Displaying 2 results from an estimated 2 matches for "end16".
Did you mean:
end12
2012 Dec 25
2
[LLVMdev] [DragonEgg] Strange call to @"\01__isoc99_fscanf"
...= call %struct._IO_FILE* @fopen(i8* %15, i8* getelementptr
inbounds ([2 x i8]* @.str6, i32 0, i32 0))
store %struct._IO_FILE* %call12, %struct._IO_FILE** %rfile, align 8
%16 = load %struct._IO_FILE** %rfile, align 8
%tobool13 = icmp ne %struct._IO_FILE* %16, null
br i1 %tobool13, label %if.end16, label %if.then14
if.then14: ; preds = %for.end
%17 = load %struct._IO_FILE** @stderr, align 8
%18 = load i8** @filename, align 8
%call15 = call i32 (%struct._IO_FILE*, i8*, ...)*
@fprintf(%struct._IO_FILE* %17, i8* getelementptr inbounds ([27 x i8]*
@....
2012 Jul 30
0
[LLVMdev] IR optimization pass ideas for backend porting before ISel
....inc14
for.inc14: ; preds = %for.inc,
%for.cond1.preheader
;%inc15 = add nsw i32 %i.020, 1
%indvars.iv.next = add i32 %indvars.iv, -1
;%exitcond21 = icmp eq i32 %inc15, %sub
%exitcond21 = icmp eq i32 %indvars.iv.next, 0
br i1 %exitcond21, label %for.end16, label %for.cond1.preheader
I think, the condition for applying such PHI optimization are
1) the same entry points for several PHI instructions in one basic block,
such as %for.inc14 and %entry.
2) the same step but may in two different directions, such as %indvars.iv
and %i.020 are both chan...