[llvm-commits] [llvm] r172534 - /llvm/trunk/test/Transforms/LoopStrengthReduce/post-inc-icmpzero.ll

[Dmitri Gribenko]

On Tue, Jan 15, 2013 at 8:33 PM, Renato Golin <rengolin at systemcall.org> wrote:
> On 15 January 2013 18:21, Rafael Espíndola <rafael.espindola at gmail.com> wrote:
>> I don't think this is true, we do want to produce the same output.
>
> There are limits to that phrase, and I agree with you up to a limit.
>
> The output has been consistent enough for me not to notice, but that
> doesn't mean I can trust it to be consistent (if you get my scientific
> bias). :)
>
> We're not testing for output equality (across architectures) and we
> haven't fixed the host/target description in Clang/LLVM, so I can't
> guarantee that the IR or assembly generated will be identical.
>
> Back to the issue at hand, I couldn't find any difference on the
> outputs on x86_64, A9 and A15, (literally, identical outputs), which
> shows that the issue has nothing to do with ARM and the test should
> also be failing on other architectures as of r171697-ish.
>
> All outputs had:
>
>   %4 = sub i64 %sub.ptr.lhs.cast, %sub.ptr.rhs.cast

OK, but the outputs are different *now*.

Dmitri

-- 
main(i,j){for(i=2;;i++){for(j=2;j<i;j++){if(!(i%j)){j=0;break;}}if
(j){printf("%d\n",i);}}} /*Dmitri Gribenko <gribozavr at gmail.com>*/