Displaying 3 results from an estimated 3 matches for "r123838".
2011 Jan 20
0
[LLVMdev] induction variable computation not preserving scev
On Jan 19, 2011, at 2:03 PM, Nick Lewycky wrote:
> On 19 January 2011 13:01, Dan Gohman <gohman at apple.com> wrote:
>
> On Jan 18, 2011, at 12:32 AM, Nick Lewycky wrote:
>
> > Hi,
> >
> > I tracked down a bug in indvars where we weren't updating SCEV properly. The attached patch shows the fix to this bug with a testcase, but it also causes five new test
2011 Jan 20
1
[LLVMdev] induction variable computation not preserving scev
...nt.ll
and as far as I can see, they're not just silly errors because we changed
the SCEVs slightly; we're actually generating worse code when with
forgetLoop() in indvars. That's what I don't understand.
I debugged one example and that led to the simplification I added to SCEV in
r123838, but we're still not eliminating the sext instructions that the
testcases expect.
Nick
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.llvm.org/pipermail/llvm-dev/attachments/20110120/ddb95b83/attachment.html>
2011 Jan 19
2
[LLVMdev] induction variable computation not preserving scev
On 19 January 2011 13:01, Dan Gohman <gohman at apple.com> wrote:
>
> On Jan 18, 2011, at 12:32 AM, Nick Lewycky wrote:
>
> > Hi,
> >
> > I tracked down a bug in indvars where we weren't updating SCEV properly.
> The attached patch shows the fix to this bug with a testcase, but it also
> causes five new test failures.
>
> Indvars isn't