On Tue, Mar 17, 2015 at 10:56 PM, Matt Turner wrote:
> Thanks. Looked through stats and at some of the regressions.
I disabled the pow(x, y) optimization in NIR, turned off the SEL
peephole in the fs backend, and added neg/abs to
block_check_for_allowed_instrs() (will test and send a patch) to ma
https://bugs.freedesktop.org/show_bug.cgi?id=81992
--- Comment #1 from comicfans44 ---
Qt OpenGL use dynamic allocated memory to test GL extensions.
this bug makes every Qt OpenGL app memory corrupted.
I'm not sure a simple strdup is the correct way
(maybe leak instead of memory corrupt).
with
Reviewed-by: Tapani Pälli
On 03/11/2015 10:44 PM, Ian Romanick wrote:
From: Ian Romanick
On platforms that do not natively generate 0u and ~0u for Boolean
results, b2f expressions that look like
f = b2f(expr cmp 0)
will generate better code by pretending the expression is
f = ir_t
On Thu, Mar 19, 2015 at 12:35 AM, Jason Ekstrand wrote:
> On Wed, Mar 18, 2015 at 9:31 PM, Matt Turner wrote:
>> On Wed, Mar 18, 2015 at 9:27 PM, Jason Ekstrand wrote:
>>> On Wed, Mar 18, 2015 at 8:59 PM, Connor Abbott wrote:
On Wed, Mar 18, 2015 at 11:35 PM, Jason Ekstrand
wrote:
>
Shader-db results on HSW:
total instructions in shared programs: 4112172 -> 4111347 (-0.02%)
instructions in affected programs: 18265 -> 17440 (-4.52%)
helped:137
HURT: 0
GAINED:0
LOST:
On Tue, Mar 10, 2015 at 4:18 AM, Kenneth Graunke wrote:
> This especially helps with NIR because we currently emit MOVs at the top
> of the shader to copy from various ATTR registers to a giant VGRF array
> of all inputs. (This could potentially be done better, but since
> there's only ever one w
On Wed, Mar 18, 2015 at 9:31 PM, Matt Turner wrote:
> On Wed, Mar 18, 2015 at 9:27 PM, Jason Ekstrand wrote:
>> On Wed, Mar 18, 2015 at 8:59 PM, Connor Abbott wrote:
>>> On Wed, Mar 18, 2015 at 11:35 PM, Jason Ekstrand
>>> wrote:
On Mar 18, 2015 8:32 PM, "Matt Turner" wrote:
>
>
On Wed, Mar 18, 2015 at 8:43 PM, Kenneth Graunke wrote:
> On Wednesday, March 18, 2015 11:27:19 PM Connor Abbott wrote:
>> On Wed, Mar 18, 2015 at 11:24 PM, Kenneth Graunke
>> wrote:
>> > On Tuesday, March 17, 2015 07:17:19 PM Jason Ekstrand wrote:
>> >> ---
>> >> src/mesa/drivers/dri/i965/brw_
On Wed, Mar 18, 2015 at 9:27 PM, Jason Ekstrand wrote:
> On Wed, Mar 18, 2015 at 8:59 PM, Connor Abbott wrote:
>> On Wed, Mar 18, 2015 at 11:35 PM, Jason Ekstrand
>> wrote:
>>>
>>> On Mar 18, 2015 8:32 PM, "Matt Turner" wrote:
On Wed, Mar 18, 2015 at 7:39 PM, Jason Ekstrand
wro
On Wed, Mar 18, 2015 at 8:59 PM, Connor Abbott wrote:
> On Wed, Mar 18, 2015 at 11:35 PM, Jason Ekstrand wrote:
>>
>> On Mar 18, 2015 8:32 PM, "Matt Turner" wrote:
>>>
>>> On Wed, Mar 18, 2015 at 7:39 PM, Jason Ekstrand
>>> wrote:
>>> > On Wed, Mar 18, 2015 at 11:37 AM, Matt Turner
>>> > wrote
On Wed, Mar 18, 2015 at 11:35 PM, Jason Ekstrand wrote:
>
> On Mar 18, 2015 8:32 PM, "Matt Turner" wrote:
>>
>> On Wed, Mar 18, 2015 at 7:39 PM, Jason Ekstrand
>> wrote:
>> > On Wed, Mar 18, 2015 at 11:37 AM, Matt Turner
>> > wrote:
>> >> Transform this into b2f(and(a, b)).
>> >>
>> >> total in
On Wednesday, March 18, 2015 11:27:19 PM Connor Abbott wrote:
> On Wed, Mar 18, 2015 at 11:24 PM, Kenneth Graunke
> wrote:
> > On Tuesday, March 17, 2015 07:17:19 PM Jason Ekstrand wrote:
> >> ---
> >> src/mesa/drivers/dri/i965/brw_fs_nir.cpp | 8 +++-
> >> 1 file changed, 7 insertions(+), 1
On Wed, Mar 18, 2015 at 8:35 PM, Jason Ekstrand wrote:
> On Mar 18, 2015 8:32 PM, "Matt Turner" wrote:
>>
>> On Wed, Mar 18, 2015 at 7:39 PM, Jason Ekstrand
>> wrote:
>> > On Wed, Mar 18, 2015 at 11:37 AM, Matt Turner
>> > wrote:
>> >> Transform this into b2f(and(a, b)).
>> >>
>> >> total instr
On Mar 18, 2015 8:32 PM, "Matt Turner" wrote:
>
> On Wed, Mar 18, 2015 at 7:39 PM, Jason Ekstrand
wrote:
> > On Wed, Mar 18, 2015 at 11:37 AM, Matt Turner
wrote:
> >> Transform this into b2f(and(a, b)).
> >>
> >> total instructions in shared programs: 6205448 -> 6204391 (-0.02%)
> >> instruction
On Wed, Mar 18, 2015 at 7:39 PM, Jason Ekstrand wrote:
> On Wed, Mar 18, 2015 at 11:37 AM, Matt Turner wrote:
>> Transform this into b2f(and(a, b)).
>>
>> total instructions in shared programs: 6205448 -> 6204391 (-0.02%)
>> instructions in affected programs: 284030 -> 282973 (-0.37%)
>> help
On Tue, Mar 10, 2015 at 4:18 AM, Kenneth Graunke wrote:
> This especially helps with NIR because we currently emit MOVs at the top
> of the shader to copy from various ATTR registers to a giant VGRF array
> of all inputs. (This could potentially be done better, but since
> there's only ever one w
On Wed, Mar 18, 2015 at 11:24 PM, Kenneth Graunke wrote:
> On Tuesday, March 17, 2015 07:17:19 PM Jason Ekstrand wrote:
>> ---
>> src/mesa/drivers/dri/i965/brw_fs_nir.cpp | 8 +++-
>> 1 file changed, 7 insertions(+), 1 deletion(-)
>>
>> diff --git a/src/mesa/drivers/dri/i965/brw_fs_nir.cpp
>
On Tuesday, March 17, 2015 07:17:19 PM Jason Ekstrand wrote:
> ---
> src/mesa/drivers/dri/i965/brw_fs_nir.cpp | 8 +++-
> 1 file changed, 7 insertions(+), 1 deletion(-)
>
> diff --git a/src/mesa/drivers/dri/i965/brw_fs_nir.cpp
> b/src/mesa/drivers/dri/i965/brw_fs_nir.cpp
> index 5da8423..41f
On Wed, Mar 18, 2015 at 8:52 PM, Emil Velikov wrote:
> On 13 March 2015 at 04:56, Ilia Mirkin wrote:
>> We only program in one layer size per texture, so that means that all
>> levels must share one size. This makes the piglit test
>>
>> bin/texelFetch fs sampler2DArray
>>
>> have the same breaka
On Wed, Mar 18, 2015 at 11:37 AM, Matt Turner wrote:
> Transform this into b2f(or(a, b)).
>
> instructions in affected programs: 432 -> 430 (-0.46%)
> helped:2
> ---
> src/glsl/nir/nir_opt_algebraic.py | 1 +
> 1 file changed, 1 insertion(+)
>
> diff --git a/sr
On Wed, Mar 18, 2015 at 11:37 AM, Matt Turner wrote:
> Transform this into b2f(and(a, b)).
>
> total instructions in shared programs: 6205448 -> 6204391 (-0.02%)
> instructions in affected programs: 284030 -> 282973 (-0.37%)
> helped:903
> HURT:
On Wed, Mar 18, 2015 at 7:11 PM, Connor Abbott wrote:
> Given that the way we pack inputs/outputs/uniforms is probably only
> going to be useful for scalar backends, can we rename this to
> nir_assign_var_locations_scalar() and in the next patch
> nir_assign_var_locations_scalar_direct_first()? Th
Given that the way we pack inputs/outputs/uniforms is probably only
going to be useful for scalar backends, can we rename this to
nir_assign_var_locations_scalar() and in the next patch
nir_assign_var_locations_scalar_direct_first()? That makes more sense
to me than "packed," since on vec4 "packed"
On Wed, Mar 18, 2015 at 6:40 PM, Connor Abbott wrote:
> On Wed, Mar 18, 2015 at 9:10 PM, Jason Ekstrand wrote:
>> On Wed, Mar 18, 2015 at 6:07 PM, Connor Abbott wrote:
>>> Overall, I think it might be better to do this as a SSA pass right
>>> before lowering to source modifiers. I don't think we
Both seems to be excessively long, namely:
ClientAPIString can get up-to 47 based on current code, while the name
of the driver can dictate the length of the VersionString, currently it
is around 11. Let's pad each to 100, rather than the current 1000.
Signed-off-by: Emil Velikov
---
src/egl/ma
On Wed, Mar 18, 2015 at 9:10 PM, Jason Ekstrand wrote:
> On Wed, Mar 18, 2015 at 6:07 PM, Connor Abbott wrote:
>> Overall, I think it might be better to do this as a SSA pass right
>> before lowering to source modifiers. I don't think we would be running
>> any optimizations after it that depend
So in GLSL, add and mul can take a combination of a scalar and a
vector, and it'll splat the scalar automatically, and with this this
series that means that and/or can now take a combination of a scalar
and a vector as well, and we'll have to splat them as well. Are you
sure that the various backen
On Wed, Mar 18, 2015 at 6:07 PM, Connor Abbott wrote:
> Overall, I think it might be better to do this as a SSA pass right
> before lowering to source modifiers. I don't think we would be running
> any optimizations after it that depend on the output being all 0's or
> all 1's, but if you're conce
Overall, I think it might be better to do this as a SSA pass right
before lowering to source modifiers. I don't think we would be running
any optimizations after it that depend on the output being all 0's or
all 1's, but if you're concerned about that then we could add separate
gen5-compare compare
Hello list,
A candidate for the Mesa 10.4.7 release is now available. The current
patch queue is as follows:
- 24 queued
- 3 nominated (outstanding)
- and 7 rejected/obsolete patches
The current series includes bugfixes for the loaders - libEGL and libGL,
a some for the i965, freedreno, r300g
Previously, we put all the uniforms into one big array. The problem with
this approach is that, as soon as there was one indirect array acces, the
backend would decide that the entire large array should be pull constants.
This commit splits the array in half: first direct-only uniforms and then
po
On 13 March 2015 at 04:56, Ilia Mirkin wrote:
> We only program in one layer size per texture, so that means that all
> levels must share one size. This makes the piglit test
>
> bin/texelFetch fs sampler2DArray
>
> have the same breakage as its non-array version instead of being
> completely off.
v2: Delete the set of indirectly accessed variables when we're done with it
---
src/glsl/nir/nir.h | 4 +++
src/glsl/nir/nir_lower_io.c | 71 +
2 files changed, 75 insertions(+)
diff --git a/src/glsl/nir/nir.h b/src/glsl/nir/nir.h
index 0015e
Reviewed-by: Connor Abbott
A case of premature optimization I guess... I think I did it for
linking, but we'll cross that bridge when we come to it.
On Wed, Mar 18, 2015 at 7:45 PM, Jason Ekstrand wrote:
> We never did a single hash table lookup in the entire NIR code base that I
> found so the
Forgot the shader-db numbers:
total instructions in shared programs: 4114840 -> 4112172 (-0.06%)
instructions in affected programs: 43316 -> 40648 (-6.16%)
helped:116
HURT: 0
GAINED:0
LOST:
Previously, we just assigned variable locations in nir_lower_io. Now, we
force the user to assign variable locations for us. This gives the backend
a bit more control over where variables are placed.
---
src/glsl/nir/nir.h | 3 +++
src/glsl/nir/nir_lower_io.c
---
src/glsl/nir/nir.h | 4 +++
src/glsl/nir/nir_lower_io.c | 69 +
2 files changed, 73 insertions(+)
diff --git a/src/glsl/nir/nir.h b/src/glsl/nir/nir.h
index 0015ec2..ba48eea 100644
--- a/src/glsl/nir/nir.h
+++ b/src/glsl/nir/nir.h
@@ -1579
We never did a single hash table lookup in the entire NIR code base that I
found so there was no real benifit to doing it that way. I suppose that
for linking, we'll probably want to be able to lookup by name but we can
leave building that hash table to the linker. In the mean time this was
causi
Previously, we put all the uniforms into one big array. The problem with
this approach is that, as soon as there was one indirect array acces, the
backend would decide that the entire large array should be pull constants.
This commit splits the array in half: first direct-only uniforms and then
po
https://bugs.freedesktop.org/show_bug.cgi?id=89662
--- Comment #6 from Brian Paul ---
I can't reproduce this. Did you try a make clean first?
--
You are receiving this mail because:
You are the QA Contact for the bug.
You are the assignee for the bug.
__
On Wed, Mar 18, 2015 at 4:27 PM, Connor Abbott wrote:
> On Wed, Mar 18, 2015 at 2:58 PM, Matt Turner wrote:
>> On Thu, Mar 12, 2015 at 9:46 PM, Matt Turner wrote:
>>> Here are some ideas I think might be reasonable GSoC ideas.
>>>
>>> - GLSL linking in NIR
>>> - Would allow us to stop doin
On 03/18/2015 05:18 PM, Fredrik Höglund wrote:
This saves the cost of repeated hash table lookups when the same
vertex array object is referenced in a sequence of calls such as:
glVertexArrayAttribFormat(vao, ...);
glVertexArrayAttribBinding(vao, ...);
glEnableVertexArrayAttrib(va
On Wed, Mar 18, 2015 at 2:58 PM, Matt Turner wrote:
> On Thu, Mar 12, 2015 at 9:46 PM, Matt Turner wrote:
>> Here are some ideas I think might be reasonable GSoC ideas.
>>
>> - GLSL linking in NIR
>> - Would allow us to stop doing optimizations and other expensive
>> things on GLSL IR
>
> I
---
src/mapi/glapi/gen/ARB_direct_state_access.xml | 6 +++
src/mesa/main/tests/dispatch_sanity.cpp| 1 +
src/mesa/main/varray.c | 65 +++---
src/mesa/main/varray.h | 2 +
4 files changed, 58 insertions(+), 16 deletions
---
src/mapi/glapi/gen/ARB_direct_state_access.xml | 6 +
src/mesa/main/arrayobj.c | 34 ++
src/mesa/main/arrayobj.h | 2 ++
src/mesa/main/tests/dispatch_sanity.cpp| 1 +
4 files changed, 43 insertions(+)
diff --gi
---
docs/GL3.txt | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/docs/GL3.txt b/docs/GL3.txt
index 93fa60d..04a9e1e 100644
--- a/docs/GL3.txt
+++ b/docs/GL3.txt
@@ -198,7 +198,7 @@ GL 4.5, GLSL 4.50:
- Framebuffer object started (Laura Ekstran
This is needed to implement VertexArrayBindingDivisor.
---
src/mesa/main/varray.c | 10 ++
1 file changed, 6 insertions(+), 4 deletions(-)
diff --git a/src/mesa/main/varray.c b/src/mesa/main/varray.c
index 6cfe240..c9a328b 100644
--- a/src/mesa/main/varray.c
+++ b/src/mesa/main/varray.c
@
This series implements the VAO portion of GL_ARB_direct_state_access.
I started working on this before the decision was made to only expose
the extension in the core profile, so this series implements both the
core and compatibility behavior. The compatibility support actually
only amounts to a c
---
src/mapi/glapi/gen/ARB_direct_state_access.xml | 5 +++
src/mesa/main/tests/dispatch_sanity.cpp| 1 +
src/mesa/main/varray.c | 45 --
src/mesa/main/varray.h | 4 +++
4 files changed, 46 insertions(+), 9 deletion
---
src/mapi/glapi/gen/ARB_direct_state_access.xml | 9 +++
src/mesa/main/tests/dispatch_sanity.cpp| 1 +
src/mesa/main/varray.c | 90 ++
src/mesa/main/varray.h | 5 ++
4 files changed, 78 insertions(+), 27 deletion
This is needed to implement VertexArrayAttribBinding.
---
src/mesa/main/varray.c | 15 +--
1 file changed, 9 insertions(+), 6 deletions(-)
diff --git a/src/mesa/main/varray.c b/src/mesa/main/varray.c
index cdda6f3..4347a92 100644
--- a/src/mesa/main/varray.c
+++ b/src/mesa/main/varray
---
src/mapi/glapi/gen/ARB_direct_state_access.xml | 8 +++
src/mesa/main/tests/dispatch_sanity.cpp| 1 +
src/mesa/main/varray.c | 83 ++
src/mesa/main/varray.h | 4 ++
4 files changed, 71 insertions(+), 25 deletion
---
src/mapi/glapi/gen/ARB_direct_state_access.xml | 5 +++
src/mesa/main/tests/dispatch_sanity.cpp| 1 +
src/mesa/main/varray.c | 46 --
src/mesa/main/varray.h | 4 +++
4 files changed, 46 insertions(+), 10 deletio
The only difference between these functions is the legal types and
sizes, so consolidate the code into a single vertex_attrib_format()
function and call it from all three entry points.
---
src/mesa/main/varray.c | 143 +++--
1 file changed, 43 insertions
This is needed to implement VertexArrayVertexBuffer and
VertexArrayVertexBuffers.
---
src/mesa/main/varray.c | 16 +---
1 file changed, 9 insertions(+), 7 deletions(-)
diff --git a/src/mesa/main/varray.c b/src/mesa/main/varray.c
index 2017167..e41d32f 100644
--- a/src/mesa/main/varray
---
src/mesa/main/arrayobj.c | 50
src/mesa/main/arrayobj.h | 3 +++
2 files changed, 53 insertions(+)
diff --git a/src/mesa/main/arrayobj.c b/src/mesa/main/arrayobj.c
index 0c15630..ac081b0 100644
--- a/src/mesa/main/arrayobj.c
+++ b/src/mesa/mai
---
src/mapi/glapi/gen/ARB_direct_state_access.xml | 5
src/mesa/main/arrayobj.c | 34 ++
src/mesa/main/arrayobj.h | 2 ++
src/mesa/main/tests/dispatch_sanity.cpp| 1 +
4 files changed, 42 insertions(+)
diff --git
---
src/mapi/glapi/gen/ARB_direct_state_access.xml | 6 ++
src/mesa/main/tests/dispatch_sanity.cpp| 1 +
src/mesa/main/varray.c | 77 ++
src/mesa/main/varray.h | 4 ++
4 files changed, 66 insertions(+), 22 deletions
This is needed to implement glGetVertexArrayIndexediv and
glGetVertexArrayIndexed64iv.
---
src/mesa/main/varray.c | 22 ++
1 file changed, 14 insertions(+), 8 deletions(-)
diff --git a/src/mesa/main/varray.c b/src/mesa/main/varray.c
index 0faf158..13b9c01 100644
--- a/src/mesa
This is needed to implement VertexArrayAttrib*Format.
---
src/mesa/main/varray.c | 15 +--
1 file changed, 9 insertions(+), 6 deletions(-)
diff --git a/src/mesa/main/varray.c b/src/mesa/main/varray.c
index a6b66a0..93cc0e3 100644
--- a/src/mesa/main/varray.c
+++ b/src/mesa/main/varray
When true, this field indicates that double-precision floating point
attribute values are not converted to single-precision floats. It is
set to true by glVertexAttribLFormat and glVertexArrayAttribLFormat.
Based on a patch by Dave Airlie.
---
src/mesa/main/arrayobj.c | 1 +
src/mesa/main/mtype
---
src/mapi/glapi/gen/ARB_direct_state_access.xml | 7 +++
src/mesa/main/arrayobj.c | 20 +++-
src/mesa/main/arrayobj.h | 2 ++
src/mesa/main/tests/dispatch_sanity.cpp| 1 +
4 files changed, 25 insertions(+), 5 deletions(-
This saves the cost of repeated hash table lookups when the same
vertex array object is referenced in a sequence of calls such as:
glVertexArrayAttribFormat(vao, ...);
glVertexArrayAttribBinding(vao, ...);
glEnableVertexArrayAttrib(vao, ...);
...
Note that VAO's are container obje
---
src/mapi/glapi/gen/ARB_direct_state_access.xml | 25 +
src/mesa/main/tests/dispatch_sanity.cpp| 3 +
src/mesa/main/varray.c | 78 ++
src/mesa/main/varray.h | 15 +
4 files changed, 121 insertions(+)
d
---
src/mapi/glapi/gen/ARB_direct_state_access.xml | 14 +++
src/mesa/main/tests/dispatch_sanity.cpp| 2 +
src/mesa/main/varray.c | 117 +
src/mesa/main/varray.h | 10 +++
4 files changed, 143 insertions(+)
diff -
This parameter was added in OpenGL 4.3 and GL_ARB_direct_state_access.
---
src/mesa/main/varray.c | 5 +
1 file changed, 5 insertions(+)
diff --git a/src/mesa/main/varray.c b/src/mesa/main/varray.c
index bbfe947..ec768f0 100644
--- a/src/mesa/main/varray.c
+++ b/src/mesa/main/varray.c
@@ -872
On Wed, Mar 18 2015, Matt Turner wrote:
> This patch implements _mesa_roundeven{f,}, a function similar to the
> roundeven function added by a yet unimplemented technical specification
> (ISO/IEC TS 18661-1:2014), with a small difference in behavior -- we
> don't bother raising the inexact exceptio
On Wed, Mar 18, 2015 at 3:27 PM, Dave Airlie wrote:
> On 17 March 2015 at 01:39, Matt Turner wrote:
>> On Sun, Mar 15, 2015 at 10:22 PM, Dave Airlie wrote:
>>> We never reset the string on eglTerminate, so it grows
>>> for ever on multiple eglInitialise.
>>>
>>> Signed-off-by: Dave Airlie
>>> -
On 17 March 2015 at 01:39, Matt Turner wrote:
> On Sun, Mar 15, 2015 at 10:22 PM, Dave Airlie wrote:
>> We never reset the string on eglTerminate, so it grows
>> for ever on multiple eglInitialise.
>>
>> Signed-off-by: Dave Airlie
>> ---
>> src/egl/main/eglapi.c | 1 +
>> 1 file changed, 1 inse
Yeah, that looks better. V2 R-B me.
--Jason
On Wed, Mar 18, 2015 at 2:56 PM, Matt Turner wrote:
> total instructions in shared programs: 6263270 -> 6203091 (-0.96%)
> instructions in affected programs: 2606529 -> 2546350 (-2.31%)
> helped:14301
> GAINED:
On Wed, Mar 18, 2015 at 3:17 PM, Mark Janes wrote:
> Reviewed-by: Mark Janes
Thanks! Noted.
>
> Jason Ekstrand writes:
>
>> Before, we enabled NIR if you set INTEL_USE_NIR to anything which mean that
>> INTEL_USE_NIR=false would actually turn on NIR. In preparation for turning
>> NIR on by d
Reviewed-by: Mark Janes
Jason Ekstrand writes:
> Before, we enabled NIR if you set INTEL_USE_NIR to anything which mean that
> INTEL_USE_NIR=false would actually turn on NIR. In preparation for turning
> NIR on by default, this commit makes it smarter by allowing the
> INTEL_USE_NIR variable t
On Wed, Mar 18, 2015 at 1:06 PM, Laura Ekstrand wrote:
> So it turns out that this doesn't actually fix any bugs or add any features,
> stictly speaking. However, it does avoid a lot of kludginess. Previously, if
> you called
>
> glCopyTextureSubImage3D(texcube, 0, 0, 0, zoffset = 3, ...
>
> it w
On Tue, Mar 17, 2015 at 7:17 PM, Jason Ekstrand wrote:
> From: Matt Turner
>
> Shader-db results for FS instructions with NIR on HSW:
>
> total instructions in shared programs: 4186747 -> 4129871 (-1.36%)
> instructions in affected programs: 2438094 -> 2381218 (-2.33%)
> helped:
total instructions in shared programs: 6263270 -> 6203091 (-0.96%)
instructions in affected programs: 2606529 -> 2546350 (-2.31%)
helped:14301
GAINED:5
LOST: 3
Revewed-by: Jason Ekstrand [v1]
---
Reviewed-by: Carl Worth
---
src/util/Makefile.am | 4 +-
src/util/roundeven_test.c | 140 ++
2 files changed, 143 insertions(+), 1 deletion(-)
create mode 100644 src/util/roundeven_test.c
diff --git a/src/util/Makefile.am b/src/util/Makefile.am
The SSE 4.1 ROUND instructions let us implement roundeven directly.
Otherwise we assume that the rounding mode has not been modified (as we
do in the rest of Mesa) and use rint().
glibc uses the ROUND instruction in rint() after a cpuid check. This
patch just lets us inline it directly when we're
Eric's initial patch adding constant expression evaluation for
ir_unop_round_even used nearbyint. The open-coded _mesa_round_to_even
implementation came about without much explanation after a reviewer
asked whether nearbyint depended on the application not modifying the
rounding mode. Of course (as
Acked-by: Matt Turner
Feel free to push it.
___
mesa-dev mailing list
mesa-dev@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/mesa-dev
On 03/18/2015 01:29 PM, Ilia Mirkin wrote:
> On Wed, Mar 18, 2015 at 4:25 PM, Ian Romanick wrote:
>> From: Ian Romanick
>>
>> We use the idoim
>>
>>ir_foo *x = y->as_foo();
>>if (x == NULL)
>> return;
>>
>> all over the place. GCC generates some quite lovely code for this.
>> One s
https://bugs.freedesktop.org/show_bug.cgi?id=89662
--- Comment #5 from Christoph Haag ---
Created attachment 114451
--> https://bugs.freedesktop.org/attachment.cgi?id=114451&action=edit
similar error in line 525
I seem to get the same error in make install, with
627991dbf74ce5aee9ce75155fc27a4
On Wed, Mar 18, 2015 at 1:25 PM, Ian Romanick wrote:
> From: Ian Romanick
>
> We use the idoim
typo: idiom
>
>ir_foo *x = y->as_foo();
>if (x == NULL)
> return;
>
> all over the place. GCC generates some quite lovely code for this.
> One such example:
>
> 340a5b: 83 7d 18
---
docs/GL3.txt | 4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)
diff --git a/docs/GL3.txt b/docs/GL3.txt
index 5d59341..fcbb213 100644
--- a/docs/GL3.txt
+++ b/docs/GL3.txt
@@ -194,10 +194,10 @@ GL 4.5, GLSL 4.50:
GL_ARB_derivative_controlDONE (i965, nv50
On Wed, Mar 18, 2015 at 4:25 PM, Ian Romanick wrote:
> From: Ian Romanick
>
> We use the idoim
>
>ir_foo *x = y->as_foo();
>if (x == NULL)
> return;
>
> all over the place. GCC generates some quite lovely code for this.
> One such example:
>
> 340a5b: 83 7d 18 04
From: Ian Romanick
We use the idoim
ir_foo *x = y->as_foo();
if (x == NULL)
return;
all over the place. GCC generates some quite lovely code for this.
One such example:
340a5b: 83 7d 18 04 cmpl $0x4,0x18(%rbp)
340a5f: 0f 85 06 04 00 00 jne34
So it turns out that this doesn't actually fix any bugs or add any features,
stictly speaking. However, it does avoid a lot of kludginess. Previously, if
you called
glCopyTextureSubImage3D(texcube, 0, 0, 0, zoffset = 3, ...
it would grab the texture image object for face = 0 in teximage.c instea
Hi,
I am looking for a GSOC project. I am interested in working on driver.
Hardware available:
All AMD gpus (+1 APU)
NVIDIA k40, GTX780, Quadro 3500, GT755M
Intel: Haswell (buying on next pay date)
On Wed, Mar 18, 2015 at 2:58 PM, Matt Turner wrote:
> On Thu, Mar 12, 2015 at 9:46 PM, Matt Turne
On Mar 18, 2015 12:39 PM, "Ilia Mirkin" wrote:
>
> On Wed, Mar 18, 2015 at 3:36 PM, Jason Ekstrand
wrote:
> > diff --git a/src/glsl/nir/glsl_to_nir.cpp b/src/glsl/nir/glsl_to_nir.cpp
> > index 047cb51..e3c085a 100644
> > --- a/src/glsl/nir/glsl_to_nir.cpp
> > +++ b/src/glsl/nir/glsl_to_nir.cpp
>
On Wed, Mar 18, 2015 at 3:36 PM, Jason Ekstrand wrote:
> diff --git a/src/glsl/nir/glsl_to_nir.cpp b/src/glsl/nir/glsl_to_nir.cpp
> index 047cb51..e3c085a 100644
> --- a/src/glsl/nir/glsl_to_nir.cpp
> +++ b/src/glsl/nir/glsl_to_nir.cpp
> @@ -98,6 +98,9 @@ private:
> /* whether the IR we're ope
We never did a single hash table lookup in the entire NIR code base that I
found so there was no real benifit to doing it that way. I suppose that
for linking, we'll probably want to be able to lookup by name but we can
leave building that hash table to the linker. In the mean time this was
causi
On 03/17/2015 08:13 PM, Dave Airlie wrote:
This add primitive restart support to the prim conversion.
This involves changing the API for the translate functions
as we need to pass the prim restart index and the original
number of indices into the translate functions.
primitive restart is suppor
On Thu, Mar 12, 2015 at 9:46 PM, Matt Turner wrote:
> Here are some ideas I think might be reasonable GSoC ideas.
>
> - GLSL linking in NIR
> - Would allow us to stop doing optimizations and other expensive
> things on GLSL IR
Ian said this should wait until everything is using NIR so that
On 03/18/2015 09:34 AM, Jose Fonseca wrote:
> Note that GLboolean is an alias for unsigned char, which lacks the
> implicit true/false semantics that C++/C99 bool have.
> ---
> src/glsl/linker.cpp | 7 ---
> 1 file changed, 4 insertions(+), 3 deletions(-)
>
> diff --git a/src/glsl/linker.cpp
On Wed, Mar 18, 2015 at 11:51 AM, Ian Romanick wrote:
> Did this patch help anything?
Nope. I guess NIR is able to recognize something here that the GLSL IR can't.
Thanks for reviewing!
___
mesa-dev mailing list
mesa-dev@lists.freedesktop.org
http://li
Patches 1, 2, and 3 are
Reviewed-by: Ian Romanick
Patches 4 and 5 are
Acked-by: Ian Romanick
On 03/18/2015 11:37 AM, Matt Turner wrote:
> They're not accessible from the source language, but optimizations are
> allowed to generate them.
> ---
> src/glsl/ir_validate.cpp
Did this patch help anything?
On 03/18/2015 11:37 AM, Matt Turner wrote:
> Transform this into b2f(or(a, b)).
> ---
> src/glsl/opt_algebraic.cpp | 12
> 1 file changed, 12 insertions(+)
>
> diff --git a/src/glsl/opt_algebraic.cpp b/src/glsl/opt_algebraic.cpp
> index 0d1f3fc..98c852a
https://bugs.freedesktop.org/show_bug.cgi?id=89662
Brian Paul changed:
What|Removed |Added
Status|NEW |RESOLVED
Resolution|---
Reviewed-by: Mark Janes
Brian Paul writes:
> I wasn't aware of these _glapi_ stub functions when I committed
> 4bdbb588a9d385509f9168e38bfdb76952ba469c. Fixes "make check"
>
> Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=89662
> ---
> src/mesa/drivers/dri/common/dri_test.c | 11
Transform this into b2f(and(a, b)).
total instructions in shared programs: 6205448 -> 6204391 (-0.02%)
instructions in affected programs: 284030 -> 282973 (-0.37%)
helped:903
HURT: 6
---
src/glsl/nir/nir_opt_algebraic.py | 2 ++
Transform this into b2f(or(a, b)).
---
src/glsl/opt_algebraic.cpp | 12
1 file changed, 12 insertions(+)
diff --git a/src/glsl/opt_algebraic.cpp b/src/glsl/opt_algebraic.cpp
index 0d1f3fc..98c852a 100644
--- a/src/glsl/opt_algebraic.cpp
+++ b/src/glsl/opt_algebraic.cpp
@@ -421,6 +421
1 - 100 of 155 matches
Mail list logo
