2
+; CHECK-NEXT: # BB#1: # %cond.false
+; CHECK-NEXT: bsfq %rdi, %rax
+; CHECK-NEXT: .LBB15_2: # %cond.end
+; CHECK-NEXT: retq
+ %tmp1 = call i64 @llvm.cttz.i64(i64 %n, i1 false)
+ ret i64 %tmp1
+}
+
define i32 @ctlz_i32_fold_cmov(i32 %n) {
; Don't generate the cmovne when the source is known non-zero (and bsr would
; not set ZF).
@@ -115,11 +232,11 @@ define i32 @ctlz_i32_fold_cmov(i32 %n) {
; CHECK: # BB#0:
; CHECK-NEXT: orl $1, %edi
; CHECK-NEXT: movl $32, %eax
-; CHECK-NEXT: je .LBB9_2
+; CHECK-NEXT: je .LBB16_2
; CHECK-NEXT: # BB#1: # %cond.false
; CHECK-NEXT: bsrl %edi, %eax
; CHECK-NEXT: xorl $31, %eax
-; CHECK-NEXT: .LBB9_2: # %cond.end
+; CHECK-NEXT: .LBB16_2: # %cond.end
; CHECK-NEXT: retq
%or = or i32 %n, 1
%tmp1 = call i32 @llvm.ctlz.i32(i32 %or, i1 false)
@@ -148,11 +265,11 @@ define i32 @ctlz_bsr_zero_test(i32 %n) {
; CHECK: # BB#0:
; CHECK-NEXT: movl $32, %eax
; CHECK-NEXT: testl %edi, %edi
-; CHECK-NEXT: je .LBB11_2
+; CHECK-NEXT: je .LBB18_2
; CHECK-NEXT: # BB#1: # %cond.false
; CHECK-NEXT: bsrl %edi, %eax
; CHECK-NEXT: xorl $31, %eax
-; CHECK-NEXT: .LBB11_2: # %cond.end
+; CHECK-NEXT: .LBB18_2: # %cond.end
; CHECK-NEXT: xorl $31, %eax
; CHECK-NEXT: retq
%ctlz = call i32 @llvm.ctlz.i32(i32 %n, i1 false)
More information about the llvm-commits
mailing list