mirror of
https://github.com/RPCS3/llvm-mirror.git
synced 2024-10-26 06:22:56 +02:00
0a98597e80
This should hopefully have fixed the stage2/stage3 miscompare on the dragonegg
testers.
"LoopVectorize: Use the dependence test utility class
We now no longer need alias analysis - the cases that alias analysis would
handle are now handled as accesses with a large dependence distance.
We can now vectorize loops with simple constant dependence distances.
for (i = 8; i < 256; ++i) {
a[i] = a[i+4] * a[i+8];
}
for (i = 8; i < 256; ++i) {
a[i] = a[i-4] * a[i-8];
}
We would be able to vectorize about 200 more loops (in many cases the cost model
instructs us no to) in the test suite now. Results on x86-64 are a wash.
I have seen one degradation in ammp. Interestingly, the function in which we
now vectorize a loop is never executed so we probably see some instruction
cache effects. There is a 2% improvement in h264ref. There is one or the other
TSCV loop kernel that speeds up.
radar://13681598"
llvm-svn: 184724
223 lines
6.3 KiB
LLVM
223 lines
6.3 KiB
LLVM
; RUN: opt < %s -loop-vectorize -force-vector-width=2 -force-vector-unroll=1 -S | FileCheck %s
|
|
; RUN: opt < %s -loop-vectorize -force-vector-width=4 -force-vector-unroll=1 -S | FileCheck %s -check-prefix=WIDTH
|
|
|
|
target datalayout = "e-p:64:64:64-i1:8:8-i8:8:8-i16:16:16-i32:32:32-i64:64:64-f32:32:32-f64:64:64-v64:64:64-v128:128:128-a0:0:64-s0:64:64-f80:128:128-n8:16:32:64-S128"
|
|
|
|
; Vectorization with dependence checks.
|
|
|
|
; No plausible dependence - can be vectorized.
|
|
; for (i = 0; i < 1024; ++i)
|
|
; A[i] = A[i + 1] + 1;
|
|
|
|
; CHECK: f1_vec
|
|
; CHECK: <2 x i32>
|
|
|
|
define void @f1_vec(i32* %A) {
|
|
entry:
|
|
br label %for.body
|
|
|
|
for.body:
|
|
%indvars.iv = phi i32 [ 0, %entry ], [ %indvars.iv.next, %for.body ]
|
|
%indvars.iv.next = add i32 %indvars.iv, 1
|
|
%arrayidx = getelementptr inbounds i32* %A, i32 %indvars.iv.next
|
|
%0 = load i32* %arrayidx, align 4
|
|
%add1 = add nsw i32 %0, 1
|
|
%arrayidx3 = getelementptr inbounds i32* %A, i32 %indvars.iv
|
|
store i32 %add1, i32* %arrayidx3, align 4
|
|
%exitcond = icmp ne i32 %indvars.iv.next, 1024
|
|
br i1 %exitcond, label %for.body, label %for.end
|
|
|
|
for.end:
|
|
ret void
|
|
}
|
|
|
|
; Plausible dependence of distance 1 - can't be vectorized.
|
|
; for (i = 0; i < 1024; ++i)
|
|
; A[i+1] = A[i] + 1;
|
|
|
|
; CHECK: f2_novec
|
|
; CHECK-NOT: <2 x i32>
|
|
|
|
define void @f2_novec(i32* %A) {
|
|
entry:
|
|
br label %for.body
|
|
|
|
for.body:
|
|
%indvars.iv = phi i32 [ 0, %entry ], [ %indvars.iv.next, %for.body ]
|
|
%arrayidx = getelementptr inbounds i32* %A, i32 %indvars.iv
|
|
%0 = load i32* %arrayidx, align 4
|
|
%add = add nsw i32 %0, 1
|
|
%indvars.iv.next = add i32 %indvars.iv, 1
|
|
%arrayidx3 = getelementptr inbounds i32* %A, i32 %indvars.iv.next
|
|
store i32 %add, i32* %arrayidx3, align 4
|
|
%exitcond = icmp ne i32 %indvars.iv.next, 1024
|
|
br i1 %exitcond, label %for.body, label %for.end
|
|
|
|
for.end:
|
|
ret void
|
|
}
|
|
|
|
; Plausible dependence of distance 2 - can be vectorized with a width of 2.
|
|
; for (i = 0; i < 1024; ++i)
|
|
; A[i+2] = A[i] + 1;
|
|
|
|
; CHECK: f3_vec_len
|
|
; CHECK: <2 x i32>
|
|
|
|
; WIDTH: f3_vec_len
|
|
; WIDTH-NOT: <4 x i32>
|
|
|
|
define void @f3_vec_len(i32* %A) {
|
|
entry:
|
|
br label %for.body
|
|
|
|
for.body:
|
|
%i.01 = phi i32 [ 0, %entry ], [ %inc, %for.body ]
|
|
%idxprom = sext i32 %i.01 to i64
|
|
%arrayidx = getelementptr inbounds i32* %A, i64 %idxprom
|
|
%0 = load i32* %arrayidx, align 4
|
|
%add = add nsw i32 %0, 1
|
|
%add1 = add nsw i32 %i.01, 2
|
|
%idxprom2 = sext i32 %add1 to i64
|
|
%arrayidx3 = getelementptr inbounds i32* %A, i64 %idxprom2
|
|
store i32 %add, i32* %arrayidx3, align 4
|
|
%inc = add nsw i32 %i.01, 1
|
|
%cmp = icmp slt i32 %inc, 1024
|
|
br i1 %cmp, label %for.body, label %for.end
|
|
|
|
for.end:
|
|
ret void
|
|
}
|
|
|
|
; Plausible dependence of distance 1 - cannot be vectorized (without reordering
|
|
; accesses).
|
|
; for (i = 0; i < 1024; ++i) {
|
|
; B[i] = A[i];
|
|
; A[i] = B[i + 1];
|
|
; }
|
|
|
|
; CHECK: f5
|
|
; CHECK-NOT: <2 x i32>
|
|
|
|
define void @f5(i32* %A, i32* %B) {
|
|
entry:
|
|
br label %for.body
|
|
|
|
for.body:
|
|
%indvars.iv = phi i64 [ 0, %entry ], [ %indvars.iv.next, %for.body ]
|
|
%arrayidx = getelementptr inbounds i32* %A, i64 %indvars.iv
|
|
%0 = load i32* %arrayidx, align 4
|
|
%arrayidx2 = getelementptr inbounds i32* %B, i64 %indvars.iv
|
|
store i32 %0, i32* %arrayidx2, align 4
|
|
%indvars.iv.next = add nsw i64 %indvars.iv, 1
|
|
%arrayidx4 = getelementptr inbounds i32* %B, i64 %indvars.iv.next
|
|
%1 = load i32* %arrayidx4, align 4
|
|
store i32 %1, i32* %arrayidx, align 4
|
|
%lftr.wideiv = trunc i64 %indvars.iv.next to i32
|
|
%exitcond = icmp ne i32 %lftr.wideiv, 1024
|
|
br i1 %exitcond, label %for.body, label %for.end
|
|
|
|
for.end:
|
|
ret void
|
|
}
|
|
|
|
; Dependence through a phi node - must not vectorize.
|
|
; for (i = 0; i < 1024; ++i) {
|
|
; a[i+1] = tmp;
|
|
; tmp = a[i];
|
|
; }
|
|
|
|
; CHECK: f6
|
|
; CHECK-NOT: <2 x i32>
|
|
|
|
define i32 @f6(i32* %a, i32 %tmp) {
|
|
entry:
|
|
br label %for.body
|
|
|
|
for.body:
|
|
%indvars.iv = phi i64 [ 0, %entry ], [ %indvars.iv.next, %for.body ]
|
|
%tmp.addr.08 = phi i32 [ %tmp, %entry ], [ %0, %for.body ]
|
|
%indvars.iv.next = add nsw i64 %indvars.iv, 1
|
|
%arrayidx = getelementptr inbounds i32* %a, i64 %indvars.iv.next
|
|
store i32 %tmp.addr.08, i32* %arrayidx, align 4
|
|
%arrayidx3 = getelementptr inbounds i32* %a, i64 %indvars.iv
|
|
%0 = load i32* %arrayidx3, align 4
|
|
%lftr.wideiv = trunc i64 %indvars.iv.next to i32
|
|
%exitcond = icmp ne i32 %lftr.wideiv, 1024
|
|
br i1 %exitcond, label %for.body, label %for.end
|
|
|
|
for.end:
|
|
ret i32 undef
|
|
}
|
|
|
|
; Don't vectorize true loop carried dependencies that are not a multiple of the
|
|
; vector width.
|
|
; Example:
|
|
; for (int i = ...; ++i) {
|
|
; a[i] = a[i-3] + ...;
|
|
; It is a bad idea to vectorize this loop because store-load forwarding will not
|
|
; happen.
|
|
;
|
|
|
|
; CHECK: @nostoreloadforward
|
|
; CHECK-NOT: <2 x i32>
|
|
|
|
define void @nostoreloadforward(i32* %A) {
|
|
entry:
|
|
br label %for.body
|
|
|
|
for.body:
|
|
%indvars.iv = phi i64 [ 16, %entry ], [ %indvars.iv.next, %for.body ]
|
|
%0 = add nsw i64 %indvars.iv, -3
|
|
%arrayidx = getelementptr inbounds i32* %A, i64 %0
|
|
%1 = load i32* %arrayidx, align 4
|
|
%2 = add nsw i64 %indvars.iv, 4
|
|
%arrayidx2 = getelementptr inbounds i32* %A, i64 %2
|
|
%3 = load i32* %arrayidx2, align 4
|
|
%add3 = add nsw i32 %3, %1
|
|
%arrayidx5 = getelementptr inbounds i32* %A, i64 %indvars.iv
|
|
store i32 %add3, i32* %arrayidx5, align 4
|
|
%indvars.iv.next = add i64 %indvars.iv, 1
|
|
%lftr.wideiv = trunc i64 %indvars.iv.next to i32
|
|
%exitcond = icmp ne i32 %lftr.wideiv, 128
|
|
br i1 %exitcond, label %for.body, label %for.end
|
|
|
|
for.end:
|
|
ret void
|
|
}
|
|
|
|
; Example:
|
|
; for (int i = ...; ++i) {
|
|
; a[i] = b[i];
|
|
; c[i] = a[i-3] + ...;
|
|
; It is a bad idea to vectorize this loop because store-load forwarding will not
|
|
; happen.
|
|
;
|
|
|
|
; CHECK: @nostoreloadforward2
|
|
; CHECK-NOT: <2 x i32>
|
|
|
|
define void @nostoreloadforward2(i32* noalias %A, i32* noalias %B, i32* noalias %C) {
|
|
entry:
|
|
br label %for.body
|
|
|
|
for.body:
|
|
%indvars.iv = phi i64 [ 16, %entry ], [ %indvars.iv.next, %for.body ]
|
|
%arrayidx = getelementptr inbounds i32* %B, i64 %indvars.iv
|
|
%0 = load i32* %arrayidx, align 4
|
|
%arrayidx2 = getelementptr inbounds i32* %A, i64 %indvars.iv
|
|
store i32 %0, i32* %arrayidx2, align 4
|
|
%1 = add nsw i64 %indvars.iv, -3
|
|
%arrayidx4 = getelementptr inbounds i32* %A, i64 %1
|
|
%2 = load i32* %arrayidx4, align 4
|
|
%arrayidx6 = getelementptr inbounds i32* %C, i64 %indvars.iv
|
|
store i32 %2, i32* %arrayidx6, align 4
|
|
%indvars.iv.next = add i64 %indvars.iv, 1
|
|
%lftr.wideiv = trunc i64 %indvars.iv.next to i32
|
|
%exitcond = icmp ne i32 %lftr.wideiv, 128
|
|
br i1 %exitcond, label %for.body, label %for.end
|
|
|
|
for.end:
|
|
ret void
|
|
}
|