Re: [llvm-commits] [llvm] r47007 - in /llvm/trunk: lib/Analysis/ScalarEvolution.cpp test/Analysis/ScalarEvolution/2008-02-12-SMAXTripCount.ll

2008-02-13 Thread Wojciech Matyjewicz
Chris Lattner wrote:

 
 Very nice,  please add a comment above the code explaining what is  
 going on though :)
 
Done.

Wojtek
___
llvm-commits mailing list
llvm-commits@cs.uiuc.edu
http://lists.cs.uiuc.edu/mailman/listinfo/llvm-commits


Re: [llvm-commits] [llvm] r47007 - in /llvm/trunk: lib/Analysis/ScalarEvolution.cpp test/Analysis/ScalarEvolution/2008-02-12-SMAXTripCount.ll

2008-02-13 Thread Chris Lattner

On Feb 13, 2008, at 3:51 AM, Wojciech Matyjewicz wrote:

 Chris Lattner wrote:


 Very nice,  please add a comment above the code explaining what is
 going on though :)

 Done.

Very nice, thanks!

-Chris
___
llvm-commits mailing list
llvm-commits@cs.uiuc.edu
http://lists.cs.uiuc.edu/mailman/listinfo/llvm-commits


Re: [llvm-commits] [llvm] r47007 - in /llvm/trunk: lib/Analysis/ScalarEvolution.cpp test/Analysis/ScalarEvolution/2008-02-12-SMAXTripCount.ll

2008-02-12 Thread Chris Lattner

On Feb 12, 2008, at 7:09 AM, Wojciech Matyjewicz wrote:

 Author: wmat
 Date: Tue Feb 12 09:09:36 2008
 New Revision: 47007

 URL: http://llvm.org/viewvc/llvm-project?rev=47007view=rev
 Log:
 Fix PR2002. Suppose n is the initial value for the induction
 variable (with step 1) and m is its final value. Then, the correct  
 trip
 count is SMAX(m,n)-n. Previously, we used SMAX(0,m-n), but m-n may
 overflow and can't in general be interpreted as signed.

Very nice,  please add a comment above the code explaining what is  
going on though :)

-Chris

___
llvm-commits mailing list
llvm-commits@cs.uiuc.edu
http://lists.cs.uiuc.edu/mailman/listinfo/llvm-commits


[llvm-commits] [llvm] r47007 - in /llvm/trunk: lib/Analysis/ScalarEvolution.cpp test/Analysis/ScalarEvolution/2008-02-12-SMAXTripCount.ll

2008-02-12 Thread Wojciech Matyjewicz
Author: wmat
Date: Tue Feb 12 09:09:36 2008
New Revision: 47007

URL: http://llvm.org/viewvc/llvm-project?rev=47007view=rev
Log:
Fix PR2002. Suppose n is the initial value for the induction 
variable (with step 1) and m is its final value. Then, the correct trip 
count is SMAX(m,n)-n. Previously, we used SMAX(0,m-n), but m-n may 
overflow and can't in general be interpreted as signed.

Patch by Nick Lewycky.

Added:
llvm/trunk/test/Analysis/ScalarEvolution/2008-02-12-SMAXTripCount.ll
Modified:
llvm/trunk/lib/Analysis/ScalarEvolution.cpp

Modified: llvm/trunk/lib/Analysis/ScalarEvolution.cpp
URL: 
http://llvm.org/viewvc/llvm-project/llvm/trunk/lib/Analysis/ScalarEvolution.cpp?rev=47007r1=47006r2=47007view=diff

==
--- llvm/trunk/lib/Analysis/ScalarEvolution.cpp (original)
+++ llvm/trunk/lib/Analysis/ScalarEvolution.cpp Tue Feb 12 09:09:36 2008
@@ -2527,19 +2527,17 @@
   if (AddRec-isAffine()) {
 // The number of iterations for {n,+,1}  m, is m-n.  However, we don't
 // know that m is = n on input to the loop.  If it is, the condition
-// returns true zero times.  To handle both cases, we return SMAX(0, m-n).
+// returns true zero times.  To handle both cases, we return SMAX(m, n)-n.
 
 // FORNOW: We only support unit strides.
 SCEVHandle One = SE.getIntegerSCEV(1, RHS-getType());
 if (AddRec-getOperand(1) != One)
   return UnknownValue;
 
-SCEVHandle Iters = SE.getMinusSCEV(RHS, AddRec-getOperand(0));
+SCEVHandle Start = AddRec-getOperand(0);
+SCEVHandle End = isSigned ? SE.getSMaxExpr(RHS, Start) : (SCEVHandle)RHS;
 
-if (isSigned)
-  return SE.getSMaxExpr(SE.getIntegerSCEV(0, RHS-getType()), Iters);
-else
-  return Iters;
+return SE.getMinusSCEV(End, Start);
   }
 
   return UnknownValue;

Added: llvm/trunk/test/Analysis/ScalarEvolution/2008-02-12-SMAXTripCount.ll
URL: 
http://llvm.org/viewvc/llvm-project/llvm/trunk/test/Analysis/ScalarEvolution/2008-02-12-SMAXTripCount.ll?rev=47007view=auto

==
--- llvm/trunk/test/Analysis/ScalarEvolution/2008-02-12-SMAXTripCount.ll (added)
+++ llvm/trunk/test/Analysis/ScalarEvolution/2008-02-12-SMAXTripCount.ll Tue 
Feb 12 09:09:36 2008
@@ -0,0 +1,16 @@
+; RUN: llvm-as  %s | opt -scalar-evolution -analyze | grep {Loop loop: ( 100 
+ ( -100 smax  %n)) iterations!}
+; PR2002
+
+define void @foo(i8 %n) {
+entry:
+   br label %loop
+loop:
+   %i = phi i8 [ -100, %entry ], [ %i.inc, %next ]
+   %cond = icmp slt i8 %i, %n
+   br i1 %cond, label %next, label %return
+next:
+%i.inc = add i8 %i, 1
+   br label %loop
+return:
+   ret void
+}


___
llvm-commits mailing list
llvm-commits@cs.uiuc.edu
http://lists.cs.uiuc.edu/mailman/listinfo/llvm-commits