Quoting Rob Clark (2017-10-14 11:41:49)
> Signed-off-by: Rob Clark
> ---
> v1: actually build freedreno
> v2: corrected installed ${driver}_dri.so name
>
> meson.build | 6 +
> meson_options.txt| 2 +-
> src/gallium/drivers/freedre
Quoting Jakob Bornecrantz (2017-10-14 13:03:14)
> On Sat, Oct 14, 2017 at 1:36 AM, Dylan Baker wrote:
> > I'm not sure about this approach, we would need a way to add depends to
> > meson,
> > but I'm also worried that calling make adds another dependency that could be
> > problematic for windows
The series looks good to me, with the comment on patch 5 fixed.
Reviewed-by: Bas Nieuwenhuizen
for the series. However, I'd be a lot more comfortable landing this if
it sees a CTS run. Will kick one off tonight and see what happens.
On Sun, Oct 15, 2017 at 1:36 AM, Timothy Arceri wrote:
> This
On Sun, Oct 15, 2017 at 1:36 AM, Timothy Arceri wrote:
> From: Bas Nieuwenhuizen
>
> This is a combined shader creation helper than will help us to
> create the shaders for each stage at once. This will allow us to
> do some link time optimisations.
>
> Signed-off-by: Timothy Arceri
> ---
> src
From: Bas Nieuwenhuizen
Here we make use of NIR the linking helpers to remove unused
varyings.
Sascha Willems demo results:
computecullandlod 39 -> 41 fps
pipelines ~6100 -> ~6200 fps
Signed-off-by: Bas Nieuwenhuizen
Signed-off-by: Timothy Arceri
---
src/amd/vulkan/radv_pipeline.c | 42
---
src/amd/vulkan/radv_pipeline.c | 220 ---
src/amd/vulkan/radv_pipeline_cache.c | 26 -
src/amd/vulkan/radv_private.h| 8 --
3 files changed, 254 deletions(-)
diff --git a/src/amd/vulkan/radv_pipeline.c b/src/amd/vulkan/radv_pipeline.c
index
From: Bas Nieuwenhuizen
This will be used to create a hash of the combined shaders in the
pipeline.
Signed-off-by: Timothy Arceri
---
src/amd/vulkan/radv_pipeline_cache.c | 33 +
src/amd/vulkan/radv_private.h| 7 +++
2 files changed, 40 insertions(+
This initial linking support removes unused varyings across
stages.
Future improvements include enabling removing unused varying
components [1], varying packing (in progress), and varying
array splitting.
I've tried to run the series on the public Vulkan CTS but I
seem to be hitting an existing i
---
src/amd/vulkan/radv_pipeline.c | 4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)
diff --git a/src/amd/vulkan/radv_pipeline.c b/src/amd/vulkan/radv_pipeline.c
index 5e409ce767..415a6cd62b 100644
--- a/src/amd/vulkan/radv_pipeline.c
+++ b/src/amd/vulkan/radv_pipeline.c
@@ -1759,30 +1759
---
src/amd/vulkan/radv_pipeline.c | 114 +++--
1 file changed, 29 insertions(+), 85 deletions(-)
diff --git a/src/amd/vulkan/radv_pipeline.c b/src/amd/vulkan/radv_pipeline.c
index f923027036..446e14226f 100644
--- a/src/amd/vulkan/radv_pipeline.c
+++ b/src/amd
---
src/amd/vulkan/radv_pipeline.c | 27 +---
src/amd/vulkan/radv_pipeline_cache.c | 127 ---
src/amd/vulkan/radv_private.h| 12 +---
3 files changed, 17 insertions(+), 149 deletions(-)
diff --git a/src/amd/vulkan/radv_pipeline.c b/src/amd/vulka
From: Bas Nieuwenhuizen
Signed-off-by: Timothy Arceri
---
src/amd/vulkan/radv_pipeline_cache.c | 155 +++
src/amd/vulkan/radv_private.h| 14
2 files changed, 169 insertions(+)
diff --git a/src/amd/vulkan/radv_pipeline_cache.c
b/src/amd/vulkan/radv
From: Bas Nieuwenhuizen
This is a combined shader creation helper than will help us to
create the shaders for each stage at once. This will allow us to
do some link time optimisations.
Signed-off-by: Timothy Arceri
---
src/amd/vulkan/radv_pipeline.c | 130
From: Bas Nieuwenhuizen
Signed-off-by: Timothy Arceri
---
src/amd/vulkan/radv_pipeline_cache.c | 102 ++-
1 file changed, 64 insertions(+), 38 deletions(-)
diff --git a/src/amd/vulkan/radv_pipeline_cache.c
b/src/amd/vulkan/radv_pipeline_cache.c
index fc99b43fff
Andy Furniss wrote:
Since
commit 13a53c4f5cdd664fd155c9e78fb46a4387af006c
Author: Emil Velikov
Date: Thu Oct 5 11:19:05 2017 +0100
configure.ac: rework llvm libs handling for 3.9+
I am getting 00s of
/mesa/src/amd/common/. undefined reference to LLVM..
Using git llvm hav
Since
commit 13a53c4f5cdd664fd155c9e78fb46a4387af006c
Author: Emil Velikov
Date: Thu Oct 5 11:19:05 2017 +0100
configure.ac: rework llvm libs handling for 3.9+
I am getting 00s of
/mesa/src/amd/common/. undefined reference to LLVM..
Using git llvm have tried with -DLLVM_AP
On 13 October 2017 at 19:44, Jason Ekstrand wrote:
> ssize_t is a GNU extension and is not available on Windows or MacOS.
Not to argue against the patch in any way, but ssize_t is POSIX.
___
mesa-dev mailing list
mesa-dev@lists.freedesktop.org
https://l
On Sat, Oct 14, 2017 at 1:36 AM, Dylan Baker wrote:
> I'm not sure about this approach, we would need a way to add depends to meson,
> but I'm also worried that calling make adds another dependency that could be
> problematic for windows, and I really don't like the idea of having a
> half-and-hal
Signed-off-by: Rob Clark
---
v1: actually build freedreno
v2: corrected installed ${driver}_dri.so name
meson.build | 6 +
meson_options.txt| 2 +-
src/gallium/drivers/freedreno/meson.build| 221 +++
src
Signed-off-by: Rob Clark
---
As before, applies on top of of Dylan's wip/meson-radeonsi branch. But
now it actually builds freedreno (and fixes some mistakes in v1 that I
hadn't noticed because freedreno wasn't actually getting added to the
build)
meson.build |
Similar to 848da662224326ccfbe6647bc82f4f89ca22c762, pass an arg to
ir3_nir_trig.py to add to python path, rather than using $PYTHONPATH,
to prep for meson build support.
Signed-off-by: Rob Clark
---
src/gallium/drivers/freedreno/Makefile.am | 2 +-
src/gallium/drivers/freedreno/ir3/ir3
The new A1B5G5R5_UNORM, X1B5G5R5_UNORM formats were added in the
wrong place in commit ef874ee450b18e "gallium: Add support for 5551
with the 1-bit field in the low bit."
---
src/gallium/drivers/svga/svga_format.c | 4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)
diff --git a/src/gallium/
Signed-off-by: Rob Clark
---
This applies on top of Dylan's wip/meson-radeonsi branch
meson.build | 6 +
meson_options.txt| 2 +-
src/gallium/drivers/freedreno/meson.build| 212 +++
src/gallium/meson.bui
On 14/10/17 10:02, Pohjolainen, Topi wrote:
> On Thu, Oct 12, 2017 at 08:38:08PM +0200, Jose Maria Casanova Crespo wrote:
>> We enable the use of 16-bit values in push constants
>> modifying the assign_constant_locations function to work
>> with 16-bit types.
>>
>> The API to access buffers in Vu
On 14/10/17 09:55, Pohjolainen, Topi wrote:
> On Thu, Oct 12, 2017 at 08:38:03PM +0200, Jose Maria Casanova Crespo wrote:
>> From: Alejandro Piñeiro
>>
>> Conversions to 16-bit need having aligment between the 16-bit
>> and 32-bit types. So the conversion operations unpack 16-bit types
>> to wit
On 14/10/17 09:49, Pohjolainen, Topi wrote:
> On Thu, Oct 12, 2017 at 08:38:05PM +0200, Jose Maria Casanova Crespo wrote:
>> From: Alejandro Piñeiro
>>
>> Although it is possible to emit them directly as AND/OR on brw_fs_nir,
>> having a specific opcode makes it easier to remove duplicate settin
2017-10-12 16:07 GMT+02:00 Lucas Stach :
> Draw operations should not use the TS if the TS buffer content is invalid,
> as this leads to wrong rendering or even GPU hangs. As the TS valid status
> can change between draws (clear operations changing it to valid, blits using
> the RS to the color or
2017-10-12 16:07 GMT+02:00 Lucas Stach :
> When not all of the vertex attributes are actually used in the shader,
> we end up with some inputs without an assigned reg. Those are marked
> as invalid and must be skipped when assigning the inputs, as those would
> overwrite other valid inputs otherwis
On Thu, Oct 12, 2017 at 04:07:48PM +0200, Lucas Stach wrote:
> Draw operations should not use the TS if the TS buffer content is invalid,
> as this leads to wrong rendering or even GPU hangs. As the TS valid status
> can change between draws (clear operations changing it to valid,
> blits using th
On Thu, Oct 12, 2017 at 04:07:47PM +0200, Lucas Stach wrote:
> When not all of the vertex attributes are actually used in the shader,
> we end up with some inputs without an assigned reg. Those are marked
> as invalid and must be skipped when assigning the inputs, as those would
> overwrite other v
Hi Emil,
regarding this patch, I now think we can indeed drop the second part
(the one that adds the g++4.4 test), because it would only check
whether one sets the -std=c++11 flag somewhere accidently. Checking
that no c++11 code makes it into core mesa is already happening as long
as travis uses
On Thu, Oct 12, 2017 at 08:38:08PM +0200, Jose Maria Casanova Crespo wrote:
> We enable the use of 16-bit values in push constants
> modifying the assign_constant_locations function to work
> with 16-bit types.
>
> The API to access buffers in Vulkan use multiples of 4-byte for
> offsets and sizes
On Thu, Oct 12, 2017 at 08:38:03PM +0200, Jose Maria Casanova Crespo wrote:
> From: Alejandro Piñeiro
>
> Conversions to 16-bit need having aligment between the 16-bit
> and 32-bit types. So the conversion operations unpack 16-bit types
> to with an stride=2 and then applies a MOV with the conver
On Thu, Oct 12, 2017 at 08:38:05PM +0200, Jose Maria Casanova Crespo wrote:
> From: Alejandro Piñeiro
>
> Although it is possible to emit them directly as AND/OR on brw_fs_nir,
> having a specific opcode makes it easier to remove duplicate settings
> later.
>
> v2: (Curro)
> - Set thread contr
34 matches
Mail list logo
