Re: [Mesa-dev] [PATCH 20/20] nir: Narrow unnecessary 64-bit operations to 32-bits

2017-07-19 Thread Connor Abbott
On Wed, Jul 19, 2017 at 1:37 PM, Matt Turner wrote: > On Mon, Jul 17, 2017 at 8:14 PM, Connor Abbott wrote: >> >> This function seems to be doing two subtly different things: >> >> 1. If the instruction defining the source is nir_op_pack_64_2x32_split >>

Re: [Mesa-dev] [PATCH 20/20] nir: Narrow unnecessary 64-bit operations to 32-bits

2017-07-19 Thread Matt Turner
On Mon, Jul 17, 2017 at 8:14 PM, Connor Abbott wrote: > On Thu, Jul 6, 2017 at 4:48 PM, Matt Turner wrote: >> If we know the high bits are zero, we can just do a 32-bit comparison on >> the low bytes instead. >> --- >>

Re: [Mesa-dev] [PATCH 20/20] nir: Narrow unnecessary 64-bit operations to 32-bits

2017-07-17 Thread Connor Abbott
On Thu, Jul 6, 2017 at 4:48 PM, Matt Turner wrote: > If we know the high bits are zero, we can just do a 32-bit comparison on > the low bytes instead. > --- > src/compiler/nir/nir_opt_algebraic.py | 14 +- > src/compiler/nir/nir_search_helpers.h | 48 >

Re: [Mesa-dev] [PATCH 20/20] nir: Narrow unnecessary 64-bit operations to 32-bits

2017-07-17 Thread Connor Abbott
On Thu, Jul 6, 2017 at 4:48 PM, Matt Turner wrote: > If we know the high bits are zero, we can just do a 32-bit comparison on > the low bytes instead. > --- > src/compiler/nir/nir_opt_algebraic.py | 14 +- > src/compiler/nir/nir_search_helpers.h | 48 >

Re: [Mesa-dev] [PATCH 20/20] nir: Narrow unnecessary 64-bit operations to 32-bits

2017-07-17 Thread Matt Turner
On Mon, Jul 17, 2017 at 5:57 PM, Kenneth Graunke wrote: > On Thursday, July 6, 2017 4:48:30 PM PDT Matt Turner wrote: >> If we know the high bits are zero, we can just do a 32-bit comparison on >> the low bytes instead. >> --- >> src/compiler/nir/nir_opt_algebraic.py | 14

Re: [Mesa-dev] [PATCH 20/20] nir: Narrow unnecessary 64-bit operations to 32-bits

2017-07-17 Thread Kenneth Graunke
On Thursday, July 6, 2017 4:48:30 PM PDT Matt Turner wrote: > If we know the high bits are zero, we can just do a 32-bit comparison on > the low bytes instead. > --- > src/compiler/nir/nir_opt_algebraic.py | 14 +- > src/compiler/nir/nir_search_helpers.h | 48 >

[Mesa-dev] [PATCH 20/20] nir: Narrow unnecessary 64-bit operations to 32-bits

2017-07-06 Thread Matt Turner
If we know the high bits are zero, we can just do a 32-bit comparison on the low bytes instead. --- src/compiler/nir/nir_opt_algebraic.py | 14 +- src/compiler/nir/nir_search_helpers.h | 48 +++ 2 files changed, 61 insertions(+), 1 deletion(-) diff --git