On Sat, Aug 15, 2015 at 10:20 PM, Serge Martin (EdB)
edb+m...@sigluy.net wrote:
This looks like what I've send a few weeks ago [0]. It's seems we should come
with something a little different as explain by curro in [1]
Ok, didn't notice that :)
Sorry for the duplicate.
On 8/14/15, Zoltan Gilian zoltan.gil...@gmail.com wrote:
Remove unnecessary use of cb_target_mask (Marek).
This is an error, it should be Remove unnecessary use of
util_range_add. I'll fix this in a moment.
___
mesa-dev mailing list
On 8/11/15, Marek Olšák mar...@gmail.com wrote:
Do compute sampler states need RADEON_CP_PACKET3_COMPUTE_MODE as well?
Oops, I've forgotten to add it. My mistake.
I'm not sure if needed (I have a working test which uses linear
filtering), but the SET_SAMPLER packet too has a header with the
Hi!
I've noticed that clover::kernel::image_wr_argument::unbind destroys
the surface. What kind of behaviour is expected from the driver to
allow later access to the resource (e.g. by clEnqueueReadImage)?
Zoltan
___
mesa-dev mailing list
thanks, this fixes the build for me.
Should I commit this marking you as reviewer?
On Wed, Aug 5, 2015 at 5:21 AM, Jan Vesely jan.ves...@rutgers.edu wrote:
On Tue, 2015-08-04 at 22:40 +0200, Zoltan Gilian wrote:
There is no MDOperand in llvm 3.5.
thanks, this fixes the build for me.
jan
:
Zoltán Gilián zoltan.gil...@gmail.com writes:
thanks, this fixes the build for me.
Should I commit this marking you as reviewer?
fixes the build for me hardly qualifies as a Tested-by. Also I've
noticed that you've pushed a few patches [1][2][3][4] without
Reviewed-by tag. Please stop
Meanwhile I've obtained commit permission, so committed as d2cd2c6.
On Fri, Jul 31, 2015 at 4:06 PM, Zoltán Gilián zoltan.gil...@gmail.com wrote:
Could you please commit this?
On Mon, Jul 27, 2015 at 1:20 PM, Francisco Jerez curroje...@riseup.net
wrote:
Zoltan Gilian zoltan.gil...@gmail.com
Committed as 9ef5b7a.
On Fri, Jul 31, 2015 at 4:07 PM, Zoltán Gilián zoltan.gil...@gmail.com wrote:
Could you please commit this?
On Mon, Jul 27, 2015 at 1:28 PM, Francisco Jerez curroje...@riseup.net
wrote:
Zoltan Gilian zoltan.gil...@gmail.com writes:
Read-only and write-only image
Fixed it. Committed as aa46fba.
On Mon, Aug 3, 2015 at 1:36 PM, Zoltán Gilián zoltan.gil...@gmail.com wrote:
Any reason you didn't fix this other branch to do the same? Or maybe
just init it to one after the if?
I've only observed problems with images if that line is missing (the
surface
Committed as be3622d.
On Fri, Jul 31, 2015 at 4:09 PM, Francisco Jerez curroje...@riseup.net wrote:
Zoltán Gilián zoltan.gil...@gmail.com writes:
Could you please commit this? I don't have permissions.
Sure, I'll put them into my queue.
On Fri, Jul 31, 2015 at 3:55 PM, Francisco Jerez
Any reason you didn't fix this other branch to do the same? Or maybe
just init it to one after the if?
I've only observed problems with images if that line is missing (the
surface is rejected because of 0 buffer size), but it makes sense to
set the array_size outside the if.
On Mon, Aug 3,
Committed as 44e90f2.
On Fri, Jul 31, 2015 at 12:11 PM, Marek Olšák mar...@gmail.com wrote:
Reviewed-by: Marek Olšák marek.ol...@amd.com
Marek
On Fri, Jul 31, 2015 at 11:59 AM, Zoltan Gilian zoltan.gil...@gmail.com
wrote:
To circumvent a problem occuring when LINEAR_ALIGNED array mode is
Could you please commit this? I don't have permissions.
On Fri, Jul 31, 2015 at 3:55 PM, Francisco Jerez curroje...@riseup.net wrote:
Zoltan Gilian zoltan.gil...@gmail.com writes:
---
src/gallium/state_trackers/clover/core/kernel.cpp | 15 +++
1 file changed, 15 insertions(+)
Could you please commit this?
On Mon, Jul 27, 2015 at 1:20 PM, Francisco Jerez curroje...@riseup.net wrote:
Zoltan Gilian zoltan.gil...@gmail.com writes:
---
.../state_trackers/clover/llvm/invocation.cpp | 28
--
1 file changed, 15 insertions(+), 13 deletions(-)
Could you please commit this?
On Mon, Jul 27, 2015 at 1:28 PM, Francisco Jerez curroje...@riseup.net wrote:
Zoltan Gilian zoltan.gil...@gmail.com writes:
Read-only and write-only image arguments are recognized and
distinguished.
Attributes of the image arguments are passed to the kernel as
Also, does the same restriction apply to SI and newer GPUs?
I don't know. I can only test this on an evergreen (juniper) GPU.
On Fri, Jul 31, 2015 at 5:15 AM, Michel Dänzer mic...@daenzer.net wrote:
On 31.07.2015 07:06, Zoltan Gilian wrote:
To circumvent a problem occuring when LINEAR_ALIGNED
value to
the input vector breaks something I don't know about?
On Sun, Jul 26, 2015 at 11:37 PM, Zoltán Gilián zoltan.gil...@gmail.com wrote:
Why? Your module::argument::image_size and ::image_format cases don't
touch the explicit_arg iterator at all AFAICT, so it will be left
pointing one past
But samplers aren't necessary for the image attributes, so I'll break
this patch into two.
On Mon, Jul 27, 2015 at 10:25 AM, Zoltán Gilián zoltan.gil...@gmail.com wrote:
Hmmm... So you only need it as padding? Wouldn't it be possible for
you to declare samplers to be 0 bytes?
Maybe it can
auto img = dynamic_castimage_argument (**(explicit_arg - 1))
Ok, so it should be (explicit_arg - 2) for the image format, I
presume. This will be incorrect, however, if the targets that need
implicit argument for format metadata are indeed a strict superset of
the ones that need dimension, as
it be possible for
you to declare samplers to be 0 bytes?
Maybe it can be done by changing the type of the sampler arg from i32
to an empty struct. I'll have to try this, I don't know if it will
work.
On Sun, Jul 26, 2015 at 2:40 PM, Francisco Jerez curroje...@riseup.net wrote:
Zoltán Gilián
I have the suspicion
that it would simplify both the OpenCL front-end and compiler back-end
code if the image metadata was interleaved with images themselves.
In fact this complicates the back-end, since the location of each
argument following an image argument changes because of the metadata
This seems to be doing essentially the same thing as v1? Is it the
right patch?
The llvm pass was invoked in clover in v1. This patch relies on llvm
to perform that task (). What this patch does basically is that it
adds the image attributes to the end of the kernel input vector.
The commit
Hint: you'll need new
module::argument::semantic enums
I see. Reworked it a bit.
On Mon, Jul 6, 2015 at 1:13 PM, Francisco Jerez curroje...@riseup.net wrote:
Zoltán Gilián zoltan.gil...@gmail.com writes:
This seems to be doing essentially the same thing as v1? Is it the
right patch
there is no need to perform the llvm pass.
On Tue, Jun 9, 2015 at 11:09 AM, Zoltán Gilián zoltan.gil...@gmail.com
wrote:
Ok, thanks for the feedback.
On Mon, Jun 8, 2015 at 2:18 PM, Francisco Jerez curroje...@riseup.net
wrote:
Zoltan Gilian zoltan.gil...@gmail.com writes:
Image attributes
Ok, thanks for the feedback.
On Mon, Jun 8, 2015 at 2:18 PM, Francisco Jerez curroje...@riseup.net
wrote:
Zoltan Gilian zoltan.gil...@gmail.com writes:
Image attributes are passed to the kernel as hidden parameters after the
image attribute itself. An llvm pass replaces the getter builtins
Why keeping the assertion commented out?
Because to test the image support functionality I'm working on, I have to
create an image and pass it to a kernel, which triggers the assertion ATM.
Can you just fix evergreen_set_cs_sampler_view so that it supports
views==NULL?
Fair point, will do
26 matches
Mail list logo
