search for: check_2

...lvmdev at cs.uiuc.edu>, "Ashutosh Nema" <Ashutosh.Nema at amd.com>, "Hal > Finkel" <hfinkel at anl.gov> > Sent: Saturday, May 23, 2015 5:45:27 AM > Subject: RE: [LLVMdev] Alias-based Loop Versioning > > > > > I am not clear what check_1+check_2 really means. For ex: if check_2 > is a subset of check_1 and check_1+check_2 implies union, then the > new check_1+check_2 may be overly conservative resulting in lost > opportunities. > That's correct. However, it is not clear that these opportunities are important in practice....

...if multiple passes would like to check different accesses. E.g. rather than: Checks_1 / \ / \ OrigLoop Checks_2 \ / \ \ / \ \ NoAlias_1 NoAlias_2 \ | / \ | / \ | / Join But instead: Checks_1+Check_2 / \ / \ OrigLoop NoAlias_2 \ / \ / \ / Join This is effectively creating a fast-path and a slow-path version of the loop. We would probably need some metadata annotation so that subsequent passes could amend the same checking blo...

...lvmdev at cs.uiuc.edu>, "Ashutosh Nema" <Ashutosh.Nema at amd.com>, "Hal > Finkel" <hfinkel at anl.gov> > Sent: Saturday, May 23, 2015 5:45:27 AM > Subject: RE: [LLVMdev] Alias-based Loop Versioning > > > > > I am not clear what check_1+check_2 really means. For ex: if check_2 > is a subset of check_1 and check_1+check_2 implies union, then the > new check_1+check_2 may be overly conservative resulting in lost > opportunities. > That's correct. However, it is not clear that these opportunities are important in practice....

...;Ashutosh Nema" <Ashutosh.Nema at amd.com>, "Hal >> Finkel" <hfinkel at anl.gov> >> Sent: Saturday, May 23, 2015 5:45:27 AM >> Subject: RE: [LLVMdev] Alias-based Loop Versioning >> >> >> >> >> I am not clear what check_1+check_2 really means. For ex: if check_2 >> is a subset of check_1 and check_1+check_2 implies union, then the >> new check_1+check_2 may be overly conservative resulting in lost >> opportunities. >> > > That's correct. However, it is not clear that these opportunities a...