c83f23d6ab
The existing heuristics were assuming that every core behaves like an Apple A7, where any extend/shift costs an extra micro-op... but in reality, nothing else behaves like that. On some older Cortex designs, shifts by 1 or 4 cost extra, but all other shifts/extensions are free. On all other cores, as far as I can tell, all shifts/extensions for integer loads are free (i.e. the same cost as an unshifted load). To reflect this, this patch: - Enables aggressive folding of shifts into loads by default. - Removes the old AddrLSLFast feature, since it applies to everything except A7 (and even if you are explicitly targeting A7, we want to assume extensions are free because the code will almost always run on a newer core). - Adds a new feature AddrLSLSlow14 that applies specifically to the Cortex cores where shifts by 1 or 4 cost extra. I didn't add support for AddrLSLSlow14 on the GlobalISel side because it would require a bunch of refactoring to work correctly. Someone can pick this up as a followup.
172 lines
6.0 KiB
LLVM
172 lines
6.0 KiB
LLVM
; RUN: llc -O3 -mtriple arm64-apple-ios3 -aarch64-enable-gep-opt=false %s -o - | FileCheck %s
|
|
; <rdar://problem/13621857>
|
|
|
|
@block = common global ptr null, align 8
|
|
|
|
define i32 @fct(i32 %i1, i32 %i2) {
|
|
; CHECK: @fct
|
|
; Sign extension is used more than once, thus it should not be folded.
|
|
; CodeGenPrepare is not sharing sext across uses, thus this is folded because
|
|
; of that.
|
|
; _CHECK-NOT: , sxtw]
|
|
entry:
|
|
%idxprom = sext i32 %i1 to i64
|
|
%0 = load ptr, ptr @block, align 8
|
|
%arrayidx = getelementptr inbounds i8, ptr %0, i64 %idxprom
|
|
%1 = load i8, ptr %arrayidx, align 1
|
|
%idxprom1 = sext i32 %i2 to i64
|
|
%arrayidx2 = getelementptr inbounds i8, ptr %0, i64 %idxprom1
|
|
%2 = load i8, ptr %arrayidx2, align 1
|
|
%cmp = icmp eq i8 %1, %2
|
|
br i1 %cmp, label %if.end, label %if.then
|
|
|
|
if.then: ; preds = %entry
|
|
%cmp7 = icmp ugt i8 %1, %2
|
|
%conv8 = zext i1 %cmp7 to i32
|
|
br label %return
|
|
|
|
if.end: ; preds = %entry
|
|
%inc = add nsw i32 %i1, 1
|
|
%inc9 = add nsw i32 %i2, 1
|
|
%idxprom10 = sext i32 %inc to i64
|
|
%arrayidx11 = getelementptr inbounds i8, ptr %0, i64 %idxprom10
|
|
%3 = load i8, ptr %arrayidx11, align 1
|
|
%idxprom12 = sext i32 %inc9 to i64
|
|
%arrayidx13 = getelementptr inbounds i8, ptr %0, i64 %idxprom12
|
|
%4 = load i8, ptr %arrayidx13, align 1
|
|
%cmp16 = icmp eq i8 %3, %4
|
|
br i1 %cmp16, label %if.end23, label %if.then18
|
|
|
|
if.then18: ; preds = %if.end
|
|
%cmp21 = icmp ugt i8 %3, %4
|
|
%conv22 = zext i1 %cmp21 to i32
|
|
br label %return
|
|
|
|
if.end23: ; preds = %if.end
|
|
%inc24 = add nsw i32 %i1, 2
|
|
%inc25 = add nsw i32 %i2, 2
|
|
%idxprom26 = sext i32 %inc24 to i64
|
|
%arrayidx27 = getelementptr inbounds i8, ptr %0, i64 %idxprom26
|
|
%5 = load i8, ptr %arrayidx27, align 1
|
|
%idxprom28 = sext i32 %inc25 to i64
|
|
%arrayidx29 = getelementptr inbounds i8, ptr %0, i64 %idxprom28
|
|
%6 = load i8, ptr %arrayidx29, align 1
|
|
%cmp32 = icmp eq i8 %5, %6
|
|
br i1 %cmp32, label %return, label %if.then34
|
|
|
|
if.then34: ; preds = %if.end23
|
|
%cmp37 = icmp ugt i8 %5, %6
|
|
%conv38 = zext i1 %cmp37 to i32
|
|
br label %return
|
|
|
|
return: ; preds = %if.end23, %if.then34, %if.then18, %if.then
|
|
%retval.0 = phi i32 [ %conv8, %if.then ], [ %conv22, %if.then18 ], [ %conv38, %if.then34 ], [ 1, %if.end23 ]
|
|
ret i32 %retval.0
|
|
}
|
|
|
|
define i32 @fct1(i32 %i1, i32 %i2) optsize {
|
|
; CHECK: @fct1
|
|
; Addressing are folded when optimizing for code size.
|
|
; CHECK: , sxtw]
|
|
; CHECK: , sxtw]
|
|
entry:
|
|
%idxprom = sext i32 %i1 to i64
|
|
%0 = load ptr, ptr @block, align 8
|
|
%arrayidx = getelementptr inbounds i8, ptr %0, i64 %idxprom
|
|
%1 = load i8, ptr %arrayidx, align 1
|
|
%idxprom1 = sext i32 %i2 to i64
|
|
%arrayidx2 = getelementptr inbounds i8, ptr %0, i64 %idxprom1
|
|
%2 = load i8, ptr %arrayidx2, align 1
|
|
%cmp = icmp eq i8 %1, %2
|
|
br i1 %cmp, label %if.end, label %if.then
|
|
|
|
if.then: ; preds = %entry
|
|
%cmp7 = icmp ugt i8 %1, %2
|
|
%conv8 = zext i1 %cmp7 to i32
|
|
br label %return
|
|
|
|
if.end: ; preds = %entry
|
|
%inc = add nsw i32 %i1, 1
|
|
%inc9 = add nsw i32 %i2, 1
|
|
%idxprom10 = sext i32 %inc to i64
|
|
%arrayidx11 = getelementptr inbounds i8, ptr %0, i64 %idxprom10
|
|
%3 = load i8, ptr %arrayidx11, align 1
|
|
%idxprom12 = sext i32 %inc9 to i64
|
|
%arrayidx13 = getelementptr inbounds i8, ptr %0, i64 %idxprom12
|
|
%4 = load i8, ptr %arrayidx13, align 1
|
|
%cmp16 = icmp eq i8 %3, %4
|
|
br i1 %cmp16, label %if.end23, label %if.then18
|
|
|
|
if.then18: ; preds = %if.end
|
|
%cmp21 = icmp ugt i8 %3, %4
|
|
%conv22 = zext i1 %cmp21 to i32
|
|
br label %return
|
|
|
|
if.end23: ; preds = %if.end
|
|
%inc24 = add nsw i32 %i1, 2
|
|
%inc25 = add nsw i32 %i2, 2
|
|
%idxprom26 = sext i32 %inc24 to i64
|
|
%arrayidx27 = getelementptr inbounds i8, ptr %0, i64 %idxprom26
|
|
%5 = load i8, ptr %arrayidx27, align 1
|
|
%idxprom28 = sext i32 %inc25 to i64
|
|
%arrayidx29 = getelementptr inbounds i8, ptr %0, i64 %idxprom28
|
|
%6 = load i8, ptr %arrayidx29, align 1
|
|
%cmp32 = icmp eq i8 %5, %6
|
|
br i1 %cmp32, label %return, label %if.then34
|
|
|
|
if.then34: ; preds = %if.end23
|
|
%cmp37 = icmp ugt i8 %5, %6
|
|
%conv38 = zext i1 %cmp37 to i32
|
|
br label %return
|
|
|
|
return: ; preds = %if.end23, %if.then34, %if.then18, %if.then
|
|
%retval.0 = phi i32 [ %conv8, %if.then ], [ %conv22, %if.then18 ], [ %conv38, %if.then34 ], [ 1, %if.end23 ]
|
|
ret i32 %retval.0
|
|
}
|
|
|
|
; CHECK: @test
|
|
; CHECK: , uxtw #2]
|
|
define i32 @test(ptr %array, i8 zeroext %c, i32 %arg) {
|
|
entry:
|
|
%conv = zext i8 %c to i32
|
|
%add = sub i32 0, %arg
|
|
%tobool = icmp eq i32 %conv, %add
|
|
br i1 %tobool, label %if.end, label %if.then
|
|
|
|
if.then: ; preds = %entry
|
|
%idxprom = zext i8 %c to i64
|
|
%arrayidx = getelementptr inbounds i32, ptr %array, i64 %idxprom
|
|
%0 = load volatile i32, ptr %arrayidx, align 4
|
|
%1 = load volatile i32, ptr %arrayidx, align 4
|
|
%add3 = add nsw i32 %1, %0
|
|
br label %if.end
|
|
|
|
if.end: ; preds = %entry, %if.then
|
|
%res.0 = phi i32 [ %add3, %if.then ], [ 0, %entry ]
|
|
ret i32 %res.0
|
|
}
|
|
|
|
|
|
; CHECK: @test2
|
|
; CHECK: , uxtw #2]
|
|
; CHECK: , uxtw #2]
|
|
define i32 @test2(ptr %array, i8 zeroext %c, i32 %arg) optsize {
|
|
entry:
|
|
%conv = zext i8 %c to i32
|
|
%add = sub i32 0, %arg
|
|
%tobool = icmp eq i32 %conv, %add
|
|
br i1 %tobool, label %if.end, label %if.then
|
|
|
|
if.then: ; preds = %entry
|
|
%idxprom = zext i8 %c to i64
|
|
%arrayidx = getelementptr inbounds i32, ptr %array, i64 %idxprom
|
|
%0 = load volatile i32, ptr %arrayidx, align 4
|
|
%1 = load volatile i32, ptr %arrayidx, align 4
|
|
%add3 = add nsw i32 %1, %0
|
|
br label %if.end
|
|
|
|
if.end: ; preds = %entry, %if.then
|
|
%res.0 = phi i32 [ %add3, %if.then ], [ 0, %entry ]
|
|
ret i32 %res.0
|
|
}
|