based on 9f65c0a26 "cl: Add tests for 24-bit div / rem optimization" by Matt
Arsenault
Signed-off-by: Jan Vesely
---
tests/cl/program/execute/scalar-arithmetic-uint.cl | 1657
1 file changed, 1657 insertions(+)
diff --git
Yes, it passes here as well but I believe the test allows a lot of
variance - that was just visual inspection. (I believe if you'd just
implement that in the driver as ignore alpha-to-one if alpha-to-coverage
is also active only then you'd get a perfect match.)
Roland
Am 04.08.2017 um 19:28
Brian Paul writes:
> The glean fbo test exercises rendering to 1D, 2D, 3D and cube images
> attached to FBOs. For the 1D texture case, the rendering is done with
> a line, rather than a polygon. Since the 1D image is one pixel high,
> the slightest mis-rasterization of the
It passes here for me w/ NVIDIA's driver but I'll hold off on this until
I can take a closer look.
-Brina
On 08/04/2017 11:11 AM, Roland Scheidegger wrote:
I think should mention that the reference probably isn't quite right (if
I try it here, the reference is the same with or without
I think should mention that the reference probably isn't quite right (if
I try it here, the reference is the same with or without alpha-to-one,
but with nvidia with alpha-to-one it's a bit different (darker) than
without). I think it's because the reference drawing doesn't really take
into account
Am 04.08.2017 um 16:43 schrieb Brian Paul:
> On 08/04/2017 08:32 AM, srol...@vmware.com wrote:
>> From: Roland Scheidegger
>>
>> I believe the sole reason alpha-to-one exists at all is that it can be
>> used together with alpha-to-coverage (otherwise why not just output alpha
Alpha-to-one only makes sense when used with alpha-to-coverage. This
patch adds a flag to the alpha-to-coverage test to exercise alpha-to-one.
This will fail with drivers which emulate alpha-to-one in the FS.
---
tests/all.py | 4
On Fri, Aug 4, 2017 at 12:40 PM, Brian Paul wrote:
> On 08/04/2017 10:13 AM, Ilia Mirkin wrote:
>>
>> On Fri, Aug 4, 2017 at 11:36 AM, Brian Paul wrote:
>>>
>>> On 08/04/2017 09:30 AM, Brian Paul wrote:
On 08/04/2017 08:51 AM, Roland
On 08/04/2017 10:13 AM, Ilia Mirkin wrote:
On Fri, Aug 4, 2017 at 11:36 AM, Brian Paul wrote:
On 08/04/2017 09:30 AM, Brian Paul wrote:
On 08/04/2017 08:51 AM, Roland Scheidegger wrote:
Am 04.08.2017 um 16:43 schrieb Brian Paul:
On 08/04/2017 08:32 AM,
On Fri, Aug 4, 2017 at 11:36 AM, Brian Paul wrote:
> On 08/04/2017 09:30 AM, Brian Paul wrote:
>>
>> On 08/04/2017 08:51 AM, Roland Scheidegger wrote:
>>>
>>> Am 04.08.2017 um 16:43 schrieb Brian Paul:
On 08/04/2017 08:32 AM, srol...@vmware.com wrote:
>
>
On 08/04/2017 08:51 AM, Roland Scheidegger wrote:
Am 04.08.2017 um 16:43 schrieb Brian Paul:
On 08/04/2017 08:32 AM, srol...@vmware.com wrote:
From: Roland Scheidegger
I believe the sole reason alpha-to-one exists at all is that it can be
used together with
On 08/04/2017 09:30 AM, Brian Paul wrote:
On 08/04/2017 08:51 AM, Roland Scheidegger wrote:
Am 04.08.2017 um 16:43 schrieb Brian Paul:
On 08/04/2017 08:32 AM, srol...@vmware.com wrote:
From: Roland Scheidegger
I believe the sole reason alpha-to-one exists at all is that
On 08/04/2017 08:32 AM, srol...@vmware.com wrote:
From: Roland Scheidegger
I believe the sole reason alpha-to-one exists at all is that it can be
used together with alpha-to-coverage (otherwise why not just output alpha
1.0 in the fs in the first place...).
Therefore, it
I applied your patch and compiled it. There were compilation warnings
about passing GLubyte * to strncmp() and strcmp() which expects char *.
I've fixed that, adjusted the formatting and wrapped the lines to 78
columns. See attachemnt. If it looks OK to you, Sanda, I'll push it.
-Brian
Explain why these tests don't need to enable GL_SAMPLE_SHADING_ARB.
---
.../arb_sample_shading/execution/builtin-gl-sample-id.cpp | 11 +++
.../execution/builtin-gl-sample-position.cpp | 11 +++
2 files changed, 22 insertions(+)
diff --git
On 01/08/17 15:06, Arkadiusz Hiler wrote:
Currently, if a test from provided testlist fails to be discovered by
the framework, piglit blows up with an exception.
This is both good - for consistency/early errors - and bad - for
handling some CI/automation scenarios (e.g autobisecting the tests).
Hi,
thanks for the answer. Then, can anyone review this please?
It would be a shame if this was not pushed, because a lot of people spent
their time to comment this patch and WGL_EXT_swap_control is in the
extension string on Windows Nvidia, AMD and Intel drivers.
Regards,
Sandra
On Wed, Jul 26,
This test verifies that the register merge step keeps a register alive
for the whole loop if at the beginning of this loop at least one component
is written unconditionally, and later another component is written
conditionally within the same loop, and the last read to the register also
happens
On 08/04/2017 03:43 AM, Timothy Arceri wrote:
This only tests for errors when is 0, but its a start.
---
tests/all.py | 6 +
tests/spec/CMakeLists.txt | 1 +
tests/spec/ext_memory_object/CMakeLists.gl.txt | 14 ++
On 04/08/17 18:56, Samuel Pitoiset wrote:
1/2 got lost?
Yeah its in moderation (to big).
See https://github.com/tarceri/Piglit/compare/memobj
On 08/04/2017 03:43 AM, Timothy Arceri wrote:
This only tests for errors when is 0, but its a start.
---
tests/all.py
1/2 got lost?
On 08/04/2017 03:43 AM, Timothy Arceri wrote:
This only tests for errors when is 0, but its a start.
---
tests/all.py | 6 +
tests/spec/CMakeLists.txt | 1 +
tests/spec/ext_memory_object/CMakeLists.gl.txt | 14 ++
21 matches
Mail list logo