[llvm-dev] PostRAScheduler

Previous Topic Next Topic
 
classic Classic list List threaded Threaded
3 messages Options
Reply | Threaded
Open this post in threaded view
|

[llvm-dev] PostRAScheduler

Robin Eklind via llvm-dev
Is there any specific documentation on this? Is there a point of contact for this file that I might bug?

So I have an instruction that needs noops inserted and it appears that the noops are being inserted in the EmitSchedule function of the PostRAScheduler. From the loop in EmitSchedule it looks like it's inserting noops whenever SUnit is NULL. I don't see anything in the DAG that appears odd or any empty nodes, how are these NULL SUnits inserted into the Sequence?

Thanks,

Ryan

_______________________________________________
LLVM Developers mailing list
[hidden email]
http://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-dev
Reply | Threaded
Open this post in threaded view
|

Re: [llvm-dev] PostRAScheduler

Robin Eklind via llvm-dev
On 20.06.2018 17:51, Ryan Taylor via llvm-dev wrote:
> Is there any specific documentation on this? Is there a point of contact
> for this file that I might bug?
>
> So I have an instruction that needs noops inserted and it appears that
> the noops are being inserted in the EmitSchedule function of the
> PostRAScheduler. From the loop in EmitSchedule it looks like it's
> inserting noops whenever SUnit is NULL. I don't see anything in the DAG
> that appears odd or any empty nodes, how are these NULL SUnits inserted
> into the Sequence?

I'm not an expert on this, but I believe they are inserted via the
target's ScheduleHazardRecognizer implementation.

Cheers,
Nicolai

>
> Thanks,
>
> Ryan
>
>
> _______________________________________________
> LLVM Developers mailing list
> [hidden email]
> http://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-dev
>


--
Lerne, wie die Welt wirklich ist,
Aber vergiss niemals, wie sie sein sollte.
_______________________________________________
LLVM Developers mailing list
[hidden email]
http://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-dev
Reply | Threaded
Open this post in threaded view
|

Re: [llvm-dev] PostRAScheduler

Robin Eklind via llvm-dev
Yes thanks, I saw that. I missed the target dependent pass. 

On Thu, Jun 21, 2018, 3:17 AM Nicolai Hähnle via llvm-dev <[hidden email]> wrote:
On 20.06.2018 17:51, Ryan Taylor via llvm-dev wrote:
> Is there any specific documentation on this? Is there a point of contact
> for this file that I might bug?
>
> So I have an instruction that needs noops inserted and it appears that
> the noops are being inserted in the EmitSchedule function of the
> PostRAScheduler. From the loop in EmitSchedule it looks like it's
> inserting noops whenever SUnit is NULL. I don't see anything in the DAG
> that appears odd or any empty nodes, how are these NULL SUnits inserted
> into the Sequence?

I'm not an expert on this, but I believe they are inserted via the
target's ScheduleHazardRecognizer implementation.

Cheers,
Nicolai

>
> Thanks,
>
> Ryan
>
>
> _______________________________________________
> LLVM Developers mailing list
> [hidden email]
> http://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-dev
>


--
Lerne, wie die Welt wirklich ist,
Aber vergiss niemals, wie sie sein sollte.
_______________________________________________
LLVM Developers mailing list
[hidden email]
http://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-dev

_______________________________________________
LLVM Developers mailing list
[hidden email]
http://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-dev