maxjhandsome
/
llvm-project
forked from OSchip/llvm-project
1
0
Fork 0
Code Issues Pull Requests Packages Releases Wiki Activity
llvm-project/polly/test/ScopInfo/wraping_signed_expr_slow_2.ll

87 lines
2.7 KiB
LLVM
Raw Normal View History

tests: Drop -polly-detect-unprofitable and -polly-no-early-exit These flags are now always passed to all tests and need to be disabled if not needed. Disabling these flags, rather than passing them to almost all tests, significantly simplfies our RUN: lines. llvm-svn: 249422
2015-10-06 23:36:44 +08:00
; RUN: opt %loadPolly -polly-scops -analyze < %s | FileCheck %s
Use modulo semantic to generate non-integer-overflow assumptions This will allow to generate non-wrap assumptions for integer expressions that are part of the SCoP. We compare the common isl representation of the expression with one computed with modulo semantic. For all parameter combinations they are not equal we can have integer overflows. The nsw flags are respected when the modulo representation is computed, nuw and nw flags are ignored for now. In order to not increase compile time to much, the non-wrap assumptions are collected in a separate boundary context instead of the assumed context. This helps compile time as the boundary context can become complex and it is therefor not advised to use it in other operations except runtime check generation. However, the assumed context is e.g., used to tighten dependences. While the boundary context might help to tighten the assumed context it is doubtful that it will help in practice (it does not effect lnt much) as the boundary (or no-wrap assumptions) only restrict the very end of the possible value range of parameters. PET uses a different approach to compute the no-wrap context, though lnt runs have shown that this version performs slightly better for us. llvm-svn: 247732
2015-09-16 06:52:53 +08:00
;
; This checks that the no-wraps checks will be computed fast as some example
; already showed huge slowdowns even though the inbounds and nsw flags were
; all in place.
;
; // Inspired by itrans8x8 in transform8x8.c from the ldecode benchmark.
; void fast(char *A, char N, char M) {
; for (char i = 0; i < 8; i++) {
; char index0 = i + N;
; char index1 = index0 * 16;
; char index2 = index1 + M;
; A[(short)index2]++;
; }
; }
;
; void slow(char *A, char N, char M) {
; for (char i = 0; i < 8; i++) {
; char index0 = i + N;
; char index1 = index0 * 16;
; short index2 = ((short)index1) + ((short)M);
; A[index2]++;
; }
; }
;
; CHECK: Function: fast
; CHECK: Function: slow
;
target datalayout = "e-m:e-i64:64-f80:128-n8:16:32:64-S128"
define void @fast(i8* %A, i8 %N, i8 %M) {
entry:
br label %for.cond
for.cond: ; preds = %for.inc, %entry
%indvars.iv = phi i8 [ %indvars.iv.next, %for.inc ], [ 0, %entry ]
%exitcond = icmp ne i8 %indvars.iv, 8
br i1 %exitcond, label %for.body, label %for.end
for.body: ; preds = %for.cond
%tmp3 = add nsw i8 %indvars.iv, %N
%mul = mul nsw i8 %tmp3, 16
%add2 = add nsw i8 %mul, %M
%add2ext = sext i8 %add2 to i16
%arrayidx = getelementptr inbounds i8, i8* %A, i16 %add2ext
%tmp4 = load i8, i8* %arrayidx, align 4
%inc = add nsw i8 %tmp4, 1
store i8 %inc, i8* %arrayidx, align 4
br label %for.inc
for.inc: ; preds = %for.body
%indvars.iv.next = add nuw nsw i8 %indvars.iv, 1
br label %for.cond
for.end: ; preds = %for.cond
ret void
}
define void @slow(i8* %A, i8 %N, i8 %M) {
entry:
br label %for.cond
for.cond: ; preds = %for.inc, %entry
%indvars.iv = phi i8 [ %indvars.iv.next, %for.inc ], [ 0, %entry ]
%exitcond = icmp ne i8 %indvars.iv, 8
br i1 %exitcond, label %for.body, label %for.end
for.body: ; preds = %for.cond
%tmp3 = add nsw i8 %indvars.iv, %N
%mul = mul nsw i8 %tmp3, 16
%mulext = sext i8 %mul to i16
%Mext = sext i8 %M to i16
%add2 = add nsw i16 %mulext, %Mext
%arrayidx = getelementptr inbounds i8, i8* %A, i16 %add2
%tmp4 = load i8, i8* %arrayidx, align 4
%inc = add nsw i8 %tmp4, 1
store i8 %inc, i8* %arrayidx, align 4
br label %for.inc
for.inc: ; preds = %for.body
%indvars.iv.next = add nuw nsw i8 %indvars.iv, 1
br label %for.cond
for.end: ; preds = %for.cond
ret void
}