Re: RISC-V: Folding memory for FP + constant case

[Vineet Gupta <vineetg@rivosinc.com>]

On 8/1/23 16:06, Philipp Tomsich wrote:
> Very helpful! Looks as if regprop for stack_pointer is now either too
> conservative — or one of our patches is missing in everyone's test
> setup; we'll take a closer look.

FWIW, all 5 of them involve a SH2ADD have SP as source in the fold FP 
case which f-m-o seems to be generating a MV for.

>
> On Wed, 2 Aug 2023 at 01:03, Vineet Gupta <vineetg@rivosinc.com> wrote:
>>
>>
>> On 8/1/23 15:07, Philipp Tomsich wrote:
>>> +Manolis Tsamis
>>>
>>> On Tue, 1 Aug 2023 at 23:56, Jeff Law via Gcc-patches
>>> <gcc-patches@gcc.gnu.org> wrote:
>>>>
>>>> On 8/1/23 13:14, Vineet Gupta wrote:
>>>>
>>>>> I have some numbers for f-m-o v3 vs this. Attached here (vs. inline to
>>>>> avoid the Thunderbird mangling the test formatting)
>>>> Thanks.  Of particular importance is the leela change.  My recollection
>>>> was that the f-m-o work also picked up that case.  But if my memory is
>>>> faulty (always a possibility), then that shows a clear case where
>>>> Jivan's work picks up a case not handled by Manolis's work.
>>> f-m-o originally targeted (and benefited) the leela-case.  I wonder if
>>> other optimizations/changes over the last year interfere with this and
>>> what needs to be changed to accomodate this... looks like we need to
>>> revisit against trunk.
>>>
>>> Philipp.
>>>
>>>> And on the other direction we can see that deepsjeng isn't helped by
>>>> Jivan's work, but is helped by Manolis's new pass.
>>>>
>>>> I'd always hoped/expected we'd have cases where one patch clearly helped
>>>> over the other.  While the .25% to .37% improvements for the three most
>>>> impacted benchmarks doesn't move the needle much across the whole suite
>>>> they do add up over time.
>>>>
>>>> Jeff
>> I took a quick look at Leela, the significant difference is from
>> additional insns with SP not getting propagated.
>>
>> e.g.
>>
>>      231b6:    mv    a4,sp
>>      231b8:    sh2add    a5,a5,a4
>>
>> vs.
>>
>>      1e824:    sh2add    a5,a5,sp
>>
>> There are 5 such instances which more or less make up for the delta.
>>
>> -Vineet
>>