Re: [PATCH 9/17][ARM] Add NEON FP16 arithmetic instructions.

2016-07-28 Thread Ramana Radhakrishnan
On Mon, Jul 4, 2016 at 3:09 PM, Matthew Wahab wrote: > On 18/05/16 01:58, Joseph Myers wrote: >> On Tue, 17 May 2016, Matthew Wahab wrote: >> >>> As with the VFP FP16 arithmetic instructions, operations on __fp16 >>> values are done by conversion to single-precision. Any new optimization >>> suppo

Re: [PATCH 9/17][ARM] Add NEON FP16 arithmetic instructions.

2016-07-04 Thread Matthew Wahab
On 18/05/16 01:58, Joseph Myers wrote: > On Tue, 17 May 2016, Matthew Wahab wrote: > >> As with the VFP FP16 arithmetic instructions, operations on __fp16 >> values are done by conversion to single-precision. Any new optimization >> supported by the instruction descriptions can only apply to code

Re: [PATCH 9/17][ARM] Add NEON FP16 arithmetic instructions.

2016-06-08 Thread Joseph Myers
On Wed, 8 Jun 2016, James Greenhalgh wrote: > My question is whether you consider the different behaviour between scalar > float16_t and vector-of-float16_t types to be a bug? I can think of some No, because it matches how things work for vectors of integer types. E.g.: typedef unsigned char v

Re: [PATCH 9/17][ARM] Add NEON FP16 arithmetic instructions.

2016-06-08 Thread James Greenhalgh
On Thu, May 19, 2016 at 05:29:16PM +, Joseph Myers wrote: > On Thu, 19 May 2016, Jiong Wang wrote: > > > Then, > > > > * if we add scalar HF mode to standard patterns, vector HF modes operation > > will be > > turned into scalar HF operations instead of scalar SF operations. > > > >

Re: [PATCH 9/17][ARM] Add NEON FP16 arithmetic instructions.

2016-05-19 Thread Joseph Myers
On Thu, 19 May 2016, Jiong Wang wrote: > Then, > > * if we add scalar HF mode to standard patterns, vector HF modes operation > will be > turned into scalar HF operations instead of scalar SF operations. > > * if we add vector HF mode to standard patterns, vector HF modes operations > wi

Re: [PATCH 9/17][ARM] Add NEON FP16 arithmetic instructions.

2016-05-19 Thread Jiong Wang
On 18/05/16 01:58, Joseph Myers wrote: On Tue, 17 May 2016, Matthew Wahab wrote: As with the VFP FP16 arithmetic instructions, operations on __fp16 values are done by conversion to single-precision. Any new optimization supported by the instruction descriptions can only apply to code generate

Re: [PATCH 9/17][ARM] Add NEON FP16 arithmetic instructions.

2016-05-17 Thread Joseph Myers
On Tue, 17 May 2016, Matthew Wahab wrote: > As with the VFP FP16 arithmetic instructions, operations on __fp16 > values are done by conversion to single-precision. Any new optimization > supported by the instruction descriptions can only apply to code > generated using intrinsics added in this pat