On 02/06/16 07:43, Timothy Arceri wrote:
> On Thu, 2016-06-02 at 07:22 +0200, Samuel Iglesias Gonsálvez wrote:
>> On 26/05/16 07:56, Samuel Iglesias Gonsálvez wrote:
>>>
>>> Hello,
>>>
>>> Timothy found that tests with unpacked double vector input varyings
>>> were failing in i965 driver. For
On Thu, 2016-06-02 at 07:22 +0200, Samuel Iglesias Gonsálvez wrote:
> On 26/05/16 07:56, Samuel Iglesias Gonsálvez wrote:
> >
> > Hello,
> >
> > Timothy found that tests with unpacked double vector input varyings
> > were failing in i965 driver. For example, this is happening when
> > using
Matt Turner writes:
> Coverity warns in multiple places about the potential for division by
> zero, caused by this function's default case.
>
> Cc: Francisco Jerez
Reviewed-by: Francisco Jerez
> ---
>
---
src/compiler/glsl/link_uniforms.cpp | 20 +---
1 file changed, 13 insertions(+), 7 deletions(-)
diff --git a/src/compiler/glsl/link_uniforms.cpp
b/src/compiler/glsl/link_uniforms.cpp
index a7f136c..571c49f 100644
--- a/src/compiler/glsl/link_uniforms.cpp
+++
On 26/05/16 07:56, Samuel Iglesias Gonsálvez wrote:
> Hello,
>
> Timothy found that tests with unpacked double vector input varyings
> were failing in i965 driver. For example, this is happening when
> using explicit locations because Mesa disables varying packing for
> that case.
>
> These
Looks good. R-B
On Wed, Jun 1, 2016 at 5:21 PM, Francisco Jerez
wrote:
> Skipping the temporary allocation and copy instructions is easy (just
> return dst), but the conditions used to find out whether the copy can
> be optimized out safely without breaking the program
On Wed, Jun 1, 2016 at 12:01 PM, Kenneth Graunke wrote:
> This writes linked shader programs to .shader_test files to
> $MESA_SHADER_CAPTURE_PATH in the format used by shader-db
> (http://cgit.freedesktop.org/mesa/shader-db).
>
> It supports both GLSL shaders and ARB
Signed-off-by: Kenneth Graunke
Cc: "12.0"
---
src/mesa/drivers/dri/i965/gen6_sf_state.c | 2 +-
src/mesa/drivers/dri/i965/gen7_sf_state.c | 2 +-
src/mesa/drivers/dri/i965/gen8_sf_state.c | 2 +-
3 files changed, 3 insertions(+), 3
I believe we were using the state-based point size when either the
tessellation evaluation or geometry shader, but the vertex shader
didn't write it. If the last enabled stage writes it (corresponding
to brw_vue_map_geom_out), we should use the shader's value, otherwise
state. This should handle
From: Dave Airlie
The specs says INVALID_VALUE for exceeding dimensions,
which is really what is happening here.
This fixes:
GL45-CTS.copy_image.non_existent_mipmap
Signed-off-by: Dave Airlie
---
src/mesa/main/copyimage.c | 2 +-
1 file changed, 1
Signed-off-by: Jason Ekstrand
Cc: "12.0"
---
src/intel/vulkan/gen7_pipeline.c | 4 +++-
src/intel/vulkan/gen8_pipeline.c | 4 +++-
src/intel/vulkan/genX_pipeline_util.h | 30 +-
3 files changed, 35
Signed-off-by: Jason Ekstrand
Cc: "12.0"
---
src/intel/vulkan/gen7_pipeline.c | 10 +-
src/intel/vulkan/gen8_pipeline.c | 9 -
2 files changed, 17 insertions(+), 2 deletions(-)
diff --git
We were using this briefly in the i965 driver to trigger recompiles but we
haven't been using it since we switched to the NIR y-transform lowering
pass.
Signed-off-by: Jason Ekstrand
---
src/compiler/glsl/glsl_to_nir.cpp | 3 ---
src/compiler/nir/nir.h | 3 ---
This is more accurate than calling
_mesa_active_fragment_shader_has_side_effects because it looks at whether
or not the SSBOs, images, or atomic buffers are actually written rather
than just existing in the program.
Signed-off-by: Jason Ekstrand
Cc: "12.0"
It is no longer used.
Signed-off-by: Jason Ekstrand
---
src/mesa/main/mtypes.h | 18 --
1 file changed, 18 deletions(-)
diff --git a/src/mesa/main/mtypes.h b/src/mesa/main/mtypes.h
index 2233526..4f93a39 100644
--- a/src/mesa/main/mtypes.h
+++
Signed-off-by: Jason Ekstrand
---
src/mesa/drivers/dri/i965/gen7_wm_state.c | 8 +++-
src/mesa/drivers/dri/i965/gen8_ps_state.c | 7 +++
2 files changed, 6 insertions(+), 9 deletions(-)
diff --git a/src/mesa/drivers/dri/i965/gen7_wm_state.c
From: Dave Airlie
This test was only happening for textures, but there is
nothing in the spec to say this, so test it for all cases.
This fixes:
GL45-CTS.copy_image.invalid_target
Signed-off-by: Dave Airlie
---
src/mesa/main/copyimage.c | 11
On Wed, 2016-06-01 at 14:36 +0200, Serge Martin wrote:
> This fix getting the size of a struct arg. vec3 types still work ok.
> Only buit-in args need to have power of two alignement, getTypeAllocSize
> reports the corect size.
"alignment", "correct"
otherwise LGTM. you probably want to cc
On Wednesday, June 1, 2016 2:44:53 PM PDT Jason Ekstrand wrote:
> This fixes about 100 of the new Vulkan CTS tests.
>
> Signed-off-by: Jason Ekstrand
> Cc: "12.0"
> Cc: Connor Abbott
> Cc: Ian Romanick
On Wednesday, June 1, 2016 2:44:58 PM PDT Jason Ekstrand wrote:
> Signed-off-by: Jason Ekstrand
> Cc: "12.0"
> Cc: Ian Romanick
> ---
> src/intel/vulkan/gen7_pipeline.c | 12 ++--
>
On Wed, 2016-06-01 at 16:22 +1000, Timothy Arceri wrote:
> This generates the program ids at cache upload time rather than at
> program creation time.
>
> Moving the id generation here will be useful for on-disk shader
> cache support because it means we don't generate ids if there was
> a cache
On Wed, Jun 1, 2016 at 9:57 PM, Marek Olšák wrote:
> Hi,
>
> This is version 2 of the previous series. This time it's been tested!!
>
> Tested cards:
> - RV670
> - RV730
> - EG/REDWOOD
> - CAYMAN
All good on JUNIPER now (piglit and a few random games).
Tested-by: Grazvydas
Skipping the temporary allocation and copy instructions is easy (just
return dst), but the conditions used to find out whether the copy can
be optimized out safely without breaking the program are rather
complex: The destination must be exactly one component of at most the
execution width of the
On 06/01/2016 08:57 PM, Marek Olšák wrote:
From: Marek Olšák
Submitting a DMA IB flushes the GFX IB and all GPU caches.
Reviewed-by: Alex Deucher
On Tonga 380X, this improves The Talos Principle from 8.3 fps to 28.3
fps (all graphics
On Wednesday, June 1, 2016 2:44:51 PM PDT Jason Ekstrand wrote:
> I recently grabbed the latest dev version of the Vulkan CTS and ran it on
> our driver. This series fixes a bunch of the bugs that it exposed. In an
> effort to get more people involved in Vulkan development and in the hopes
> of
The conditional mod of these instructions determines the semantics of
the comparison itself (rather than being evaluated based on the result
of the instruction as is usually the case for most other instructions
that allow conditional mods), so it's in general not legal to
propagate a conditional
On Wed, 2016-06-01 at 14:55 -0700, Ian Romanick wrote:
> On 05/31/2016 04:45 PM, Ian Romanick wrote:
> >
> > On 05/31/2016 02:44 PM, Timothy Arceri wrote:
> > >
> > > On Tue, 2016-05-31 at 11:52 -0700, Ian Romanick wrote:
> > > >
> > > > From: Ian Romanick
> > > >
>
On Jun 1, 2016 4:21 PM, "Jason Ekstrand" wrote:
>
>
> On May 31, 2016 9:27 AM, "Emil Velikov" wrote:
> >
> > From: Emil Velikov
> >
> > Signed-off-by: Emil Velikov
> > ---
> >
On Jun 1, 2016 4:20 PM, "Jason Ekstrand" wrote:
>
>
> On May 31, 2016 9:27 AM, "Emil Velikov" wrote:
> >
> > From: Emil Velikov
> >
> > Including the file in both ISL_FILES and ISL_GENERATED_FILES makes
> > the actual
On May 31, 2016 9:27 AM, "Emil Velikov" wrote:
>
> From: Emil Velikov
>
> Signed-off-by: Emil Velikov
> ---
> src/gallium/drivers/swr/Makefile.am | 1 +
> src/intel/isl/Makefile.am | 1 +
> 2 files
On May 31, 2016 9:27 AM, "Emil Velikov" wrote:
>
> From: Emil Velikov
>
> Including the file in both ISL_FILES and ISL_GENERATED_FILES makes
> the actual dependency list less obvious.
>
> Signed-off-by: Emil Velikov
On Wed, Jun 1, 2016 at 3:58 PM, Francisco Jerez
wrote:
> Jason Ekstrand writes:
>
> > On Wed, Jun 1, 2016 at 3:33 PM, Francisco Jerez
> > wrote:
> >
> >> Jason Ekstrand writes:
> >>
> >> > This fixes 64
Jason Ekstrand writes:
> On Wed, Jun 1, 2016 at 3:33 PM, Francisco Jerez
> wrote:
>
>> Jason Ekstrand writes:
>>
>> > This fixes 64 Vulkan CTS tests per gen
>> >
>> > Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=96299
I made a few cosmetic comments on patch 9 and I made a comment on patch 5
that might make a good future cleanup. However, I think it's good enough
to land.
Reviewed-by: Jason Ekstrand
Don't forget to cc 12.0
--Jason
On Wed, Jun 1, 2016 at 3:04 PM, Jordan Justen
On Wed, Jun 1, 2016 at 3:04 PM, Jordan Justen
wrote:
> The cross thread constant support appears on Haswell. It allows us to
> upload a set of uniform data for all threads without duplicating it
> per thread.
>
> We also support per-thread data which allows us to store
On Wed, Jun 1, 2016 at 3:04 PM, Jordan Justen
wrote:
> Signed-off-by: Jordan Justen
> ---
> src/intel/vulkan/anv_cmd_buffer.c | 7 +++
> src/intel/vulkan/anv_private.h| 1 -
> src/intel/vulkan/gen7_cmd_buffer.c
On Wed, Jun 1, 2016 at 3:33 PM, Francisco Jerez
wrote:
> Jason Ekstrand writes:
>
> > This fixes 64 Vulkan CTS tests per gen
> >
> > Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=96299
> > Cc: "12.0"
> >
Jason Ekstrand writes:
> This fixes 64 Vulkan CTS tests per gen
>
> Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=96299
> Cc: "12.0"
> Cc: Francisco Jerez
> Cc: Mark Janes
> ---
>
On Wed, Jun 1, 2016 at 3:04 PM, Jordan Justen
wrote:
> This thread ID uniform will be used to compute the
> gl_LocalInvocationIndex and gl_LocalInvocationID values.
>
> It is important for this uniform to be added in the last push constant
> register.
This fixes 64 Vulkan CTS tests per gen
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=96299
Cc: "12.0"
Cc: Francisco Jerez
Cc: Mark Janes
---
src/mesa/drivers/dri/i965/brw_fs.cpp | 8
1 file
v2:
* simd16/32 fixes (curro)
Signed-off-by: Jordan Justen
---
src/compiler/nir/nir_intrinsics.h| 1 +
src/mesa/drivers/dri/i965/brw_fs_nir.cpp | 15 +++
2 files changed, 16 insertions(+)
diff --git a/src/compiler/nir/nir_intrinsics.h
We need information about push constants in a few places for the GL
driver, and another couple places for the vulkan driver.
When we add support for uploading both a common (cross-thread) set of
push constants, combined with the previous per-thread push constant
data, things are going to get even
v4:
* Force thread_local_id_index to -1 for now, and have
fs_visitor::setup_cs_payload look at thread_local_id_index. This
enables us to more easily cut over from the old local ID layout to
the new layout, as suggested by Jason.
Signed-off-by: Jordan Justen
We add a lowering pass for nir intrinsics. This pass can replace nir
intrinsics with driver specific nir lower code.
We lower the gl_LocalInvocationIndex intrinsic based on a uniform
which is loaded with a thread specific ID.
We also lower the gl_LocalInvocationID based on
The cross thread constant support appears on Haswell. It allows us to
upload a set of uniform data for all threads without duplicating it
per thread.
One complication is that cross-thread constants are loaded into
registers before per-thread constants. Previously, our local IDs were
loaded before
Signed-off-by: Jordan Justen
---
src/intel/vulkan/anv_cmd_buffer.c | 7 +++
src/intel/vulkan/anv_private.h| 1 -
src/intel/vulkan/gen7_cmd_buffer.c| 2 +-
src/intel/vulkan/gen8_cmd_buffer.c| 2 +-
The cross thread constant support appears on Haswell. It allows us to
upload a set of uniform data for all threads without duplicating it
per thread.
We also support per-thread data which allows us to store a per-thread
ID in one of the uniforms that can be used to calculate the
This thread ID uniform will be used to compute the
gl_LocalInvocationIndex and gl_LocalInvocationID values.
It is important for this uniform to be added in the last push constant
register. fs_visitor::assign_constant_locations is updated to make
sure this happens.
The reason this is important is
The cross thread constant support appears on Haswell. It allows us to
upload a set of uniform data for all threads without duplicating it
per thread.
We also support per-thread data which allows us to store a per-thread
ID in one of the uniforms that can be used to calculate the
The old method pushed data for each channels uvec3 data of
gl_LocalInvocationID.
The new method pushes 1 dword of data that is a 'thread local ID'
value. Based on that value, we can generate gl_LocalInvocationIndex
and gl_LocalInvocationID with some calculations.
Signed-off-by: Jordan Justen
Signed-off-by: Jordan Justen
---
src/compiler/nir/nir.c | 4
src/compiler/nir/nir.h | 2 ++
src/compiler/nir/nir_gather_info.c | 1 +
src/compiler/nir/nir_intrinsics.h | 1 +
git://people.freedesktop.org/~jljusten/mesa hsw-cs-cross-thread-constants-v4
v4:
* Support both the old and new layouts until the switch-over to the
new layout. This minimizes the size of the switch over patch.
(Jason)
v3:
*
v2:
* Move lower flag to context constants. (Ken)
Signed-off-by: Jordan Justen
Reviewed-by: Kenneth Graunke (v1)
---
src/compiler/glsl/builtin_variables.cpp | 29 ++---
src/compiler/glsl/glsl_parser_extras.cpp | 2 +-
On 05/31/2016 04:45 PM, Ian Romanick wrote:
> On 05/31/2016 02:44 PM, Timothy Arceri wrote:
>> On Tue, 2016-05-31 at 11:52 -0700, Ian Romanick wrote:
>>> From: Ian Romanick
>>>
>>> Our piglit tests for geometry and tessellation shader inputs were
>>> incorrect. Array
Signed-off-by: Jason Ekstrand
Cc: "12.0"
Cc: Ian Romanick
---
src/intel/vulkan/gen7_pipeline.c | 12 ++--
src/intel/vulkan/gen8_pipeline.c | 12 ++--
src/intel/vulkan/genX_pipeline_util.h |
Signed-off-by: Jason Ekstrand
Cc: "12.0"
Cc: Ian Romanick
---
src/compiler/spirv/spirv_to_nir.c | 7 ---
1 file changed, 4 insertions(+), 3 deletions(-)
diff --git a/src/compiler/spirv/spirv_to_nir.c
This fixes about 100 of the new Vulkan CTS tests.
Signed-off-by: Jason Ekstrand
Cc: "12.0"
Cc: Connor Abbott
Cc: Ian Romanick
Cc: Kenneth Graunke
---
I recently grabbed the latest dev version of the Vulkan CTS and ran it on
our driver. This series fixes a bunch of the bugs that it exposed. In an
effort to get more people involved in Vulkan development and in the hopes
of actually getting reviews, I've CC'd at least one person on each patch.
Signed-off-by: Jason Ekstrand
Cc: "12.0"
Cc: Jordan Justen
---
src/intel/vulkan/anv_pipeline.c | 9 -
1 file changed, 4 insertions(+), 5 deletions(-)
diff --git a/src/intel/vulkan/anv_pipeline.c
Signed-off-by: Jason Ekstrand
Cc: "12.0"
Cc: Jordan Justen
---
src/intel/vulkan/gen7_pipeline.c | 37 +-
src/intel/vulkan/gen8_pipeline.c | 49 -
This fixes the 7 dEQP-VK.pipeline.spec_constant.compute.local_size.* tests
in the latest dev version of the Vulkan CTS.
Signed-off-by: Jason Ekstrand
Cc: "12.0"
Cc: Ian Romanick
---
This is more consistent with gen8+
Signed-off-by: Jason Ekstrand
Cc: "12.0"
Cc: Jordan Justen
---
src/intel/genxml/gen6.xml | 4 ++--
src/intel/genxml/gen7.xml | 4 ++--
src/intel/genxml/gen75.xml
Signed-off-by: Jason Ekstrand
Cc: "12.0"
Cc: Jordan Justen
---
src/intel/vulkan/gen7_pipeline.c | 8 +++-
src/intel/vulkan/gen8_pipeline.c | 7 ++-
2 files changed, 13 insertions(+), 2 deletions(-)
diff --git
Signed-off-by: Jason Ekstrand
Cc: "12.0"
Cc: Nanley Chery
---
src/intel/vulkan/anv_meta_clear.c | 6 --
1 file changed, 4 insertions(+), 2 deletions(-)
diff --git a/src/intel/vulkan/anv_meta_clear.c
On Wed, Jun 1, 2016 at 2:57 PM, Marek Olšák wrote:
> Hi,
>
> This is version 2 of the previous series. This time it's been tested!!
>
> Tested cards:
> - RV670
> - RV730
> - EG/REDWOOD
> - CAYMAN
>
> This patch series:
> - fixes several bugs around making 3D and CP DMA idle with
On 05/13/2016 07:57 AM, Alejandro Piñeiro wrote:
> There are several internalformat that are not supported on gl es, so
> it should return -1 if that is the case. This is needed in order to
> get ARB_internalformat_query2 implementation deciding correctly if
> a resource is supported or not on
On 05/13/2016 07:57 AM, Alejandro Piñeiro wrote:
> For ES 3.0 NUM_SAMPLE_COUNTS spec points that some formats will be
> always zero. But on ES 3.1 can be different to zero.
>
> The current code is correctly checking exactly against version 3.0,
> but the comment only mentions 3.0 spec. It is
This patch is
Reviewed-by: Ian Romanick
On 05/13/2016 07:57 AM, Alejandro Piñeiro wrote:
> The comment clarifies that the driver is called only to try to get
> a preferred internalformat, and that it was already checked if the
> format is supported or not.
>
>
This patch is
Reviewed-by: Ian Romanick
On 05/13/2016 07:57 AM, Alejandro Piñeiro wrote:
> Right now the implementation only checks if the internalformat is
> supported or not. But that implementation is wrong, returning
> unsupported for some internalformats.
This writes linked shader programs to .shader_test files to
$MESA_SHADER_CAPTURE_PATH in the format used by shader-db
(http://cgit.freedesktop.org/mesa/shader-db).
It supports both GLSL shaders and ARB programs. All stages that
are linked together are written in a single .shader_test file.
This
On Wed, Jun 1, 2016 at 9:31 AM, Plamena Manolova
wrote:
>
> When validating attributes during surface creation we should account
> for the default values of texture target and format (EGL_NO_TEXTURE)
> since the user is not obligated to explicitly set both via the
>
From: Marek Olšák
Reviewed-by: Alex Deucher
---
src/gallium/drivers/r600/evergreen_state.c | 25 +++--
src/gallium/drivers/r600/r600_hw_context.c | 4
src/gallium/drivers/r600/r600_state.c| 25
From: Marek Olšák
v3: use PFP_SYNC_ME on EG-CM only when supported by the kernel,
otherwise use MEM_WRITE + WAIT_REG_MEM to emulate that
---
src/gallium/drivers/r600/evergreen_hw_context.c | 16 --
src/gallium/drivers/r600/evergreend.h | 1 +
On 06/01/2016 11:35 AM, Alejandro Piñeiro wrote:
> On 13/05/16 23:20, Alejandro Piñeiro wrote:
>>
>> On 13/05/16 17:06, Ilia Mirkin wrote:
>>> On Fri, May 13, 2016 at 10:57 AM, Alejandro Piñeiro
>>> wrote:
Earlier this year the support for ARB_internalformat_query2 has
From: Marek Olšák
The main impact is that fast color clear doesn't flush TC, CONST, DB.
Reviewed-by: Alex Deucher
---
src/gallium/drivers/r600/evergreen_hw_context.c | 15 +--
src/gallium/drivers/r600/r600_pipe.h| 17
From: Marek Olšák
Submitting a DMA IB flushes the GFX IB and all GPU caches.
Reviewed-by: Alex Deucher
---
src/gallium/drivers/radeon/r600_buffer_common.c | 9 -
1 file changed, 4 insertions(+), 5 deletions(-)
diff --git
From: Marek Olšák
---
src/gallium/drivers/r600/r600_blit.c | 6 --
1 file changed, 6 deletions(-)
diff --git a/src/gallium/drivers/r600/r600_blit.c
b/src/gallium/drivers/r600/r600_blit.c
index 76c3364..c9d7823 100644
--- a/src/gallium/drivers/r600/r600_blit.c
+++
From: Marek Olšák
This has been wrong all along. Fixing this will allow removing useless
cache flushes.
Cc: 11.1 11.2 12.0
---
src/gallium/drivers/r600/r600_hw_context.c | 19 +++
1 file changed, 11 insertions(+), 8
From: Marek Olšák
Just move the alignment parameter from u_suballocator_create
to u_suballocator_alloc.
---
src/gallium/auxiliary/util/u_suballoc.c | 22 ++
src/gallium/auxiliary/util/u_suballoc.h | 6 +++---
From: Marek Olšák
The main impact is that {upload, draw, upload, draw, ..} doesn't flush
framebuffer caches before every upload.
Reviewed-by: Alex Deucher
---
src/gallium/drivers/r600/r600_hw_context.c | 15 +--
1 file changed, 1
From: Marek Olšák
This will allow removing useless cache & IB flushes.
---
src/gallium/drivers/r600/r600_hw_context.c | 9 -
1 file changed, 8 insertions(+), 1 deletion(-)
diff --git a/src/gallium/drivers/r600/r600_hw_context.c
From: Marek Olšák
---
src/gallium/drivers/radeon/r600_pipe_common.c | 8
src/gallium/drivers/radeon/r600_pipe_common.h | 2 +-
src/gallium/drivers/radeon/r600_streamout.c | 2 +-
3 files changed, 6 insertions(+), 6 deletions(-)
diff --git
Hi,
This is version 2 of the previous series. This time it's been tested!!
Tested cards:
- RV670
- RV730
- EG/REDWOOD
- CAYMAN
This patch series:
- fixes several bugs around making 3D and CP DMA idle with respect to CP.PFP,
which allows removing a lot of cache flushes (= hacks really) and IB
Alejandro Piñeiro writes:
> Among other thigs, fix a gpu hang when using INTEL_DEBUG=shader_time
> for any shader.
>
> Signed-off-by: Jason Ekstrand
> Signed-off-by: Alejandro Piñeiro
> ---
>
> This is the change suggested by
On Wed, Jun 1, 2016 at 8:51 AM, ⚛ <0xe2.0x9a.0...@gmail.com> wrote:
> On Wed, Jun 1, 2016 at 3:50 PM, Erik Faye-Lund wrote:
>> On Wed, Jun 1, 2016 at 3:02 PM, ⚛ <0xe2.0x9a.0...@gmail.com> wrote:
>>> On Wed, Jun 1, 2016 at 2:19 PM, Marek Olšák wrote:
On 13/05/16 23:20, Alejandro Piñeiro wrote:
>
> On 13/05/16 17:06, Ilia Mirkin wrote:
>> On Fri, May 13, 2016 at 10:57 AM, Alejandro Piñeiro
>> wrote:
>>> Earlier this year the support for ARB_internalformat_query2 has landed
>>> [1][2], initially only for desktop GL.
>>>
On Wed, Jun 1, 2016 at 7:40 PM, Patrick Baggett
wrote:
>>
>>
>> No. Shader compilation can only be asynchronous if it's far enough
>> from a draw call and the app doesn't query its status. If it's next to
>> a draw call, multithreading is useless. Completely useless.
>>
On 01.06.2016 19:31, Eric Engestrom wrote:
On Wed, Jun 01, 2016 at 12:18:57AM +0200, Tobias Klausmann wrote:
Signed-off-by: Tobias Klausmann
---
src/Makefile.am | 4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)
diff --git a/src/Makefile.am
On 05/31/2016 07:39 AM, Ilia Mirkin wrote:
> This has the unfortunate side-effect of opening up cube map arrays to
> ES 3.0 implementations where the backend driver also supports texture
> cubemaps for desktop GL (I'm thinking the DX10.1 GT21x's for example).
> Perhaps we don't care? Otherwise it
>
>
> No. Shader compilation can only be asynchronous if it's far enough
> from a draw call and the app doesn't query its status. If it's next to
> a draw call, multithreading is useless. Completely useless.
>
I don't know a lot about the shader compilation/linking process, so
I'm just asking
On Wed, Jun 01, 2016 at 12:18:57AM +0200, Tobias Klausmann wrote:
> Signed-off-by: Tobias Klausmann
> ---
> src/Makefile.am | 4 ++--
> 1 file changed, 2 insertions(+), 2 deletions(-)
>
> diff --git a/src/Makefile.am b/src/Makefile.am
> index
[+RobH since he's interested in plumbing similar/same formats into GBM]
Hi Philipp,
On 26 May 2016 at 13:00, Philipp Zabel wrote:
> Am Donnerstag, den 26.05.2016, 12:43 +0100 schrieb Emil Velikov:
>> Hi gents,
>>
>> On 26 May 2016 at 11:28, Philipp Zabel
Reviewed-by: Roland Scheidegger
Am 01.06.2016 um 17:32 schrieb Nicolai Hähnle:
> From: Nicolai Hähnle
>
> v2:
> - TG4 does not calculate derivatives (Ilia)
> - also handle SAMPLE* instructions (Roland)
>
> Cc: 12.0
> >>> > MESA_COMMON_MK := $(MESA_TOP)/Android.common.mk
> >>> > MESA_PYTHON2 := python
> >>> > +MESA_PYTHON3 := python3
> >>
> >> I've just seen that while a few days ago python3 was necessary to build
> >> gen%_pack.h headers,
> >> now the .py scripts just require python2, so this MESA_PYTHON3
On 06/01/2016 05:19 AM, Marek Olšák wrote:
> On Fri, May 27, 2016 at 8:49 PM, ⚛ <0xe2.0x9a.0...@gmail.com> wrote:
>> Hello.
>>
>> http://en.cppreference.com/w/cpp/thread/future
>> http://en.cppreference.com/w/cpp/thread/async
>>
>> Assumption: Shader compilation will need run on separate
On Wed, Jun 1, 2016 at 6:06 PM, ⚛ <0xe2.0x9a.0...@gmail.com> wrote:
> On Wed, Jun 1, 2016 at 3:53 PM, Marek Olšák wrote:
>> Because of external factors you can't predict, your driver suddenly
>> receives a bunch of shaders that take 2000 ms to compile right before
>> a draw
On 1 June 2016 at 17:34, ⚛ <0xe2.0x9a.0...@gmail.com> wrote:
> On Wed, Jun 1, 2016 at 4:06 PM, Brian Paul wrote:
>> I think the issue here is someone comes along with no history in the project
>> and asserts that he has a great solution for a problem
>
> Asynchronous
On 1 June 2016 at 01:25, Mauro Rossi wrote:
> 2016-06-01 2:22 GMT+02:00 Mauro Rossi :
>> 2016-05-31 16:33 GMT+02:00 Emil Velikov :
>>>
>>> Hi Mauro,
>>>
>>> A couple of questions, nothing serious imho.
>>>
>>> On 30 May 2016
Among other thigs, fix a gpu hang when using INTEL_DEBUG=shader_time
for any shader.
Signed-off-by: Jason Ekstrand
Signed-off-by: Alejandro Piñeiro
---
This is the change suggested by Jason (so I added him as signed-off). I just
run a full piglit run
>
> > With gcc 5.3.1 I end up with lib{GL,OSMesa}.so @ 44M and
> > libswrAVX{,2}.so @ 70M. With flto turned on it drops WAY down to
> > lib{GL,OSMesa}.so @ 13M and libswrAVX{,2}.so @ 18M
>
> I assume those numbers are including debugging symbols? How do stripped
> binaries compare?
>
Silly me, I
1 - 100 of 220 matches
Mail list logo