Displaying 2 results from an estimated 2 matches for "_crit_edge5".
Did you mean:
_crit_edge
2013 Aug 16
0
[LLVMdev] [Polly] Analysis of extra compile-time overhead for simple nested loops
...2 %7, %x.04
br label %13
; <label>:13 ; preds = %._crit_edge, %.preheader
%x.1.lcssa = phi i32 [ %12, %._crit_edge ], [ %x.04, %.preheader ]
%14 = add nsw i32 %a.03, 1
%exitcond7 = icmp ne i32 %14, %7
br i1 %exitcond7, label %.preheader, label %._crit_edge5
._crit_edge5: ; preds = %13
%x.1.lcssa.lcssa = phi i32 [ %x.1.lcssa, %13 ]
br label %15
; <label>:15 ; preds = %._crit_edge5, %6
%x.0.lcssa = phi i32 [ %x.1.lcssa.lcssa, %._crit_edge5 ], [ 0, %6 ]
%16 = tail ca...
2013 Aug 15
4
[LLVMdev] [Polly] Analysis of extra compile-time overhead for simple nested loops
Hi all,
I have investigated the 6X extra compile-time overhead when Polly compiles the simple nestedloop benchmark in LLVM-testsuite. (http://188.40.87.11:8000/db_default/v4/nts/31?compare_to=28&baseline=28). Preliminary results show that such compile-time overhead is resulted by the complicated polly-dependence analysis. However, the key seems to be the polly-prepare pass, which introduces