search for: getcouldnotcomput

...tion::ComputeExitLimit(const Loop > *L, BasicBlock *ExitingBlock) { >   // >   // FIXME: we should be able to handle switch instructions (with a > single exit) >   BranchInst *ExitBr = dyn_cast<BranchInst>(ExitingBlock->getTerminator()); > + >   if (ExitBr == 0) return getCouldNotCompute(); >   assert(ExitBr->isConditional() && "If unconditional, it can't be in loop!"); > > +  BranchInst* BrFirstSucc = dyn_cast<BranchInst>(ExitBr-> > +                                            getSuccessor(0)->getTerminator()); > +  BranchInst*...

...-4293,9 +4293,15 @@ ScalarEvolution::ComputeExitLimit(const Loop *L, BasicBlock *ExitingBlock) { // // FIXME: we should be able to handle switch instructions (with a single exit) BranchInst *ExitBr = dyn_cast<BranchInst>(ExitingBlock->getTerminator()); + if (ExitBr == 0) return getCouldNotCompute(); assert(ExitBr->isConditional() && "If unconditional, it can't be in loop!"); + BranchInst* BrFirstSucc = dyn_cast<BranchInst>(ExitBr-> + getSuccessor(0)->getTerminator()); + BranchInst* BrSecondSucc = dyn_cast...

..._cast<BranchInst>(ExitBr-> + getSuccessor(0)->getTerminator()); + BranchInst* BrSecondSucc = dyn_cast<BranchInst>(ExitBr-> + getSuccessor(1)->getTerminator()); + if (ExitBr == 0) return getCouldNotCompute(); assert(ExitBr->isConditional() && "If unconditional, it can't be in loop!"); @@ -4315,8 +4320,12 @@ ScalarEvolution::ComputeExitLimit(const Loop *L, BasicBlock *ExitingBlock) { // if (ExitBr->getSuccessor(0) != L->getHeader() && ExitBr-&gt...

...{ >> >   // >> >   // FIXME: we should be able to handle switch instructions (with a >> > single exit) >> >   BranchInst *ExitBr = >> > dyn_cast<BranchInst>(ExitingBlock->getTerminator()); >> > + >> >   if (ExitBr == 0) return getCouldNotCompute(); >> >   assert(ExitBr->isConditional() && "If unconditional, it can't be in >> > loop!"); >> > >> > +  BranchInst* BrFirstSucc = dyn_cast<BranchInst>(ExitBr-> >> > + >> >  getSuccessor(0)->getTerminator())...

...gt; *L, BasicBlock *ExitingBlock) { > > // > > // FIXME: we should be able to handle switch instructions (with a > > single exit) > > BranchInst *ExitBr = > dyn_cast<BranchInst>(ExitingBlock->getTerminator()); > > + > > if (ExitBr == 0) return getCouldNotCompute(); > > assert(ExitBr->isConditional() && "If unconditional, it can't be in > loop!"); > > > > + BranchInst* BrFirstSucc = dyn_cast<BranchInst>(ExitBr-> > > + > getSuccessor(0)->getTerminator()); > > + BranchInst* BrSecon...

...E: we should be able to handle switch instructions (with a >> >> > single exit) >> >> >   BranchInst *ExitBr = >> >> > dyn_cast<BranchInst>(ExitingBlock->getTerminator()); >> >> > + >> >> >   if (ExitBr == 0) return getCouldNotCompute(); >> >> >   assert(ExitBr->isConditional() && "If unconditional, it can't be in >> >> > loop!"); >> >> > >> >> > +  BranchInst* BrFirstSucc = dyn_cast<BranchInst>(ExitBr-> >> >> > + >&g...

...;> > // FIXME: we should be able to handle switch instructions (with a > >> > single exit) > >> > BranchInst *ExitBr = > >> > dyn_cast<BranchInst>(ExitingBlock->getTerminator()); > >> > + > >> > if (ExitBr == 0) return getCouldNotCompute(); > >> > assert(ExitBr->isConditional() && "If unconditional, it can't be in > >> > loop!"); > >> > > >> > + BranchInst* BrFirstSucc = dyn_cast<BranchInst>(ExitBr-> > >> > + > >> > getSuc...

...to handle switch instructions (with a >>> >> > single exit) >>> >> >   BranchInst *ExitBr = >>> >> > dyn_cast<BranchInst>(ExitingBlock->getTerminator()); >>> >> > + >>> >> >   if (ExitBr == 0) return getCouldNotCompute(); >>> >> >   assert(ExitBr->isConditional() && "If unconditional, it can't be in >>> >> > loop!"); >>> >> > >>> >> > +  BranchInst* BrFirstSucc = dyn_cast<BranchInst>(ExitBr-> >>> &g...

...structions (with a >>>> >> > single exit) >>>> >> >   BranchInst *ExitBr = >>>> >> > dyn_cast<BranchInst>(ExitingBlock->getTerminator()); >>>> >> > + >>>> >> >   if (ExitBr == 0) return getCouldNotCompute(); >>>> >> >   assert(ExitBr->isConditional() && "If unconditional, it can't be in >>>> >> > loop!"); >>>> >> > >>>> >> > +  BranchInst* BrFirstSucc = dyn_cast<BranchInst>(ExitBr->...

On 27.02.2012, at 18:49, Eli Friedman wrote: > On Mon, Feb 27, 2012 at 9:30 AM, Benjamin Kramer > <benny.kra at googlemail.com> wrote: >> >> On 27.02.2012, at 17:13, Николай Лихогруд wrote: >> >>> Dear LLVM, >>> >>> Consider two loops with one interation - >>> First with constant lower bound, second with usual

On Mon, Feb 27, 2012 at 9:30 AM, Benjamin Kramer <benny.kra at googlemail.com> wrote: > > On 27.02.2012, at 17:13, Николай Лихогруд wrote: > >> Dear LLVM, >> >>     Consider two loops with one interation - >>     First with constant lower bound, second with usual non-constant lower bound: >> >>     int main(int argc, char ** argv) >>     {

...Rec->getLoop()->contains(L)) { 04719       // To evaluate this recurrence, we need to know how many times the AddRec 04720       // loop iterates.  Compute this now. 04721       const SCEV *BackedgeTakenCount = getBackedgeTakenCount(AddRec->getLoop()); 04722       if (BackedgeTakenCount == getCouldNotCompute()) return AddRec; 04723 04724       // Then, evaluate the AddRec. 04725       return AddRec->evaluateAtIteration(BackedgeTakenCount, *this); 04726     } The guard checks that !AddRec->getLoop()->contains(L), but not !L->contains(AddRec->getLoop()). Consequently, if getSCEVAtScope...

...(L)) { > 04719       // To evaluate this recurrence, we need to know how many > times the AddRec > 04720       // loop iterates.  Compute this now. > 04721       const SCEV *BackedgeTakenCount = > getBackedgeTakenCount(AddRec->getLoop()); > 04722       if (BackedgeTakenCount == getCouldNotCompute()) return AddRec; > 04723 > 04724       // Then, evaluate the AddRec. > 04725       return AddRec->evaluateAtIteration(BackedgeTakenCount, *this); > 04726     } > > The guard checks that !AddRec->getLoop()->contains(L), but not > !L->contains(AddRec->getLoop())....

Hi Pankaj, Chawla, Pankaj via llvm-dev wrote: > It looks like ScalarEvolution bails out of loop backedge computation if > it cannot prove the IV stride as either positive or negative (based on > loop control condition). I think this logic can be refined for signed IVs. > > Consider this simple loop- > > void foo(int *A, int n, int s) { > > int i; > >

Hi, It looks like ScalarEvolution bails out of loop backedge computation if it cannot prove the IV stride as either positive or negative (based on loop control condition). I think this logic can be refined for signed IVs. Consider this simple loop- void foo(int *A, int n, int s) { int i; for(i=0; i<n; i += s) { A[i]++; } } The IV of this loop has this SCEV form-