Re: [PATCH-1, rs6000] Put constant into pseudo at expand when it needs two insns [PR86106]

[HAO CHEN GUI <guihaoc@linux.ibm.com>]

Hi Richard,

在 2023/3/16 18:36, Richard Biener 写道:
> On Thu, Mar 16, 2023 at 10:04 AM HAO CHEN GUI <guihaoc@linux.ibm.com> wrote:
>>
>> Hi Richard,
>>
>> 在 2023/3/16 15:57, Richard Biener 写道:
>>> So this is one way around the lack of CSE/PRE of constant operands.  I'd
>>> argue that a better spot for this _might_ be LRA (split the constant out if
>>> there's a free register available), postreload-[g]cse (CSE the constants) and
>>> then maybe cprop_hardreg to combine back single-use constants?
>>>
>>> I'm not sure if careful constraints massaging like adding magic letters to
>>> alternatives with constants to pessimize them for LRA, making them
>>> more expensive than spilling the constant to a register but avoid
>>> secondary reloads with spilling a register to the stack to make room
>>> for the constant, is possible - but in theory a special constraint modifier
>>> for this purpose could be invented.
>>
>> Thanks so much for your advice.
>>
>> cse/gcse doesn't take cost of constant set (the def insn of the constant) into
>> consideration. So it won't replace the register with a constant as it costs 1
>> insn with the register and costs 2 insn with the constant.
> 
> I think it does (and should) cost the constant set (IIRC we had some
> improvements
> there, or at least proposed, during this stage1).  But sure - this is why your
> "trick" works.
> 
It's doable if post-reload gsc costs the constant set. I will draft a patch to
test it.

>> Finally, the single-
>> use constants can't be back to 2 insn.
> 
> And that's because of the issue you point out above?
No. my original concern is the constant can't be back. If post-reload gsc doen't
cost the constant set, the insn with a register always cost less than two insns
with immediates. Commonly the constant set itself costs 2 insn also.
> 
>> Not sure if I understand it correctly.
>> Looking forward to your advice.
> 
> My main point is that CSEing constants has impacts on register pressure
> and thus should probably be done after or within register allocation.  RTL
> expansion itself is probably a bad time to pro-actively split out constants
> even more if, as you say, nothing puts them back.
> 
I agree. Thank a lot.
> Richard.
> 
>> Thanks
>> Gui Haochen

Gui Haochen