[llvm-commits] [llvm] r60748 - in /llvm/trunk: lib/Analysis/ScalarEvolution.cpp test/Analysis/ScalarEvolution/2008-12-08-FiniteSGE.ll

[Evan Cheng]

Nick, this has caused a number of regressions last night. On x86, I  
see SPASS, 254.gap, and 464.h264ref failures. I'll back it out for now  
and send you a reduced bc file privately.

Thanks,

Evan

On Dec 8, 2008, at 11:25 PM, Nick Lewycky wrote:

> Author: nicholas
> Date: Tue Dec  9 01:25:04 2008
> New Revision: 60748
>
> URL: http://llvm.org/viewvc/llvm-project?rev=60748&view=rev
> Log:
> It's easy to handle SLE/SGE when the loop has a unit stride.
>
> Added:
>    llvm/trunk/test/Analysis/ScalarEvolution/2008-12-08-FiniteSGE.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=60748&r1=60747&r2=60748&view=diff
>
> = 
> = 
> = 
> = 
> = 
> = 
> = 
> = 
> ======================================================================
> --- llvm/trunk/lib/Analysis/ScalarEvolution.cpp (original)
> +++ llvm/trunk/lib/Analysis/ScalarEvolution.cpp Tue Dec  9 01:25:04  
> 2008
> @@ -2924,8 +2924,12 @@
>   if (!R)
>     return true;
>
> -  if (isSigned)
> +  if (isSigned) {
> +    if (SC->getValue()->isOne())
> +      return R->getValue()->isMaxValue(true);
> +
>     return true;  // XXX: because we don't have an sdiv scev.
> +  }
>
>   // If negative, it wraps around every iteration, but we don't care  
> about that.
>   APInt S = SC->getValue()->getValue().abs();
>
> Added: llvm/trunk/test/Analysis/ScalarEvolution/2008-12-08- 
> FiniteSGE.ll
> URL: http://llvm.org/viewvc/llvm-project/llvm/trunk/test/Analysis/ScalarEvolution/2008-12-08-FiniteSGE.ll?rev=60748&view=auto
>
> = 
> = 
> = 
> = 
> = 
> = 
> = 
> = 
> ======================================================================
> --- llvm/trunk/test/Analysis/ScalarEvolution/2008-12-08-FiniteSGE.ll  
> (added)
> +++ llvm/trunk/test/Analysis/ScalarEvolution/2008-12-08-FiniteSGE.ll  
> Tue Dec  9 01:25:04 2008
> @@ -0,0 +1,24 @@
> +; RUN: llvm-as < %s | opt -analyze -scalar-evolution | grep {255  
> iterations}
> +
> +define i32 @foo(i32 %x, i32 %y, i32* %lam, i32* %alp) nounwind {
> +bb1.thread:
> +	br label %bb1
> +
> +bb1:		; preds = %bb1, %bb1.thread
> +	%indvar = phi i32 [ 0, %bb1.thread ], [ %indvar.next, %bb1 ]		;  
> <i32> [#uses=4]
> +	%i.0.reg2mem.0 = sub i32 255, %indvar		; <i32> [#uses=2]
> +	%0 = getelementptr i32* %alp, i32 %i.0.reg2mem.0		; <i32*> [#uses=1]
> +	%1 = load i32* %0, align 4		; <i32> [#uses=1]
> +	%2 = getelementptr i32* %lam, i32 %i.0.reg2mem.0		; <i32*> [#uses=1]
> +	store i32 %1, i32* %2, align 4
> +	%3 = sub i32 254, %indvar		; <i32> [#uses=1]
> +	%4 = icmp slt i32 %3, 0		; <i1> [#uses=1]
> +	%indvar.next = add i32 %indvar, 1		; <i32> [#uses=1]
> +	br i1 %4, label %bb2, label %bb1
> +
> +bb2:		; preds = %bb1
> +	%tmp10 = mul i32 %indvar, %x		; <i32> [#uses=1]
> +	%z.0.reg2mem.0 = add i32 %tmp10, %y		; <i32> [#uses=1]
> +	%5 = add i32 %z.0.reg2mem.0, %x		; <i32> [#uses=1]
> +	ret i32 %5
> +}
>
>
> _______________________________________________
> llvm-commits mailing list
> llvm-commits at cs.uiuc.edu
> http://lists.cs.uiuc.edu/mailman/listinfo/llvm-commits