Re: [PATCH 1/x] arm: Add vld1_lane_bf16 + vldq_lane_bf16 intrinsics

[Andrea Corallo via Gcc-patches]

Christophe Lyon  writes:

> On Tue, 3 Nov 2020 at 11:27, Kyrylo Tkachov via Gcc-patches
>  wrote:
>>
>> Hi Andrea,
>>
>> > -Original Message-
>> > From: Andrea Corallo 
>> > Sent: 26 October 2020 15:59
>> > To: gcc-patches@gcc.gnu.org
>> > Cc: Kyrylo Tkachov ; Richard Earnshaw
>> > ; nd 
>> > Subject: [PATCH 1/x] arm: Add vld1_lane_bf16 + vldq_lane_bf16 intrinsics
>> >
>> > Hi all,
>> >
>> > I'd like to submit the following patch implementing the bfloat16_t
>> > neon related load intrinsics: vld1_lane_bf16, vld1q_lane_bf16.
>> >
>> > Please see refer to:
>> > ACLE 
>> > ISA  
>> >
>> > Regtested and bootstrapped.
>> >
>> > Okay for trunk?
>>
>
> I think you need to add -mfloat-abi=hard to the dg-additional-options
> otherwise vld1_lane_bf16_1.c
> fails on targets with a soft float-abi default (eg arm-linux-gnueabi).
>
> See bf16_vldn_1.c.
>
> BTW, why did you use a different naming scheme for the tests?
> (bf16_vldn_1.c vs vld1_lane_bf16_1.c)

Nothing special, it made more sense to me to use directly the name of
the intrinsic as it include already the bf16 information.  I believe we
have both schemas in the aarch64 & arm backends.  I've no problem with
renaming the tests if we feel is important.

  Andrea

Re: [PATCH 1/x] arm: Add vld1_lane_bf16 + vldq_lane_bf16 intrinsics

[Andrea Corallo via Gcc-patches]

Christophe Lyon  writes:

> On Wed, 4 Nov 2020 at 14:29, Christophe Lyon  
> wrote:
>>
>> On Tue, 3 Nov 2020 at 11:27, Kyrylo Tkachov via Gcc-patches
>>  wrote:
>> >
>> > Hi Andrea,
>> >
>> > > -Original Message-
>> > > From: Andrea Corallo 
>> > > Sent: 26 October 2020 15:59
>> > > To: gcc-patches@gcc.gnu.org
>> > > Cc: Kyrylo Tkachov ; Richard Earnshaw
>> > > ; nd 
>> > > Subject: [PATCH 1/x] arm: Add vld1_lane_bf16 + vldq_lane_bf16 intrinsics
>> > >
>> > > Hi all,
>> > >
>> > > I'd like to submit the following patch implementing the bfloat16_t
>> > > neon related load intrinsics: vld1_lane_bf16, vld1q_lane_bf16.
>> > >
>> > > Please see refer to:
>> > > ACLE 
>> > > ISA  
>> > >
>> > > Regtested and bootstrapped.
>> > >
>> > > Okay for trunk?
>> >
>>
>> I think you need to add -mfloat-abi=hard to the dg-additional-options
>> otherwise vld1_lane_bf16_1.c
>> fails on targets with a soft float-abi default (eg arm-linux-gnueabi).
>>
>> See bf16_vldn_1.c.
>
> Actually that's not sufficient because in turn we get:
> /sysroot-arm-none-linux-gnueabi/usr/include/gnu/stubs.h:10:11: fatal
> error: gnu/stubs-hard.h: No such file or directory
>
> So you should check that -mfloat-abi=hard is supported.
>
> Ditto for the vst tests.
>

Hi Christophe,

thanks for catching this, I'll prepare a patch.

  Andrea

Re: [PATCH 1/x] arm: Add vld1_lane_bf16 + vldq_lane_bf16 intrinsics

[Christophe Lyon via Gcc-patches]

On Wed, 4 Nov 2020 at 14:29, Christophe Lyon  wrote:
>
> On Tue, 3 Nov 2020 at 11:27, Kyrylo Tkachov via Gcc-patches
>  wrote:
> >
> > Hi Andrea,
> >
> > > -Original Message-
> > > From: Andrea Corallo 
> > > Sent: 26 October 2020 15:59
> > > To: gcc-patches@gcc.gnu.org
> > > Cc: Kyrylo Tkachov ; Richard Earnshaw
> > > ; nd 
> > > Subject: [PATCH 1/x] arm: Add vld1_lane_bf16 + vldq_lane_bf16 intrinsics
> > >
> > > Hi all,
> > >
> > > I'd like to submit the following patch implementing the bfloat16_t
> > > neon related load intrinsics: vld1_lane_bf16, vld1q_lane_bf16.
> > >
> > > Please see refer to:
> > > ACLE 
> > > ISA  
> > >
> > > Regtested and bootstrapped.
> > >
> > > Okay for trunk?
> >
>
> I think you need to add -mfloat-abi=hard to the dg-additional-options
> otherwise vld1_lane_bf16_1.c
> fails on targets with a soft float-abi default (eg arm-linux-gnueabi).
>
> See bf16_vldn_1.c.

Actually that's not sufficient because in turn we get:
/sysroot-arm-none-linux-gnueabi/usr/include/gnu/stubs.h:10:11: fatal
error: gnu/stubs-hard.h: No such file or directory

So you should check that -mfloat-abi=hard is supported.

Ditto for the vst tests.

>
> BTW, why did you use a different naming scheme for the tests?
> (bf16_vldn_1.c vs vld1_lane_bf16_1.c)
>
> Christophe
>
> > Ok.
> > Thanks,
> > Kyrill
> >
> >
> > >
> > >   Andrea
> >

Re: [PATCH 1/x] arm: Add vld1_lane_bf16 + vldq_lane_bf16 intrinsics

[Christophe Lyon via Gcc-patches]

On Tue, 3 Nov 2020 at 11:27, Kyrylo Tkachov via Gcc-patches
 wrote:
>
> Hi Andrea,
>
> > -Original Message-
> > From: Andrea Corallo 
> > Sent: 26 October 2020 15:59
> > To: gcc-patches@gcc.gnu.org
> > Cc: Kyrylo Tkachov ; Richard Earnshaw
> > ; nd 
> > Subject: [PATCH 1/x] arm: Add vld1_lane_bf16 + vldq_lane_bf16 intrinsics
> >
> > Hi all,
> >
> > I'd like to submit the following patch implementing the bfloat16_t
> > neon related load intrinsics: vld1_lane_bf16, vld1q_lane_bf16.
> >
> > Please see refer to:
> > ACLE 
> > ISA  
> >
> > Regtested and bootstrapped.
> >
> > Okay for trunk?
>

I think you need to add -mfloat-abi=hard to the dg-additional-options
otherwise vld1_lane_bf16_1.c
fails on targets with a soft float-abi default (eg arm-linux-gnueabi).

See bf16_vldn_1.c.

BTW, why did you use a different naming scheme for the tests?
(bf16_vldn_1.c vs vld1_lane_bf16_1.c)

Christophe

> Ok.
> Thanks,
> Kyrill
>
>
> >
> >   Andrea
>

RE: [PATCH 1/x] arm: Add vld1_lane_bf16 + vldq_lane_bf16 intrinsics

[Kyrylo Tkachov via Gcc-patches]

Hi Andrea,

> -Original Message-
> From: Andrea Corallo 
> Sent: 26 October 2020 15:59
> To: gcc-patches@gcc.gnu.org
> Cc: Kyrylo Tkachov ; Richard Earnshaw
> ; nd 
> Subject: [PATCH 1/x] arm: Add vld1_lane_bf16 + vldq_lane_bf16 intrinsics
> 
> Hi all,
> 
> I'd like to submit the following patch implementing the bfloat16_t
> neon related load intrinsics: vld1_lane_bf16, vld1q_lane_bf16.
> 
> Please see refer to:
> ACLE 
> ISA  
> 
> Regtested and bootstrapped.
> 
> Okay for trunk?

Ok.
Thanks,
Kyrill


> 
>   Andrea

Re: [PATCH 1/x] arm: Add vld1_lane_bf16 + vldq_lane_bf16 intrinsics

[Andrea Corallo via Gcc-patches]

Andrea Corallo via Gcc-patches  writes:

> Hi all,
>
> I'd like to submit the following patch implementing the bfloat16_t
> neon related load intrinsics: vld1_lane_bf16, vld1q_lane_bf16.
>
> Please see refer to:
> ACLE 
> ISA  
>
> Regtested and bootstrapped.
>
> Okay for trunk?
>
>   Andrea

Ping