Re: [Mesa-dev] [PATCH 12/59] intel/compiler: handle b2i/b2f with other integer conversion opcodes

2018-12-07 Thread Jason Ekstrand 
This'll have to be rebased on dca6cd9ce6510 but otherwise looks fine.  I've
been a bit annoyed by this myself.

Reviewed-by: Jason Ekstrand 

Incidentally, this could also be lowered in NIR Not sure if we want to
but there it is.

On Tue, Dec 4, 2018 at 1:18 AM Iago Toral Quiroga  wrote:

> Since we handle booleans as integers this makes more sense.
> ---
>  src/intel/compiler/brw_fs_nir.cpp | 10 +-
>  1 file changed, 5 insertions(+), 5 deletions(-)
>
> diff --git a/src/intel/compiler/brw_fs_nir.cpp
> b/src/intel/compiler/brw_fs_nir.cpp
> index 9f3d3bf9762..6c765fc2661 100644
> --- a/src/intel/compiler/brw_fs_nir.cpp
> +++ b/src/intel/compiler/brw_fs_nir.cpp
> @@ -801,11 +801,6 @@ fs_visitor::nir_emit_alu(const fs_builder ,
> nir_alu_instr *instr)
>inst->saturate = instr->dest.saturate;
>break;
>
> -   case nir_op_b2i:
> -   case nir_op_b2f:
> -  op[0].type = BRW_REGISTER_TYPE_D;
> -  op[0].negate = !op[0].negate;
> -  /* fallthrough */
> case nir_op_f2f64:
> case nir_op_f2i64:
> case nir_op_f2u64:
> @@ -850,6 +845,11 @@ fs_visitor::nir_emit_alu(const fs_builder ,
> nir_alu_instr *instr)
>inst->saturate = instr->dest.saturate;
>break;
>
> +   case nir_op_b2i:
> +   case nir_op_b2f:
> +  op[0].type = BRW_REGISTER_TYPE_D;
> +  op[0].negate = !op[0].negate;
> +  /* fallthrough */
> case nir_op_i2f64:
> case nir_op_i2i64:
> case nir_op_u2f64:
> --
> 2.17.1
>
> ___
> mesa-dev mailing list
> mesa-dev@lists.freedesktop.org
> https://lists.freedesktop.org/mailman/listinfo/mesa-dev
>
___
mesa-dev mailing list
mesa-dev@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/mesa-dev

Re: [Mesa-dev] [PATCH 12/59] intel/compiler: handle b2i/b2f with other integer conversion opcodes

2018-12-05 Thread Pohjolainen, Topi 
On Wed, Dec 05, 2018 at 11:23:06AM +0100, Iago Toral wrote:
> On Tue, 2018-12-04 at 18:16 +0200, Pohjolainen, Topi wrote:
> > On Tue, Dec 04, 2018 at 08:16:36AM +0100, Iago Toral Quiroga wrote:
> > > Since we handle booleans as integers this makes more sense.
> > 
> > If this is applied before patch 10, can we merge 10 and 13?
> 
> We can't apply this before patch 10 because patch 10 is the one that
> splits the f264 and {i,u}264 opcodes. However, we could merge this and
> patch 13 into patch 10 if that looks better to you.

What you have is just fine. I just didn't see all the corners involved.
Patches 11-13:

Reviewed-by: Topi Pohjolainen 

> 
> Iago
> 
> > > ---
> > >  src/intel/compiler/brw_fs_nir.cpp | 10 +-
> > >  1 file changed, 5 insertions(+), 5 deletions(-)
> > > 
> > > diff --git a/src/intel/compiler/brw_fs_nir.cpp
> > > b/src/intel/compiler/brw_fs_nir.cpp
> > > index 9f3d3bf9762..6c765fc2661 100644
> > > --- a/src/intel/compiler/brw_fs_nir.cpp
> > > +++ b/src/intel/compiler/brw_fs_nir.cpp
> > > @@ -801,11 +801,6 @@ fs_visitor::nir_emit_alu(const fs_builder
> > > , nir_alu_instr *instr)
> > >inst->saturate = instr->dest.saturate;
> > >break;
> > >  
> > > -   case nir_op_b2i:
> > > -   case nir_op_b2f:
> > > -  op[0].type = BRW_REGISTER_TYPE_D;
> > > -  op[0].negate = !op[0].negate;
> > > -  /* fallthrough */
> > > case nir_op_f2f64:
> > > case nir_op_f2i64:
> > > case nir_op_f2u64:
> > > @@ -850,6 +845,11 @@ fs_visitor::nir_emit_alu(const fs_builder
> > > , nir_alu_instr *instr)
> > >inst->saturate = instr->dest.saturate;
> > >break;
> > >  
> > > +   case nir_op_b2i:
> > > +   case nir_op_b2f:
> > > +  op[0].type = BRW_REGISTER_TYPE_D;
> > > +  op[0].negate = !op[0].negate;
> > > +  /* fallthrough */
> > > case nir_op_i2f64:
> > > case nir_op_i2i64:
> > > case nir_op_u2f64:
> > > -- 
> > > 2.17.1
> > > 
> > > ___
> > > mesa-dev mailing list
> > > mesa-dev@lists.freedesktop.org
> > > https://lists.freedesktop.org/mailman/listinfo/mesa-dev
> > 
> > 
> 
___
mesa-dev mailing list
mesa-dev@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/mesa-dev

Re: [Mesa-dev] [PATCH 12/59] intel/compiler: handle b2i/b2f with other integer conversion opcodes

2018-12-05 Thread Iago Toral 
On Tue, 2018-12-04 at 18:16 +0200, Pohjolainen, Topi wrote:
> On Tue, Dec 04, 2018 at 08:16:36AM +0100, Iago Toral Quiroga wrote:
> > Since we handle booleans as integers this makes more sense.
> 
> If this is applied before patch 10, can we merge 10 and 13?

We can't apply this before patch 10 because patch 10 is the one that
splits the f264 and {i,u}264 opcodes. However, we could merge this and
patch 13 into patch 10 if that looks better to you.

Iago

> > ---
> >  src/intel/compiler/brw_fs_nir.cpp | 10 +-
> >  1 file changed, 5 insertions(+), 5 deletions(-)
> > 
> > diff --git a/src/intel/compiler/brw_fs_nir.cpp
> > b/src/intel/compiler/brw_fs_nir.cpp
> > index 9f3d3bf9762..6c765fc2661 100644
> > --- a/src/intel/compiler/brw_fs_nir.cpp
> > +++ b/src/intel/compiler/brw_fs_nir.cpp
> > @@ -801,11 +801,6 @@ fs_visitor::nir_emit_alu(const fs_builder
> > , nir_alu_instr *instr)
> >inst->saturate = instr->dest.saturate;
> >break;
> >  
> > -   case nir_op_b2i:
> > -   case nir_op_b2f:
> > -  op[0].type = BRW_REGISTER_TYPE_D;
> > -  op[0].negate = !op[0].negate;
> > -  /* fallthrough */
> > case nir_op_f2f64:
> > case nir_op_f2i64:
> > case nir_op_f2u64:
> > @@ -850,6 +845,11 @@ fs_visitor::nir_emit_alu(const fs_builder
> > , nir_alu_instr *instr)
> >inst->saturate = instr->dest.saturate;
> >break;
> >  
> > +   case nir_op_b2i:
> > +   case nir_op_b2f:
> > +  op[0].type = BRW_REGISTER_TYPE_D;
> > +  op[0].negate = !op[0].negate;
> > +  /* fallthrough */
> > case nir_op_i2f64:
> > case nir_op_i2i64:
> > case nir_op_u2f64:
> > -- 
> > 2.17.1
> > 
> > ___
> > mesa-dev mailing list
> > mesa-dev@lists.freedesktop.org
> > https://lists.freedesktop.org/mailman/listinfo/mesa-dev
> 
> 

___
mesa-dev mailing list
mesa-dev@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/mesa-dev

Re: [Mesa-dev] [PATCH 12/59] intel/compiler: handle b2i/b2f with other integer conversion opcodes

2018-12-04 Thread Pohjolainen, Topi 
On Tue, Dec 04, 2018 at 08:16:36AM +0100, Iago Toral Quiroga wrote:
> Since we handle booleans as integers this makes more sense.

If this is applied before patch 10, can we merge 10 and 13?

> ---
>  src/intel/compiler/brw_fs_nir.cpp | 10 +-
>  1 file changed, 5 insertions(+), 5 deletions(-)
> 
> diff --git a/src/intel/compiler/brw_fs_nir.cpp 
> b/src/intel/compiler/brw_fs_nir.cpp
> index 9f3d3bf9762..6c765fc2661 100644
> --- a/src/intel/compiler/brw_fs_nir.cpp
> +++ b/src/intel/compiler/brw_fs_nir.cpp
> @@ -801,11 +801,6 @@ fs_visitor::nir_emit_alu(const fs_builder , 
> nir_alu_instr *instr)
>inst->saturate = instr->dest.saturate;
>break;
>  
> -   case nir_op_b2i:
> -   case nir_op_b2f:
> -  op[0].type = BRW_REGISTER_TYPE_D;
> -  op[0].negate = !op[0].negate;
> -  /* fallthrough */
> case nir_op_f2f64:
> case nir_op_f2i64:
> case nir_op_f2u64:
> @@ -850,6 +845,11 @@ fs_visitor::nir_emit_alu(const fs_builder , 
> nir_alu_instr *instr)
>inst->saturate = instr->dest.saturate;
>break;
>  
> +   case nir_op_b2i:
> +   case nir_op_b2f:
> +  op[0].type = BRW_REGISTER_TYPE_D;
> +  op[0].negate = !op[0].negate;
> +  /* fallthrough */
> case nir_op_i2f64:
> case nir_op_i2i64:
> case nir_op_u2f64:
> -- 
> 2.17.1
> 
> ___
> mesa-dev mailing list
> mesa-dev@lists.freedesktop.org
> https://lists.freedesktop.org/mailman/listinfo/mesa-dev
___
mesa-dev mailing list
mesa-dev@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/mesa-dev

[Mesa-dev] [PATCH 12/59] intel/compiler: handle b2i/b2f with other integer conversion opcodes

2018-12-03 Thread Iago Toral Quiroga 
Since we handle booleans as integers this makes more sense.
---
 src/intel/compiler/brw_fs_nir.cpp | 10 +-
 1 file changed, 5 insertions(+), 5 deletions(-)

diff --git a/src/intel/compiler/brw_fs_nir.cpp 
b/src/intel/compiler/brw_fs_nir.cpp
index 9f3d3bf9762..6c765fc2661 100644
--- a/src/intel/compiler/brw_fs_nir.cpp
+++ b/src/intel/compiler/brw_fs_nir.cpp
@@ -801,11 +801,6 @@ fs_visitor::nir_emit_alu(const fs_builder , 
nir_alu_instr *instr)
   inst->saturate = instr->dest.saturate;
   break;
 
-   case nir_op_b2i:
-   case nir_op_b2f:
-  op[0].type = BRW_REGISTER_TYPE_D;
-  op[0].negate = !op[0].negate;
-  /* fallthrough */
case nir_op_f2f64:
case nir_op_f2i64:
case nir_op_f2u64:
@@ -850,6 +845,11 @@ fs_visitor::nir_emit_alu(const fs_builder , 
nir_alu_instr *instr)
   inst->saturate = instr->dest.saturate;
   break;
 
+   case nir_op_b2i:
+   case nir_op_b2f:
+  op[0].type = BRW_REGISTER_TYPE_D;
+  op[0].negate = !op[0].negate;
+  /* fallthrough */
case nir_op_i2f64:
case nir_op_i2i64:
case nir_op_u2f64:
-- 
2.17.1

___
mesa-dev mailing list
mesa-dev@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/mesa-dev