Re: [Mesa-dev] [PATCH excerpt] mesa: Wrap unlikely() around conditions leading to _mesa_error().

2015-03-04 Thread Kenneth Graunke
On Monday, March 02, 2015 03:03:00 PM Eric Anholt wrote: > Kenneth Graunke writes: > > > Generated by the following Coccinelle semantic patch: > > > > @@ > > expression E; > > @@ > > if ( > > -E > > +unlikely(E) > > ) > > { ... _mesa_error(...) ... } > > > > Improves performance in OglBatch7 (a

Re: [Mesa-dev] [PATCH excerpt] mesa: Wrap unlikely() around conditions leading to _mesa_error().

2015-03-03 Thread Brian Paul
On 03/02/2015 06:56 PM, Matt Turner wrote: On Mon, Mar 2, 2015 at 4:37 PM, Brian Paul wrote: On 03/02/2015 04:03 PM, Eric Anholt wrote: Kenneth Graunke writes: Generated by the following Coccinelle semantic patch: @@ expression E; @@ if ( -E +unlikely(E) ) { ... _mesa_error(...) ... }

Re: [Mesa-dev] [PATCH excerpt] mesa: Wrap unlikely() around conditions leading to _mesa_error().

2015-03-02 Thread Matt Turner
On Mon, Mar 2, 2015 at 4:37 PM, Brian Paul wrote: > On 03/02/2015 04:03 PM, Eric Anholt wrote: >> >> Kenneth Graunke writes: >> >>> Generated by the following Coccinelle semantic patch: >>> >>> @@ >>> expression E; >>> @@ >>> if ( >>> -E >>> +unlikely(E) >>> ) >>> { ... _mesa_error(...) ... } >

Re: [Mesa-dev] [PATCH excerpt] mesa: Wrap unlikely() around conditions leading to _mesa_error().

2015-03-02 Thread Eric Anholt
Brian Paul writes: > On 03/02/2015 04:03 PM, Eric Anholt wrote: >> Kenneth Graunke writes: >> >>> Generated by the following Coccinelle semantic patch: >>> >>> @@ >>> expression E; >>> @@ >>> if ( >>> -E >>> +unlikely(E) >>> ) >>> { ... _mesa_error(...) ... } >>> >>> Improves performance in Og

Re: [Mesa-dev] [PATCH excerpt] mesa: Wrap unlikely() around conditions leading to _mesa_error().

2015-03-02 Thread Matt Turner
On Mon, Mar 2, 2015 at 4:45 PM, Dave Airlie wrote: > why can't we just do _mesa_error -> __mesa_error, #define _mesa_error > unlikely(__mesa_error) or some such. I don't think wrapping the function call in unlikely() does anything. ___ mesa-dev mailing

Re: [Mesa-dev] [PATCH excerpt] mesa: Wrap unlikely() around conditions leading to _mesa_error().

2015-03-02 Thread Dave Airlie
On 3 March 2015 at 10:37, Brian Paul wrote: > On 03/02/2015 04:03 PM, Eric Anholt wrote: >> >> Kenneth Graunke writes: >> >>> Generated by the following Coccinelle semantic patch: >>> >>> @@ >>> expression E; >>> @@ >>> if ( >>> -E >>> +unlikely(E) >>> ) >>> { ... _mesa_error(...) ... } >>> >>>

Re: [Mesa-dev] [PATCH excerpt] mesa: Wrap unlikely() around conditions leading to _mesa_error().

2015-03-02 Thread Brian Paul
On 03/02/2015 04:03 PM, Eric Anholt wrote: Kenneth Graunke writes: Generated by the following Coccinelle semantic patch: @@ expression E; @@ if ( -E +unlikely(E) ) { ... _mesa_error(...) ... } Improves performance in OglBatch7 (a CPU-bound microbenchmark) on an i7 4750HQ (Haswell) by 1.085

Re: [Mesa-dev] [PATCH excerpt] mesa: Wrap unlikely() around conditions leading to _mesa_error().

2015-03-02 Thread Brian Paul
On 03/02/2015 03:48 PM, Kenneth Graunke wrote: Generated by the following Coccinelle semantic patch: @@ expression E; @@ if ( -E +unlikely(E) ) { ... _mesa_error(...) ... } Improves performance in OglBatch7 (a CPU-bound microbenchmark) on an i7 4750HQ (Haswell) by 1.08546% +/- 0.519734% (n=29

Re: [Mesa-dev] [PATCH excerpt] mesa: Wrap unlikely() around conditions leading to _mesa_error().

2015-03-02 Thread Eric Anholt
Kenneth Graunke writes: > Generated by the following Coccinelle semantic patch: > > @@ > expression E; > @@ > if ( > -E > +unlikely(E) > ) > { ... _mesa_error(...) ... } > > Improves performance in OglBatch7 (a CPU-bound microbenchmark) on > an i7 4750HQ (Haswell) by 1.08546% +/- 0.519734% (n=29

[Mesa-dev] [PATCH excerpt] mesa: Wrap unlikely() around conditions leading to _mesa_error().

2015-03-02 Thread Kenneth Graunke
Generated by the following Coccinelle semantic patch: @@ expression E; @@ if ( -E +unlikely(E) ) { ... _mesa_error(...) ... } Improves performance in OglBatch7 (a CPU-bound microbenchmark) on an i7 4750HQ (Haswell) by 1.08546% +/- 0.519734% (n=292). Signed-off-by: Kenneth Graunke --- Here is