Sandybridge requires the post-sync non-zero workaround in a ton of
places, and if you ever miss one, the GPU usually hangs.
Currently, we try to track exactly when a workaround flush is
necessary (via the brw->batch.need_workaround_flush flag). This is
tricky to get right, and we've botched it se
On Fri, Jan 9, 2015 at 10:33 PM, Kenneth Graunke wrote:
> On Friday, January 09, 2015 09:31:35 PM Matt Turner wrote:
>> On Fri, Jan 9, 2015 at 8:46 PM, Kenneth Graunke
>> wrote:
>> > When doing repclears, we only want to use the SIMD16 program, not the
>> > SIMD8 one. Kristian added this to the
On Friday, January 09, 2015 09:31:35 PM Matt Turner wrote:
> On Fri, Jan 9, 2015 at 8:46 PM, Kenneth Graunke wrote:
> > When doing repclears, we only want to use the SIMD16 program, not the
> > SIMD8 one. Kristian added this to the Gen7+ code, but apparently we
> > missed it in the Gen6 code. Th
On Fri, Jan 9, 2015 at 8:24 PM, Connor Abbott wrote:
> On Fri, Jan 9, 2015 at 11:04 PM, Jason Ekstrand
> wrote:
> > This solves a number of problems. First is the ability to change the
> > number of sources that a texture instruction has. Second, it solves the
> > delema that may occur if a te
On Fri, Jan 9, 2015 at 8:46 PM, Kenneth Graunke wrote:
> When doing repclears, we only want to use the SIMD16 program, not the
> SIMD8 one. Kristian added this to the Gen7+ code, but apparently we
> missed it in the Gen6 code. This patch copies that code over.
>
> Approximately doubles the perfo
When doing repclears, we only want to use the SIMD16 program, not the
SIMD8 one. Kristian added this to the Gen7+ code, but apparently we
missed it in the Gen6 code. This patch copies that code over.
Approximately doubles the performance in a clear microbenchmark from
mesa-demos (clearspd -width
On Tue, Jan 6, 2015 at 1:04 AM, Jason Ekstrand wrote:
>
>
> On Sun, Jan 4, 2015 at 9:15 PM, Connor Abbott wrote:
>>
>> This is a general question for the interpolation support:
>>
>> Why are we using the variable-based intrinsics directly, instead of
>> lowering it to something index-based in the
On Fri, Jan 9, 2015 at 11:04 PM, Jason Ekstrand wrote:
> This solves a number of problems. First is the ability to change the
> number of sources that a texture instruction has. Second, it solves the
> delema that may occur if a texture instruction has more than 4 sources.
> ---
> src/glsl/nir/
Ok, with the assert removed,
Reviewed-by: Connor Abbott
On Fri, Jan 9, 2015 at 11:16 PM, Jason Ekstrand wrote:
>
>
> On Fri, Jan 9, 2015 at 8:04 PM, Jason Ekstrand wrote:
>>
>> This solves a number of problems. First is the ability to change the
>> number of sources that a texture instruction
On Fri, Jan 9, 2015 at 8:04 PM, Jason Ekstrand wrote:
> This solves a number of problems. First is the ability to change the
> number of sources that a texture instruction has. Second, it solves the
> delema that may occur if a texture instruction has more than 4 sources.
> ---
> src/glsl/nir/
This solves a number of problems. First is the ability to change the
number of sources that a texture instruction has. Second, it solves the
delema that may occur if a texture instruction has more than 4 sources.
---
src/glsl/nir/glsl_to_nir.cpp | 36
On Fri, Jan 9, 2015 at 4:38 PM, Connor Abbott wrote:
> + case nir_intrinsic_copy_var:
> + unreachable("There should be no copies whatsoever at this
> point");
> + break;
> >>>
> >>>
> >>> Are you sure about this? My impression is that lower_variables will
From: Mike Mason
Changes the initial internal format of a render buffer
to GL_RGBA4. This fixes a failure in the following
DrawElements test:
dEQP-GLES3.functional.state_query.rbo.renderbuffer_internal_format
---
src/mesa/main/renderbuffer.c | 6 +-
1 file changed, 5 insertions(+), 1 deleti
https://bugs.freedesktop.org/show_bug.cgi?id=87886
--- Comment #17 from Stéphane Travostino ---
Status as of Linux 3.19-rc3, mesa HEAD e28f9d0
Resolution 1280x800 out of native 1600x900, all graphic detail at minimum.
Radeon: min fps 20, avg fps 65, max fps 95
In game FPS averages 65 FPS, wit
+ case nir_intrinsic_copy_var:
+ unreachable("There should be no copies whatsoever at this
point");
+ break;
>>>
>>>
>>> Are you sure about this? My impression is that lower_variables will lower
>>> copies involving things that aren't indirectly referenced,
https://bugs.freedesktop.org/show_bug.cgi?id=88079
--- Comment #2 from Mike Mason ---
I submitted a similar patch for review yesterday:
http://lists.freedesktop.org/archives/mesa-dev/2015-January/074092.html
then submitted v2 today after Matt Turner's review:
http://lists.freedesktop.org/archi
On Tue, Jan 6, 2015 at 5:05 PM, Jason Ekstrand wrote:
>
>
> On Sun, Jan 4, 2015 at 7:52 PM, Connor Abbott wrote:
>
>>
>>
>> On Tue, Dec 16, 2014 at 1:11 AM, Jason Ekstrand
>> wrote:
>>
>>> ---
>>> src/glsl/Makefile.sources | 1 +
>>> src/glsl/nir/nir.h |
On Fri, Jan 9, 2015 at 5:41 PM, Jason Ekstrand wrote:
>
>
> On Fri, Jan 9, 2015 at 11:34 AM, Connor Abbott wrote:
>>
>> BTW, now that we're adding an extra possible source, do we need to
>> expand the size of the source array?
>
>
> I don't know that we've added anything that wasn't there before.
This meta path, designed for use with PBO's, creates a temporary texture
out of the PBO and uses BlitFramebuffers to do the actual texture upload.
v2 Jason Ekstrand :
- Add support for handling simple packing options
---
src/mesa/Makefile.sources | 1 +
src/mesa/drivers/commo
On Fri, Jan 9, 2015 at 11:34 AM, Connor Abbott wrote:
> BTW, now that we're adding an extra possible source, do we need to
> expand the size of the source array?
>
I don't know that we've added anything that wasn't there before. Then
again, it's not documented how you came to the number 4. Als
On Fri, Jan 9, 2015 at 11:39 AM, Neil Roberts wrote:
> Jason Ekstrand writes:
>
> > This improves texture upload performance on the PBO upload test available
> > at http://www.songho.ca/opengl/gl_pbo.html by 80% for the non-PBO case
> (due
> > to avoiding a buffer stall) and 500% for the PBO cas
From: Mike Mason
Removes commit 7894278 changes and moves fix to _mesa_GetInternalformativ().
The original commit enabled the GL_RGB and GL_RGBA unsized internal formats
as valid for render buffers in GLES3, but this is incorrect. They should
have only been enabled for GetInternalformativ()
Bugz
As far as I know, no.
Marek
On Fri, Jan 9, 2015 at 7:42 AM, Matt Turner wrote:
> On Thu, Jan 8, 2015 at 5:48 AM, Marek Olšák wrote:
>> Please avoid TGSI_SAT_MINUS_PLUS_ONE. Nobody uses it, hardly anybody
>> implements it.
>
> I'm curious -- does any hardware implement it?
__
Jason Ekstrand writes:
> This improves texture upload performance on the PBO upload test available
> at http://www.songho.ca/opengl/gl_pbo.html by 80% for the non-PBO case (due
> to avoiding a buffer stall) and 500% for the PBO case.
Just for reference, if I run this branch against the little te
BTW, now that we're adding an extra possible source, do we need to
expand the size of the source array?
Anyways,
Reviewed-by: Connor Abbott
On Wed, Jan 7, 2015 at 9:03 PM, Jason Ekstrand wrote:
> This helps a lot with things like lowering passes that may need to add
> sources.
> ---
> src/glsl
Reviewed-by: Connor Abbott
On Wed, Jan 7, 2015 at 9:03 PM, Jason Ekstrand wrote:
> In particular, we rename nir_tex_src_sampler_index to _sampler_offset and
> add a sampler_array_size field to nir_tex_instr. This way we can pass the
> size of sampler arrays through to backends even after removi
On Tue, Jan 6, 2015 at 7:34 PM, Jason Ekstrand wrote:
> Additional description was added to a variety of places. Also, we no
> longer use the term "leaf" to describe fully-qualified direct derefs.
> Instead, we simply use the term "direct" or spell it out completely.
> ---
> src/glsl/nir/nir_low
On Sun, Jan 4, 2015 at 4:01 PM, Connor Abbott wrote:
[...]
>> static bool
>> deref_may_be_aliased(nir_deref_var *deref,
>> struct lower_variables_state *state)
>> {
>>nir_deref_var var_deref = *deref;
>>var_deref.deref.child = NULL;
>>struct deref_node *node = get_
Whoops... I was looking over the actual file (instead of the patch),
and there's a comment above (at least in lower_variables) about magic
numbers that needs to be deleted or rewritten.
On Fri, Jan 9, 2015 at 1:30 PM, Connor Abbott wrote:
> As you mentioned, you should add something to the commit
With Matt's suggested commit message,
Reviewed-by: Connor Abbott
On Tue, Jan 6, 2015 at 7:34 PM, Jason Ekstrand wrote:
> ---
> src/glsl/nir/nir.h | 4
> src/glsl/nir/nir_validate.c | 7 +++
> 2 files changed, 11 insertions(+)
>
> diff --git a/src/glsl/nir/nir.h b/src/glsl/nir
As you mentioned, you should add something to the commit message
saying you're making it use a loop too.
On Tue, Jan 6, 2015 at 7:34 PM, Jason Ekstrand wrote:
> ---
> src/glsl/nir/nir_lower_locals_to_regs.c | 65
> src/glsl/nir/nir_lower_variables.c | 101
> ++
Daniel Vetter writes:
> Oh, I guess my earlier mail was too late. One issue still is picking
> the numbers, since you seem to assume here that ver >= 2 means the
> stuff actually works. But like Ken said the cmd parser in upstream
> isn't really enabled yet.
The patch only enables the predicate
https://bugs.freedesktop.org/show_bug.cgi?id=88079
--- Comment #1 from Eduardo Lima Mitev ---
Sent a patch for review:
http://lists.freedesktop.org/archives/mesa-dev/2015-January/074122.html.
It just allows GL_RGB and GL_RGBA when querying state through
glGetInternalformativ(). It now passes the
Previous patch 78942787170615c9333810cf3a4819a13c9eb8e8 globally introduced
GL_RGB and GL_RGBA as valid internal formats on GLES 3.0.4, as the spec implies
(page 112). However, page 205 of the spec states that
for RenderbufferStorageMultisample, internal format must be a sized internal
format. For
On Jan 9, 2015 6:33 AM, "Neil Roberts" wrote:
>
> This patch looks really good. I have some comments below.
>
> Jason Ekstrand writes:
>
> > This meta path, designed for use with PBO's, creates a temporary texture
> > out of the PBO and uses BlitFramebuffers to do the actual texture
upload.
> > -
On Jan 9, 2015 2:43 AM, "Samuel Iglesias Gonsalvez"
wrote:
>
> This will be used to unify code in pack.c.
>
> v2:
> - Modify pack_int_*() function generator to use c.datatype() and
> f.datatype()
>
> v3:
> - Only autogenerate pack_int_*() functions for non-normalized integer
> formats.
>
> Sig
On Jan 9, 2015 2:40 AM, "Samuel Iglesias Gonsálvez"
wrote:
>
> On Thursday, January 08, 2015 08:20:55 AM Iago Toral Quiroga wrote:
> > From: Samuel Iglesias Gonsalvez
> >
> > We will use this later on to handle uint conversion scenarios in a
master
> > convert function.
> >
> > v2:
> > - Modify p
On Friday, January 09, 2015 07:14:37 AM Jason Ekstrand wrote:
> On Jan 8, 2015 11:54 PM, "Samuel Iglesias Gonsálvez"
>
> wrote:
> > On Thursday, January 08, 2015 10:13:11 AM Jason Ekstrand wrote:
> > > On Wed, Jan 7, 2015 at 11:20 PM, Iago Toral Quiroga
> > >
> > > wrote:
> > > > From: Samuel I
On Jan 8, 2015 11:54 PM, "Samuel Iglesias Gonsálvez"
wrote:
>
> On Thursday, January 08, 2015 10:13:11 AM Jason Ekstrand wrote:
> > On Wed, Jan 7, 2015 at 11:20 PM, Iago Toral Quiroga
> >
> > wrote:
> > > From: Samuel Iglesias Gonsalvez
> > >
> > > This will be used to unify code in pack.c.
> >
This patch looks really good. I have some comments below.
Jason Ekstrand writes:
> This meta path, designed for use with PBO's, creates a temporary texture
> out of the PBO and uses BlitFramebuffers to do the actual texture upload.
> ---
> src/mesa/Makefile.sources | 1 +
>
On Thu, Jan 8, 2015 at 10:34 PM, Dave Airlie wrote:
> From: Dave Airlie
>
> After much searching and empricial testing, and reading of
> things I've no justifcation for this fix, other than it really
> appears this is what the hw is doing or close enough.
>
> It makes sense that each entry in the
On 01/09/2015 12:07 PM, Kenneth Graunke wrote:
On Friday, January 09, 2015 11:25:18 AM Tapani Pälli wrote:
Patch enables ES2 extension that utilizes existing ES3 functionality.
Changes make all the subtests to run and pass in WebGL conformance
test 'webgl-draw-buffers' when running Chrome on
On Thu, 2015-01-08 at 13:54 +0200, Aras Pranckevicius wrote:
> I see precision qualifiers being parsed and stored in the AST,
> but I
> don't see where this information is passed to the IR:
> ir_variable or
> glsl_type don't have this info, in fact,
>
This will be used to unify code in pack.c.
v2:
- Modify pack_int_*() function generator to use c.datatype() and
f.datatype()
v3:
- Only autogenerate pack_int_*() functions for non-normalized integer
formats.
Signed-off-by: Samuel Iglesias Gonsalvez
---
src/mesa/main/format_pack.h | 3 ++
On Thursday, January 08, 2015 08:20:55 AM Iago Toral Quiroga wrote:
> From: Samuel Iglesias Gonsalvez
>
> We will use this later on to handle uint conversion scenarios in a master
> convert function.
>
> v2:
> - Modify pack_uint_*() function generation to use c.datatype() and
> f.datatype().
>
On Friday, January 09, 2015 11:25:18 AM Tapani Pälli wrote:
> Patch enables ES2 extension that utilizes existing ES3 functionality.
>
> Changes make all the subtests to run and pass in WebGL conformance
> test 'webgl-draw-buffers' when running Chrome on OpenGL ES, also
> Piglit test 'draw_buffers_
On Thursday, January 08, 2015 10:03:01 AM Jason Ekstrand wrote:
> On Wed, Jan 7, 2015 at 11:20 PM, Iago Toral Quiroga
>
> wrote:
> > From: Jason Ekstrand
> >
> > We were auto-generating it before. The problem was that the
> > autogeneration
> > tool we were using was called "copy, paste, and e
Patch enables ES2 extension that utilizes existing ES3 functionality.
Changes make all the subtests to run and pass in WebGL conformance
test 'webgl-draw-buffers' when running Chrome on OpenGL ES, also
Piglit test 'draw_buffers_gles2' passes.
v2: remove unused boolean (Ilia Mirkin)
v3: proper err
On Thursday, January 08, 2015 01:31:56 PM Jason Ekstrand wrote:
> On Thu, Jan 8, 2015 at 12:02 PM, Anuj Phogat wrote:
> > On Wed, Jan 7, 2015 at 11:20 PM, Iago Toral Quiroga
> >
> > wrote:
> > > From: Jason Ekstrand
> > >
> > > Aparently, the packing/unpacking functions for these formats have
49 matches
Mail list logo
