similar to: missing simplification in ScalarEvolution?

Thanks for the suggestion but datalayout info did not solve the problem! -Pankaj -----Original Message----- From: Philip Reames <listmail at philipreames.com> Sent: Tuesday, August 20, 2019 5:26 PM To: Chawla, Pankaj <pankaj.chawla at intel.com>; llvm-dev at lists.llvm.org Subject: Re: [llvm-dev] missing simplification in ScalarEvolution? Try adding a datalayout with pointer size

Hi Sanjoy, Thanks for the reply! Your approach sounds good to me! I think 1) is legal as address wraparound in unsigned range doesn't make sense given a positive offset, but I am not sure. I think umax will not be added if we can prove the predicate as known. I am not sure whether umax will get simplified if we add nuw to the expressions. -Pankaj -----Original Message----- From: Sanjoy

Hi Sanjoy, I have attached the IR I got by compiling with -O2. This is just before we widen the IV. To get the backedge taken count info I ran indvars on it and then replaced zext with sext. I think regardless of where we decide to add this transformation in the pipeline, it should try to preserve as much information as it can. This means that we should generate sext for signed IVs and

Hi all, It looks like the induction variable simplification pass prefers doing a zero-extension to compute the wider trip count of loops when extending the IV. This can sometimes result in loss of information making ScalarEvolution's analysis conservative which can lead to missed performance opportunities. For example, consider this loopnest- int i, j; for(i=0; i< 40; i++) for(j=0;

Hi Pankaj, On April 14, 2017 at 4:55:16 PM, Chawla, Pankaj (pankaj.chawla at intel.com) wrote: > I have attached the IR I got by compiling with -O2. This is just before we widen the IV. Thanks! > To get the backedge taken count info I ran indvars on it and then replaced zext with sext. > > I think regardless of where we decide to add this transformation in the pipeline, it should

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-

Hi Sanjoy, SCEV is behaving inconsistently when forming SCEV for this zext instruction in the attached test case- %conv5 = zext i32 %dec to i64 If we request a SCEV for the instruction, it returns- (zext i32 {{-1,+,1}<nw><%for.body>,+,-1}<nw><%for.body7> to i64) This can be seen by invoking- $ opt -analyze -scalar-evolution inconsistent-scev-zext.ll But when computing

Hi, +CC Max, Serguei This looks like a textbook case of why caching is hard. We first call getZeroExtendExpr on %dec, and this call does end up returning an add rec. However, in the process of simplifying the zext, it also calls into isLoopBackedgeGuardedByCond which in turn calls getZeroExtendExpr(%dec) again. However, this second (recursive) time, we don't simplify the zext and cache a

Hi Sanjoy, Thanks for investigating the issue! I am more interested in getting the underlying problem fixed rather than making this particular test case work. I think I have more test cases where this problem crops up. I would any day prefer consistent results over compile time savings which can lead to inconsistencies. These inconsistencies require significant developer time to analyze and fix

This sounds fine to me (and sorry for the delay!). -- Sanjoy On Mon, Mar 12, 2018 at 1:09 PM, Chawla, Pankaj <pankaj.chawla at intel.com> wrote: > Hi Sanjoy, > > So what is the verdict on this issue? > > Thanks, > Pankaj > > > -----Original Message----- > From: Chawla, Pankaj > Sent: Monday, February 26, 2018 11:12 AM > To: Sanjoy Das <sanjoy at

Hi Pankaj, On Tue, Mar 13, 2018 at 1:55 PM, Chawla, Pankaj <pankaj.chawla at intel.com> wrote: > Thanks for the reply! > Would it be possible for you to implement this? I don't have cycles for this right now, but if you file a bug I can give this a shot when I have time later. Even in the best case this will have to at least wait until end of April because I'm leaving for a

Hi Sanjoy, Thanks for the reply! Would it be possible for you to implement this? You know the codebase better than I do. Thanks, Pankaj -----Original Message----- From: Sanjoy Das [mailto:sanjoy at playingwithpointers.com] Sent: Tuesday, March 13, 2018 1:34 PM To: Chawla, Pankaj <pankaj.chawla at intel.com> Cc: Maxim Kazantsev <max.kazantsev at azul.com>; Serguei Katkov

Thanks for the helpful reply! I see that we are trying to keep ScalarEvolution stable around instruction ordering. My suggestion would be to not restrict the fix by only recursing on the first operand. By "dominator logic" I meant that if all other 'cheap' checks fail, we should decide by walking the dominator tree to see which instruction's basic block is encountered

Hi Sanjoy, >> I'm a bit apprehensive of adding more caching to solve problems created by caching; but if there is no way out of adding another cache, how about adding a cache that maps SCEV expressions to their simplified versions? Then we could do something like: I may be wrong but I think caching is not an issue in itself, but caching in the presence of self-recursion is. >>

Thanks for fixing this Sanjoy! I do have a few questions/suggestions on the fix if you don't mind. 1) Would this work correctly if the values are call instructions with no operands, like this- %a = foo() %b = bar() 2) From the way this function is set up, it looks like the emphasis is on saving compile time by trading off robustness. Is compile time such a big concern here that we want to

Hi Sanjoy, So what is the verdict on this issue? Thanks, Pankaj -----Original Message----- From: Chawla, Pankaj Sent: Monday, February 26, 2018 11:12 AM To: Sanjoy Das <sanjoy at playingwithpointers.com> Cc: Maxim Kazantsev <max.kazantsev at azul.com>; Serguei Katkov <serguei.katkov at azul.com>; llvm-dev at lists.llvm.org Subject: RE: [SCEV] Inconsistent SCEV formation for

Hi Hal, In that case what is the best way to query whether there is a loop carried dependence between B[j] and A[j] at i-loop level? We were operating under the assumption of 'conservatively correct' behavior of alias analysis in the function scope? Thanks, Pankaj From: Finkel, Hal J. <hfinkel at anl.gov> Sent: Tuesday, March 17, 2020 11:50 AM To: Hiroshi Yamauchi <yamauchi at

Hi, I noticed an inconsistency in how ScalarEvolution orders instruction operands. This inconsistency can result in creation of separate (%a * %b) and (%b * %a) SCEVs as demonstrated by the example IR below (attached as gep-phi.ll)- target datalayout = "e-m:e-p:32:32-f64:32:64-f80:32-n8:16:32-S128" define void @foo(i8* nocapture %arr, i32 %n, i32* %A, i32* %B) local_unnamed_addr {

My understanding is that alias analysis returns results in the function scope, not in loop scope. Since both the phis access both global arrays, that should results in BasicAA conservatively returning MayAlias. I debugged this a little bit and narrowed it down to the section of the code in BasicAAResult::aliasPHI() which has this comment- // Analyse the PHIs' inputs under the assumption