Re: [Mesa-dev] [PATCH] i965: Avoid dependency hints on math opcodes

2014-04-15 Thread Matt Turner
On Wed, Feb 12, 2014 at 4:24 PM, wrote: > From: Mike Stroyan > > Putting NoDDClr and NoDDChk dependency control on instruction > sequences that include math opcodes can cause corruption of channels. > Treat math opcodes like send opcodes and suppress dependency hinting. > > Signed-off-by: Mike

Re: [Mesa-dev] [PATCH] i965: Avoid dependency hints on math opcodes

2014-03-21 Thread Mike Stroyan
On Wed, Mar 19, 2014 at 3:16 PM, Matt Turner wrote: > - Does the broken behavior depend on the hardware generation? I.e., > broken on Ivy Bridge and Haswell but not Ironlake? > - There are ~12 math ops. Are the dependency control hints broken for > all of them, or just exp2? The problem with NoD

Re: [Mesa-dev] [PATCH] i965: Avoid dependency hints on math opcodes

2014-03-19 Thread Matt Turner
On Wed, Mar 19, 2014 at 7:27 AM, Mike Stroyan wrote: > Ken, > > The defect depends on details of the shader compiler reusing registers. > Old register values are sometimes left in some channels when the > write-after-write hints are used. > Further reducing the GLSL test case would be tricky bec

Re: [Mesa-dev] [PATCH] i965: Avoid dependency hints on math opcodes

2014-03-19 Thread Mike Stroyan
Ken, The defect depends on details of the shader compiler reusing registers. Old register values are sometimes left in some channels when the write-after-write hints are used. Further reducing the GLSL test case would be tricky because the reuse of g8 would be likely to be changed. I did already

Re: [Mesa-dev] [PATCH] i965: Avoid dependency hints on math opcodes

2014-03-18 Thread Kenneth Graunke
On 02/28/2014 02:35 PM, Mike Stroyan wrote: > Matt, > > You haven't replied to my mail with an updated shader test that shows > the math instructions alone causing trouble. I don't think Matt has time to do that. Could you please trim down your shader test to a smaller case which demonstrates

Re: [Mesa-dev] [PATCH] i965: Avoid dependency hints on math opcodes

2014-02-28 Thread Mike Stroyan
Matt, You haven't replied to my mail with an updated shader test that shows the math instructions alone causing trouble. Do you now agree that my patch avoiding math instructions in opt_set_dependency_control is the appropriate fix? On Fri, Feb 21, 2014 at 11:30 AM, Mike Stroyan wrote: > Mat

Re: [Mesa-dev] [PATCH] i965: Avoid dependency hints on math opcodes

2014-02-23 Thread Jordan Justen
On Wed, Feb 12, 2014 at 9:25 PM, Ian Romanick wrote: > On 02/12/2014 06:59 PM, Matt Turner wrote: >> On Wed, Feb 12, 2014 at 4:24 PM, wrote: >>> From: Mike Stroyan >>> >>> Putting NoDDClr and NoDDChk dependency control on instruction >>> sequences that include math opcodes can cause corruptio

Re: [Mesa-dev] [PATCH] i965: Avoid dependency hints on math opcodes

2014-02-20 Thread Matt Turner
On Thu, Feb 20, 2014 at 4:55 PM, Mike Stroyan wrote: > Ian, > > Here is a shader_test version. It lacks the sparkling uncertainty of the > pixel values in the previous animated example program. > It just gets all vertices uniformly wrong when I run it. Thanks for the test Mike. I reproduced it

Re: [Mesa-dev] [PATCH] i965: Avoid dependency hints on math opcodes

2014-02-20 Thread Mike Stroyan
Ian, Here is a shader_test version. It lacks the sparkling uncertainty of the pixel values in the previous animated example program. It just gets all vertices uniformly wrong when I run it. On Thu, Feb 20, 2014 at 4:31 PM, Ian Romanick wrote: > On 02/12/2014 04:24 PM, m...@lunarg.com wrote:

Re: [Mesa-dev] [PATCH] i965: Avoid dependency hints on math opcodes

2014-02-20 Thread Ian Romanick
On 02/12/2014 04:24 PM, m...@lunarg.com wrote: > From: Mike Stroyan > > Putting NoDDClr and NoDDChk dependency control on instruction > sequences that include math opcodes can cause corruption of channels. > Treat math opcodes like send opcodes and suppress dependency hinting. Since you've ana

Re: [Mesa-dev] [PATCH] i965: Avoid dependency hints on math opcodes

2014-02-18 Thread Eric Anholt
m...@lunarg.com writes: > From: Mike Stroyan > > Putting NoDDClr and NoDDChk dependency control on instruction > sequences that include math opcodes can cause corruption of channels. > Treat math opcodes like send opcodes and suppress dependency hinting. > > Signed-off-by: Mike Stroyan > Teste

Re: [Mesa-dev] [PATCH] i965: Avoid dependency hints on math opcodes

2014-02-12 Thread Ian Romanick
On 02/12/2014 06:59 PM, Matt Turner wrote: > On Wed, Feb 12, 2014 at 4:24 PM, wrote: >> From: Mike Stroyan >> >> Putting NoDDClr and NoDDChk dependency control on instruction >> sequences that include math opcodes can cause corruption of channels. >> Treat math opcodes like send opcodes and su

Re: [Mesa-dev] [PATCH] i965: Avoid dependency hints on math opcodes

2014-02-12 Thread Matt Turner
On Wed, Feb 12, 2014 at 4:24 PM, wrote: > From: Mike Stroyan > > Putting NoDDClr and NoDDChk dependency control on instruction > sequences that include math opcodes can cause corruption of channels. > Treat math opcodes like send opcodes and suppress dependency hinting. > > Signed-off-by: Mike

[Mesa-dev] [PATCH] i965: Avoid dependency hints on math opcodes

2014-02-12 Thread mike
From: Mike Stroyan Putting NoDDClr and NoDDChk dependency control on instruction sequences that include math opcodes can cause corruption of channels. Treat math opcodes like send opcodes and suppress dependency hinting. Signed-off-by: Mike Stroyan Tested-by: Tony Bertapelli --- src/mesa/dr