llvm dev - Oct 2019 - Which way to lower selects on architectures without conditional moves&

Hello,

We have the architecture without conditional moves. Which way can we 
lower select?

As we know there was the special pass a long time ago, but it was deleted.

commit c3591a0d48ce045bbf5ae0d78a41f3dae4bb99db
Author: Chris Lattner <sabre at nondot.org>
Date:   Tue Feb 19 07:49:17 2008 +0000

     remove the LowerSelect pass.  The last client was the old Sparc 
backend, which is long dead by now.

     llvm-svn: 47323

The problem is that if we use this pass we lose some useful special 
optimizing cases for selects, e.g. replacing it by min/max.

Does anyone have some experience with such architecture?

We're trying to generate some "if-then-else" on SelectionDAG now.

Could anyone show the way to generate "if-then-else" constructs on 
SelectionDAG?

Thanks in advance,

Sergey.

SelectionDAG can't have control flow. For X86 we match selects on old CPUs
or on types that don't support cmov to pseudo cmov instructions during
isel. Those cmov instructions are marked as "UsesCustomInserter" in
the
InstrInfo td files. X86TargetLowering::EmitInstrWithCustomInserter will
then expand those pseudo instructions into control flow by creating new
MachineBasicBlocks and branches. Most of that code is
in X86TargetLowering::EmitLoweredSelect.


~Craig


On Fri, Oct 11, 2019 at 4:20 PM s.ignatov via llvm-dev <
llvm-dev at lists.llvm.org> wrote:
> Hello,
>
> We have the architecture without conditional moves. Which way can we
> lower select?
>
> As we know there was the special pass a long time ago, but it was deleted.
>
> commit c3591a0d48ce045bbf5ae0d78a41f3dae4bb99db
> Author: Chris Lattner <sabre at nondot.org>
> Date:   Tue Feb 19 07:49:17 2008 +0000
>
>      remove the LowerSelect pass.  The last client was the old Sparc
> backend, which is long dead by now.
>
>      llvm-svn: 47323
>
> The problem is that if we use this pass we lose some useful special
> optimizing cases for selects, e.g. replacing it by min/max.
>
> Does anyone have some experience with such architecture?
>
> We're trying to generate some "if-then-else" on SelectionDAG
now.
>
> Could anyone show the way to generate "if-then-else" constructs
on
> SelectionDAG?
>
> Thanks in advance,
>
> Sergey.
>
> _______________________________________________
> LLVM Developers mailing list
> llvm-dev at lists.llvm.org
> https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-dev
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL:
<http://lists.llvm.org/pipermail/llvm-dev/attachments/20191011/94734f22/attachment.html>

Hi Sergey,

You may also want to look at the MSP430 backend implementation. It also uses
Custom Inserters to translate selects into branches. Since it is a simpler
backend (meaning less lines of code) it is easier to follow.

You mentioned that you want to optimise some cases of selects. This is normally
done for you by the DAGCombine code. However, if your architecture has
relatively cheap branches, such ‘optimisations’ might not be always desirable.
The trunk MSP430 and the AVR targets are affected by this problem, and I worked
on a backend that suffers from that too. For those targets, deciding which
selects should not be combined into non-branching sequences is difficult, but
there are a few cases that are pretty obvious. If your target gets affected by
this, you may be interested in supporting this:
http://llvm.1065342.n5.nabble.com/llvm-dev-AVR-MSP430-Code-gen-improvements-for-8-bit-and-16-bit-targets-tc131716.html
<http://llvm.1065342.n5.nabble.com/llvm-dev-AVR-MSP430-Code-gen-improvements-for-8-bit-and-16-bit-targets-tc131716.html>

John

> On 12 Oct 2019, at 01:34, Craig Topper via llvm-dev <llvm-dev at
lists.llvm.org <mailto:llvm-dev at lists.llvm.org>> wrote:
> 
> SelectionDAG can't have control flow. For X86 we match selects on old
CPUs or on types that don't support cmov to pseudo cmov instructions during
isel. Those cmov instructions are marked as "UsesCustomInserter" in
the InstrInfo td files. X86TargetLowering::EmitInstrWithCustomInserter will then
expand those pseudo instructions into control flow by creating new
MachineBasicBlocks and branches. Most of that code is in
X86TargetLowering::EmitLoweredSelect.
> 
> 
> ~Craig
> 
> 
> On Fri, Oct 11, 2019 at 4:20 PM s.ignatov via llvm-dev <llvm-dev at
lists.llvm.org <mailto:llvm-dev at lists.llvm.org>> wrote:
> Hello,
> 
> We have the architecture without conditional moves. Which way can we 
> lower select?
> 
> As we know there was the special pass a long time ago, but it was deleted.
> 
> commit c3591a0d48ce045bbf5ae0d78a41f3dae4bb99db
> Author: Chris Lattner <sabre at nondot.org <mailto:sabre at
nondot.org>>
> Date:   Tue Feb 19 07:49:17 2008 +0000
> 
>      remove the LowerSelect pass.  The last client was the old Sparc 
> backend, which is long dead by now.
> 
>      llvm-svn: 47323
> 
> The problem is that if we use this pass we lose some useful special 
> optimizing cases for selects, e.g. replacing it by min/max.
> 
> Does anyone have some experience with such architecture?
> 
> We're trying to generate some "if-then-else" on SelectionDAG
now.
> 
> Could anyone show the way to generate "if-then-else" constructs
on
> SelectionDAG?
> 
> Thanks in advance,
> 
> Sergey.
> 
> _______________________________________________
> LLVM Developers mailing list
> llvm-dev at lists.llvm.org <mailto:llvm-dev at lists.llvm.org>
> https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-dev
<https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-dev>
> _______________________________________________
> LLVM Developers mailing list
> llvm-dev at lists.llvm.org <mailto:llvm-dev at lists.llvm.org>
> https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-dev
-------------- next part --------------
An HTML attachment was scrubbed...
URL:
<http://lists.llvm.org/pipermail/llvm-dev/attachments/20191012/ed502997/attachment.html>