Re: [ARM] PR 67591 ARM v8 Thumb IT blocks are deprecated

[Christophe Lyon <christophe dot lyon at linaro dot org>]

Hi,

On 12 October 2016 at 11:22, Christophe Lyon <christophe.lyon@linaro.org> wrote:
> On 12 October 2016 at 11:14, Kyrill Tkachov <kyrylo.tkachov@foss.arm.com> wrote:
>>
>> On 12/10/16 09:59, Christophe Lyon wrote:
>>>
>>> Hi Kyrill,
>>>
>>> On 7 October 2016 at 17:00, Kyrill Tkachov <kyrylo.tkachov@foss.arm.com>
>>> wrote:
>>>>
>>>> Hi Christophe,
>>>>
>>>>
>>>> On 07/09/16 21:05, Christophe Lyon wrote:
>>>>>
>>>>> Hi,
>>>>>
>>>>> The attached patch is a first part to solve PR 67591: it removes
>>>>> several occurrences of "IT blocks containing 32-bit Thumb
>>>>> instructions are deprecated in ARMv8" messages in the
>>>>> gcc/g++/libstdc++/fortran testsuites.
>>>>>
>>>>> It does not remove them all yet. This patch only modifies the
>>>>> *cmp_and, *cmp_ior, *ior_scc_scc, *ior_scc_scc_cmp,
>>>>> *and_scc_scc and *and_scc_scc_cmp patterns.
>>>>> Additional work is required in sub_shiftsi etc, at least.
>>>>> I've started looking at these, but I decided I could already
>>>>> post this self-contained patch to check if this implementation
>>>>> is OK.
>>>>>
>>>>> Regarding *cmp_and and *cmp_ior patterns, the addition of the
>>>>> enabled_for_depr_it attribute is aggressive in the sense that it keeps
>>>>> only the alternatives with 'l' and 'Py' constraints, while in some
>>>>> cases the constraints could be relaxed. Indeed, these 2 patterns can
>>>>> swap their input comparisons, meaning that any of them can be emitted
>>>>> in the IT-block, and is thus subject to the ARMv8 deprecation.
>>>>> The generated code is possibly suboptimal in the cases where the
>>>>> operands are not swapped, since 'r' could be used.
>>>>>
>>>>> Cross-tested on arm-none-linux-gnueabihf with -mthumb/-march=armv8-a
>>>>> and --with-cpu=cortex-a57 --with-mode=thumb, showing only improvements:
>>>>>
>>>>>
>>>>> http://people.linaro.org/~christophe.lyon/cross-validation/gcc-test-patches/239850-depr-it-4/report-build-info.html
>>>>>
>>>>> Bootstrapped OK on armv8l HW.
>>>>>
>>>>> Is this OK?
>>>>>
>>>>> Thanks,
>>>>>
>>>>> Christophe
>>>>
>>>>
>>>>   (define_insn_and_split "*ior_scc_scc"
>>>> -  [(set (match_operand:SI 0 "s_register_operand" "=Ts")
>>>> +  [(set (match_operand:SI 0 "s_register_operand" "=Ts,Ts")
>>>>          (ior:SI (match_operator:SI 3 "arm_comparison_operator"
>>>> -                [(match_operand:SI 1 "s_register_operand" "r")
>>>> -                 (match_operand:SI 2 "arm_add_operand" "rIL")])
>>>> +                [(match_operand:SI 1 "s_register_operand" "r,l")
>>>> +                 (match_operand:SI 2 "arm_add_operand" "rIL,lPy")])
>>>>                  (match_operator:SI 6 "arm_comparison_operator"
>>>> -                [(match_operand:SI 4 "s_register_operand" "r")
>>>> -                 (match_operand:SI 5 "arm_add_operand" "rIL")])))
>>>> +                [(match_operand:SI 4 "s_register_operand" "r,l")
>>>> +                 (match_operand:SI 5 "arm_add_operand" "rIL,lPy")])))
>>>>
>>>> Can you please put the more restrictive alternatives (lPy) first?
>>>> Same with the other patterns your patch touches.
>>>> Ok with that change if a rebased testing run is ok.
>>>> Sorry for the delay in reviewing.
>>>>
>>> OK, I will update my patch accordingly.
>>>
>>> However, when I discussed this with Ramana during the Cauldron,
>>> he requested benchmark results. So far, I was able to run spec2006
>>> on an APM machine, and I'm seeing performance changes in the
>>> range 11% improvement (465.tonto) to 7% degradation (433.milc).
>>>
>>> Would that be acceptable?
>>
>>
>> Those sound like quite large swings.
> Indeed, but most are in the -1%-+1% range.
>
>> Are you sure the machine was not running anything else at the time
>> or playing tricks with frequency scaling?
> No, I had no such guarantee. I used this machine temporarily,
> first to check that bootstrap worked. I planed to use another
> board with an A57 "standard" microarch for proper
> benchmarking, but I'm not sure when I'll have access to it
> (wrt to e/o gcc stage1), that's why I reported these early
> figures.
>
>> Did all iterations of SPEC show a consistent difference?
>>
>> If the changes are consistent, could you have a look at the codegen
>> to see if there are any clues to the differences?
> I will update my patch according to your comment, re-run the bench
> and have a deeper look at the codegen differences.
>

I have finally been able to run benchmarks with my patch updated
according to your comment, on new machines where we have
better control of the environment (frequency, etc...).

These machines use cortex-a57 CPUs and spec2006 shows little
difference with and without this patch.

Comparing several runs with and without the patch:
- gcc is about 1% slower
- povray about 1% faster
- omnetpp about 1.5% faster

The others are in the noise level.

OK for trunk?

Thanks,

Christophe


>> I'd like to get an explanation for these differences before committing
>> this patch. If they are just an effect of the more restrictive constraints
>> then there may be not much we can do, but I'd like to make sure there's not
>> anything else unexpected going on.
>>
> Thanks,
>
> Christophe
>
>>>
>>> The number of warnings (IT blocks containing 32-bit Thumb instructions
>>> are deprecated in ARMv8)
>>> was 712 without my patch and 122 with it. (using the hosts's binutils
>>> 2.24/ubuntu).
>>> I expected some warning, since as I said earlier other patterns need
>>> to be updated.
>>
>>
>> Understood. That's fine.
>>
>> Thanks,
>> Kyrill
>>
>>
>>>
>>> Christophe
>>>
>>>
>>>> Thanks for improving this area!
>>>> Kyrill
>>>>
>>

Attachment: depr-it.log.txt
Description: Text document

Attachment: depr-it.patch.txt
Description: Text document