Daniel Berlin via llvm-dev
2017-Apr-14 14:52 UTC
[llvm-dev] TBAA falsely reporting may alias?
On Fri, Apr 14, 2017 at 6:12 AM, Hal Finkel via llvm-dev < llvm-dev at lists.llvm.org> wrote:> > On 04/14/2017 08:03 AM, Kavon Farvardin via llvm-dev wrote: > >> Thanks for the explanation Sanjoy! >> >> Another question I have is: are there are any passes that invalidate or >> make the TBAA analysis information less precise? >> > > Some transformation drop TBAA information when they might invalidate it, > but it should be preserved in most places. > >FWIW, I believe some folks at Google in a different group have come to the conclusion it is dropped in about 20-30% of cases. (I said "file bugs") -------------- next part -------------- An HTML attachment was scrubbed... URL: <http://lists.llvm.org/pipermail/llvm-dev/attachments/20170414/179ff6bf/attachment.html>
Kavon Farvardin via llvm-dev
2017-Apr-17 10:08 UTC
[llvm-dev] TBAA falsely reporting may alias?
By "dropped" do you mean that the metadata is not propagated/preserved during the transformation? If anyone knows which passes in particular are doing this, I would be happy to try to come up with a patch! ~kavon> On Apr 14, 2017, at 3:52 PM, Daniel Berlin <dberlin at dberlin.org> wrote: > > > > On Fri, Apr 14, 2017 at 6:12 AM, Hal Finkel via llvm-dev <llvm-dev at lists.llvm.org <mailto:llvm-dev at lists.llvm.org>> wrote: > > On 04/14/2017 08:03 AM, Kavon Farvardin via llvm-dev wrote: > Thanks for the explanation Sanjoy! > > Another question I have is: are there are any passes that invalidate or make the TBAA analysis information less precise? > > Some transformation drop TBAA information when they might invalidate it, but it should be preserved in most places. > > > FWIW, I believe some folks at Google in a different group have come to the conclusion it is dropped in about 20-30% of cases. > (I said "file bugs") >-------------- next part -------------- An HTML attachment was scrubbed... URL: <http://lists.llvm.org/pipermail/llvm-dev/attachments/20170417/a4bbfd68/attachment.html>
Daniel Berlin via llvm-dev
2017-Apr-17 16:36 UTC
[llvm-dev] TBAA falsely reporting may alias?
On Mon, Apr 17, 2017, 3:08 AM Kavon Farvardin <kavon at farvard.in> wrote:> By "dropped" do you mean that the metadata is not propagated/preserved > during the transformation? >Yes> > If anyone knows which passes in particular are doing this, I would be > happy to try to come up with a patch! >+Ben> > ~kavon > > On Apr 14, 2017, at 3:52 PM, Daniel Berlin <dberlin at dberlin.org> wrote: > > > > On Fri, Apr 14, 2017 at 6:12 AM, Hal Finkel via llvm-dev < > llvm-dev at lists.llvm.org> wrote: > >> >> On 04/14/2017 08:03 AM, Kavon Farvardin via llvm-dev wrote: >> >>> Thanks for the explanation Sanjoy! >>> >>> Another question I have is: are there are any passes that invalidate or >>> make the TBAA analysis information less precise? >>> >> >> Some transformation drop TBAA information when they might invalidate it, >> but it should be preserved in most places. >> >> > FWIW, I believe some folks at Google in a different group have come to the > conclusion it is dropped in about 20-30% of cases. > (I said "file bugs") > > >-------------- next part -------------- An HTML attachment was scrubbed... URL: <http://lists.llvm.org/pipermail/llvm-dev/attachments/20170417/ce77ba36/attachment.html>