similar to: ConstantRange modelling precision?

----- Original Message ----- > From: "Daniel Berlin via llvm-dev" <llvm-dev at lists.llvm.org> > To: "Joerg Sonnenberger" <joerg at britannica.bec.de>, "llvm-dev" > <llvm-dev at lists.llvm.org> > Sent: Thursday, January 14, 2016 10:38:10 AM > Subject: Re: [llvm-dev] High memory use and LVI/Correlated Value > Propagation > On

Hi all, with the current trunk I have two major cases where clang needs more than 2GB memory for compiling programs with -O2. One is related to GVN and MemoryDependenceAnalysis and has a pending patch. The other is related to the Correlated Value Propagation and Lazy Value Information cache. Attached is a heap profile for one of the relevant test cases. Looking at the sources, I don't see any

On 01/13/2016 04:28 PM, Daniel Berlin via llvm-dev wrote: > > > On Wed, Jan 13, 2016 at 4:23 PM, Joerg Sonnenberger via llvm-dev > <llvm-dev at lists.llvm.org <mailto:llvm-dev at lists.llvm.org>> wrote: > > On Wed, Jan 13, 2016 at 03:38:24PM -0800, Philip Reames wrote: > > I don't think that arbitrary limiting the complexity of the > search

On Mon, Oct 7, 2019 at 11:32 AM Roman Lebedev <lebedev.ri at gmail.com> wrote: > > Bump. Any further thoughts here? > > To recap - i don't really see how this can be a demandedbits problem - we do > demand all those bits, we just know they must be zero. > (i would love to be proven wrong though!) > > Roman. > > On Tue, Oct 1, 2019 at 11:17 PM Roman Lebedev

On Wed, Jan 13, 2016 at 04:28:03PM -0800, Daniel Berlin wrote: > On Wed, Jan 13, 2016 at 4:23 PM, Joerg Sonnenberger via llvm-dev < > llvm-dev at lists.llvm.org> wrote: > > > On Wed, Jan 13, 2016 at 03:38:24PM -0800, Philip Reames wrote: > > > I don't think that arbitrary limiting the complexity of the search is the > > > right approach. There are

On Wed, Jan 13, 2016 at 03:38:24PM -0800, Philip Reames wrote: > I don't think that arbitrary limiting the complexity of the search is the > right approach. There are numerous ways the LVI infrastructure could be > made more memory efficient. Fixing the existing code to be memory efficient > is the right approach. Only once there's no more low hanging fruit should > we

On Sun, Feb 5, 2017 at 3:41 PM, Nuno Lopes <nunoplopes at sapo.pt> wrote: > Thanks for the answers! The plan makes sense to me. > > Regarding phis, what about diamonds, e.g.: > > > define i32 @f(i32 %x) { > br .., label %bb0, label %bb1 > bb0: > %cmp = icmp sge i32 %x, 0 ; x > 0 > br i1 %cmp, label %bb2, label %bb3 > bb1: > %x2 = add nsw

On 2016-09-27 02:28, Philip Reames wrote: > On 09/20/2016 12:05 PM, Matti Niemenmaa via llvm-dev wrote: >> I posted some questions related to implementing inference of nsw/nuw >> flags based on known icmp results to Bug 30428 ( >> https://llvm.org/bugs/show_bug.cgi?id=30428 ) and it was recommended >> that I engage a wider audience by coming here. The minimal context is

On Mon, Jul 24, 2017 at 2:29 PM, John Regehr via llvm-dev < llvm-dev at lists.llvm.org> wrote: > On 7/24/17 3:16 PM, Hongbin Zheng via llvm-dev wrote: > >> Thanks, maybe we could use ScalarEvolution in LazyValueInfo if it is >> available? >> > > This should be fairly easy to try, if you want to propose a patch and run > some experiments. The question is

Looks like LVI actually does have cases for max and min; would it be better to allow ValueTracking to use range analysis instead? - CL > On Aug 1, 2016, at 8:44 PM, Philip Reames <listmail at philipreames.com> wrote: > >> On 07/29/2016 07:08 PM, Carlos Liam via llvm-dev wrote: >> Hi all, >> >> Say we have this IR: >> >> %1 = icmp slt i16 %x, 0

The thing is, we *don't* "not demand" those high bits. We *don't* not care what's in those bits - IR shifts don't mask their shift amounts. I.e we can't replace `x >> (32-y)` with `x >> (-y)`, which would be legal transform should we not demand those bits. We very much demand them. We just know those bits to be zero. And i'm not sure how to convey

On Mon, Aug 7, 2017 at 2:14 AM, Anastasiya Ruzhanskaya via llvm-dev <llvm-dev at lists.llvm.org> wrote: > I am trying to print it like this (maybe here is smth wrong?) > > > LazyValueInfo &LV = getAnalysis<LazyValueInfoWrapperPass>().getLVI(); > DominatorTree &DT = > getAnalysis<DominatorTreeWrapperPass>().getDomTree(); > LV.printLVI(F,

I am primarily interested in phi nodes and their induction variables, in ValueTracking file there is an analysis of them, but if the upper bound is inf, it is not working? 2017-08-07 11:41 GMT+02:00 Anastasiya Ruzhanskaya < anastasiya.ruzhanskaya at frtk.ru>: > So, it is not supported to determine by this instruction : %cmp = icmp slt > i32 %i.03, 99, > that %i.03 = phi i32 [ 0,

For those not following along, startingin r301379, Simplify* in InstructionSimplify now can just take a query struct instead of 8000 optional arguments. Nothing is really new since it used the same thing under the covers. I'm slowly converting the old uses away (deletion of the old APIs is a different question). Staring at most of them, i could just directly convert them using braced list

+1 to alive2.llvm.org On Thu, Jun 18, 2020 at 8:11 AM John Regehr via llvm-dev <llvm-dev at lists.llvm.org> wrote: > > > If we can get Alive1 up there too, that would be even better. I still > > use that to try to prove things where it's not obvious how to express > > the relationships in pure LLVM IR: > > https://rise4fun.com/Alive/NDu > > I don't

Thanks, maybe we could use ScalarEvolution in LazyValueInfo if it is available? On Mon, Jul 24, 2017 at 1:44 PM, John Regehr via llvm-dev < llvm-dev at lists.llvm.org> wrote: > Since they are different static analyses there's no reason to expect them > to agree, but they should both be conservative. In other words, if you like > you can compute the intersection of the two

First of all, sorry for the long mail. Inspired by the excellent analysis Rui did for lld, I decided to do the same for llvm. I'm personally very interested in build-time for LTO configuration, with particular attention to the time spent in the optimizer. Rafael did something similar back in March, so this can be considered as an update. This tries to include a more accurate high-level

Hi all, Say we have this IR: %1 = icmp slt i16 %x, 0 %.x = select i1 %1, i16 0, i16 %x This is the canonical form of what is effectively max(x, 0). From what I can tell LLVM has no facilities to determine from this code that %.x >= 0, so (for example) an SExt on %.x will not be converted to a ZExt. I'm interested in seeing what sorts of changes would be needed to recognize this pattern

Hmmm... why? Duncan Sands wrote: > Hi Stepan, > >> Well... each case is represented as pair<BB, vector<Range> >. Right? > > after thinking about this some more I think you are right to not use > ConstantRange, and instead to build your own set abstraction. > > Ciao, Duncan. > _______________________________________________ > LLVM Developers mailing

Hi, I have a question about ConstantRange::sub(const ConstantRange &Other) at lib/Support/ConstantRange.cpp:524. The code computes the new bounds as follows. APInt NewLower = getLower() - Other.getLower(); APInt NewUpper = getUpper() - Other.getUpper() + 1; Could someone explain this to me? I was expecting something like APInt NewLower = getLower() - Other.getUpper() + 1; APInt