search for: sub259

Hi, On Fri, Dec 30, 2011 at 3:09 AM, Tobias Grosser <tobias at grosser.es> wrote: > As it seems my intuition is wrong, I am very eager to see and understand > an example where a search limit of 4000 is really needed. > To make the ball roll again, I attached a testcase that can be tuned to understand the impact on compile time for different sizes of a basic block. One can also

On 12/29/2011 06:32 PM, Hal Finkel wrote: > On Thu, 2011-12-29 at 15:00 +0100, Tobias Grosser wrote: >> On 12/14/2011 01:25 AM, Hal Finkel wrote: >> One thing that I would still like to have is a test case where >> bb-vectorize-search-limit is needed to avoid exponential compile time >> growth and another test case that is not optimized, if >>

...%add14, %mul167 %mul188 = mul nsw i32 %add14, %sub174 %sub195 = sub nsw i32 %add14, %mul188 %mul209 = mul nsw i32 %add14, %sub195 %sub216 = sub nsw i32 %add14, %mul209 %mul231 = mul nsw i32 %add14, %sub216 %sub238 = sub nsw i32 %add14, %mul231 %mul252 = mul nsw i32 %add14, %sub238 %sub259 = sub nsw i32 %add14, %mul252 %mul273 = mul nsw i32 %add14, %sub259 %sub280 = sub nsw i32 %add14, %mul273 %mul294 = mul nsw i32 %add14, %sub280 %sub301 = sub nsw i32 %add14, %mul294 %mul315 = mul nsw i32 %add14, %sub301 %sub322 = sub nsw i32 %add14, %mul315 %mul336 = mul nsw i32 %add1...

....i, i32 0, i32 12 br label %land.rhs.i land.rhs.i: ; preds = %eshup1.exit819.i, %land.rhs.lr.ph.i %digit.0967.i = phi i32 [ %conv239.i, %land.rhs.lr.ph.i ], [ %conv258.i, %eshup1.exit819.i ] %expon.11966.i = phi i32 [ %expon.10.i, %land.rhs.lr.ph.i ], [ %sub259.i, %eshup1.exit819.i ] %call244.i = call fastcc i32 @ecmp(i16* %arraydecay4.i, i16* getelementptr inbounds ([10 x i16]* @ezero, i32 0, i32 0)) nounwind %cmp245.i = icmp eq i32 %call244.i, 0 br i1 %cmp245.i, label %while.end260.i, label %for.body.i873.i for.body.i873.i:...