On 29/03/2015 04:02, Marek Olšák wrote:
On Sat, Mar 28, 2015 at 9:43 PM, Martin Peres wrote:
On 22/03/2015 17:35, Samuel Pitoiset wrote:
From: Christoph Bumiller
This is based on the original patch of Christoph Bumiller.
(source: http://people.freedesktop.org/~chrisbmr/perfmon.diff)
It wou
Hi thank you for the help, I will test
The only reason for trying to use the texture to wl_buffer directly was
just to get something working and then work back from there to see if it
was an application side issue or not.
Also just a fun experiment.
On Sun, Mar 29, 2015 at 6:04 AM, Jason Ekstran
On 29 March 2015 at 04:17, Kenneth Graunke wrote:
> On Sunday, March 29, 2015 12:14:50 AM Emil Velikov wrote:
>> On 28 March 2015 at 20:54, Emil Velikov wrote:
>> > From: Mauro Rossi
>> >
>> > Required by the i965 driver.
>> >
>> > Cc: "10.5"
>> > [Emil Velikov: Split from a larger commit]
>> >
On 28 March 2015 at 23:51, Bernd Kuhls wrote:
> Hi,
>
> Emil Velikov wrote in news:1427132964-21468-2-
> git-send-email-emil.l.veli...@gmail.com:
>
>> In case of using a distribution tarball (or a dirty git tree) one can
>> have the generated sources locally. Make configure.ac error out
>> otherw
This will allow us to finally remove python from the build time
dependencies list. Considering that you're building from a release
tarball of course :-)
Cc: Bernd Kuhls
Reported-by: Bernd Kuhls
Cc: "10.5"
Signed-off-by: Emil Velikov
---
src/mesa/drivers/dri/common/xmlpool/Makefile.am | 9
... by folding it into CLEANFILES. Don't worry about $(LANG) as it is
essentially the first folder of $(POS). With the latter already handled.
Signed-off-by: Emil Velikov
---
src/mesa/drivers/dri/common/xmlpool/Makefile.am | 10 --
1 file changed, 4 insertions(+), 6 deletions(-)
diff --
On 03/28/2015 09:43 PM, Martin Peres wrote:
On 22/03/2015 17:35, Samuel Pitoiset wrote:
From: Christoph Bumiller
This is based on the original patch of Christoph Bumiller.
(source: http://people.freedesktop.org/~chrisbmr/perfmon.diff)
It would be nice if you could add "v2: Samuel Pitoiset"
On 03/29/2015 11:13 AM, Martin Peres wrote:
On 29/03/2015 04:02, Marek Olšák wrote:
On Sat, Mar 28, 2015 at 9:43 PM, Martin Peres
wrote:
On 22/03/2015 17:35, Samuel Pitoiset wrote:
From: Christoph Bumiller
This is based on the original patch of Christoph Bumiller.
(source: http://people.f
On 29/03/2015 17:57, Samuel Pitoiset wrote:
On 03/29/2015 11:13 AM, Martin Peres wrote:
On 29/03/2015 04:02, Marek Olšák wrote:
On Sat, Mar 28, 2015 at 9:43 PM, Martin Peres
wrote:
On 22/03/2015 17:35, Samuel Pitoiset wrote:
From: Christoph Bumiller
This is based on the original patch of
On 29/03/2015 17:56, Samuel Pitoiset wrote:
On 03/28/2015 09:43 PM, Martin Peres wrote:
On 22/03/2015 17:35, Samuel Pitoiset wrote:
From: Christoph Bumiller
This is based on the original patch of Christoph Bumiller.
(source: http://people.freedesktop.org/~chrisbmr/perfmon.diff)
It would be
On 28/03/15 19:23, Emil Velikov wrote:
> Mesa 10.5.2 is now available. This release addresses bugs in the common glsl
> code-base, the libGL and glapi libraries, and the dri modules. The tarball no
> longer contains hardlinks and has all the haiku files. With this release one
> can build mesa wi
The same rationale applies here as for the hash table.
Power of two size should give better performance,
and using the algorithm hash = sh + i/2 + i*i/2
should result in only distinct hash values when hitting collisions.
Should give a performance increase as we can do bitmasking instead
of a modulo
Since a pointer is basically just an int we can use integer hashing.
This one is taken from https://gist.github.com/badboy/6267743
A google search seems to suggest this is a common and good algorithm.
Since it operates 32 bits at a time it is really effective.
assert that we are hashing 32bit align
Here's the latest round of fixup on the hash-table patches.
I think I've gotten all the review feedback incorporated now.
These patches give a nice little boost, indicated in each commit.
As a side effect of upping the minimum size of the table and set
there is now also less spamming of rzalloc and
This should give better cache locality, less memory consumption,
less code, and should also be faster since we avoid a modulo operation.
Also change table size to be power of two.
This gives better performance as we can do bitmasking instead of
modulo operations for fitting the hash in the address
On Sun, Mar 29, 2015 at 11:05 AM, Thomas Helland
wrote:
> Since a pointer is basically just an int we can use integer hashing.
> This one is taken from https://gist.github.com/badboy/6267743
> A google search seems to suggest this is a common and good algorithm.
> Since it operates 32 bits at a ti
2015-03-29 20:15 GMT+02:00 Matt Turner :
> On Sun, Mar 29, 2015 at 11:05 AM, Thomas Helland
> wrote:
>> Since a pointer is basically just an int we can use integer hashing.
>> This one is taken from https://gist.github.com/badboy/6267743
>> A google search seems to suggest this is a common and goo
On 2015-03-29 11:05:40, Thomas Helland wrote:
> Since a pointer is basically just an int we can use integer hashing.
> This one is taken from https://gist.github.com/badboy/6267743
There doesn't seem to be a license associated with this code, nor is
it indicated that it is public domain code.
-Jo
(Forgot to send to list)
That is indeed an issue.
I found the original article on the wayback machine and it
doesn't state anything particular wrt license.
However, it seems to be used in a LOT of projects.
(javascript, chromium, hiphop-php, kde, +++)
I found this webpage that gives some more in
I just remembered the existence of this series so
I went ahead and tried it on top of today's master.
With NIR enabled there is no benefit at all, didn't try without NIR.
I've marked it as rejected in patchwork, so it's not floating around in
there.
3. jan. 2015 20.21 skrev "Thomas Helland" :
> I
Does anyone have suggestions for good literature on compilers?
Since GPU's and CPU's are a bit different there are probably
books that are better suited than others for GPUs?
I have what is probably Norway's biggest library on the subject to rent
books from, so I guess I should be able to find most
Since a pointer is basically just an int we can use integer hashing.
This implementation is found on Bob Jenkins' webpage on:
http://burtleburtle.net/bob/hash/integer.html
It states that this implementation is faster than any of his algorithms.
It also statest that the algorithm is public domain.
S
Add a test to exercise a worst case collision scenario
that may cause us to not be able to find an empty
slot in the table even though it is not full.
This hits the bug in my last revision of the series
converting the hash table to quadratic probing.
Signed-off-by: Thomas Helland
---
src/util/te
On 2015-03-29 13:28:02, Thomas Helland wrote:
> (Forgot to send to list)
>
> That is indeed an issue.
> I found the original article on the wayback machine and it
> doesn't state anything particular wrt license.
> However, it seems to be used in a LOT of projects.
> (javascript, chromium, hiphop-p
From: Roland Scheidegger
Just announce support for 4 components.
While here also increase the max/min texel offsets (the limit is completely
artificial, was chosen because that's what other hardware did, however there's
other drivers using larger limits).
Over a thousand little piglits skip->pass
From: Roland Scheidegger
This is quite trivial, essentially just follow all the same code you'd
use with linear min/mag (and no mip) filter, then just skip the filtering
after looking up the texels in favor of direct assignment of the right channel
to the result. (This is though not true for the
From: Roland Scheidegger
Luckily thanks to the revamped interface this is a lot less work now...
---
src/gallium/auxiliary/gallivm/lp_bld_sample.h | 18 +
src/gallium/auxiliary/gallivm/lp_bld_sample_soa.c | 31 +--
src/gallium/auxiliary/gallivm/lp_bld_tgsi_soa
https://bugs.freedesktop.org/show_bug.cgi?id=89818
Bug ID: 89818
Summary: WebGL Conformance
conformance/textures/texture-size-limit.html ->
OUT_OF_MEMORY
Product: Mesa
Version: unspecified
Hardware: Othe
On Wednesday, March 25, 2015 05:53:43 PM Ben Widawsky wrote:
> On Sun, Mar 22, 2015 at 06:49:15PM -0700, Jordan Justen wrote:
> > These fields will be used when emitting a send for the barrier function.
> >
> > Reference: IVB PRM Volume 4, Part 2, Section 1.1.1 Message Descriptor
> >
> > Signed-o
https://bugs.freedesktop.org/show_bug.cgi?id=89819
Bug ID: 89819
Summary: WebGL Conformance swrast failure in
conformance/uniforms/uniform-default-values.html
Product: Mesa
Version: unspecified
Hardware: Other
On Monday, March 23, 2015 09:56:52 AM Antia Puentes wrote:
> From GLSL 1.30.10, section 3.3 (Preprocessor):
> "#line line source-string-number ... After processing this directive
> (including its new-line), the implementation will behave as if it is
> compiling at ... source string number source-st
On Sun, Mar 29, 2015 at 2:05 PM, Thomas Helland
wrote:
> Here's the latest round of fixup on the hash-table patches.
> I think I've gotten all the review feedback incorporated now.
> These patches give a nice little boost, indicated in each commit.
> As a side effect of upping the minimum size of
On Monday, March 23, 2015 09:56:29 AM Antia Puentes wrote:
> From GLSL 3.30 and GLSL ES 1.00 on, after processing the line
> directive (including its new-line), the implementation should
> behave as if it is compiling at the line number passed as
> argument. In previous versions, it behaved as if c
On Friday, March 20, 2015 05:28:55 PM Jordan Justen wrote:
> git://people.freedesktop.org/~jljusten/mesa i965-pipelines-v3
>
> v2:
> * Rename brw->atoms[] to render_atoms
> * Add brw->compute_atoms[]
>* Replace brw_pipeline_first_atom with brw_get_pipeline_atoms
>
> v3:
> * Avoid changing
On Friday, March 20, 2015 11:28:28 PM Carl Worth wrote:
> On Fri, Mar 20 2015, Chris Forbes wrote:
> > I think that having both the existing `struct brw_vs_compile` and a
> > function with the same name is going to cause confusion. (same with
> > the other non-fs stages)
>
> In an earlier version
On Sun, Mar 29, 2015 at 5:54 PM, Thomas Helland
wrote:
> Does anyone have suggestions for good literature on compilers?
> Since GPU's and CPU's are a bit different there are probably
> books that are better suited than others for GPUs?
> I have what is probably Norway's biggest library on the subj
On Friday, March 20, 2015 05:49:06 PM Carl Worth wrote:
> This commit splits portions of the existing brw_upload_vs_prog and
> brw_upload_gs_prog function into new brw_vs_populate_key and
> brw_gs_populate_key functions. This follows the same style as is
> already present for all other stages, (see
Emil Velikov writes:
> The header was added with commit 2a135c470e3(nir: Add an ALU op builder
> kind of like ir_builder.h) but did not made it into to the sources list,
> and its dependency of nir_builder_opcodes.h was missing.
>
> Fortunately it remained unused until resent commit faf6106c6f6(n
On Sun, Mar 29, 2015 at 9:51 PM, Connor Abbott wrote:
> On Sun, Mar 29, 2015 at 5:54 PM, Thomas Helland
> wrote:
>> Does anyone have suggestions for good literature on compilers?
>> Since GPU's and CPU's are a bit different there are probably
>> books that are better suited than others for GPUs?
https://bugs.freedesktop.org/show_bug.cgi?id=89586
Dan Sebald changed:
What|Removed |Added
Attachment #114342|0 |1
is obsolete|
https://bugs.freedesktop.org/show_bug.cgi?id=89586
Dan Sebald changed:
What|Removed |Added
Attachment #114619|0 |1
is obsolete|
https://bugs.freedesktop.org/show_bug.cgi?id=89586
Dan Sebald changed:
What|Removed |Added
Attachment #114596|0 |1
is obsolete|
https://bugs.freedesktop.org/show_bug.cgi?id=89586
Dan Sebald changed:
What|Removed |Added
Attachment #114597|0 |1
is obsolete|
https://bugs.freedesktop.org/show_bug.cgi?id=89586
Dan Sebald changed:
What|Removed |Added
Attachment #114598|0 |1
is obsolete|
https://bugs.freedesktop.org/show_bug.cgi?id=89586
Dan Sebald changed:
What|Removed |Added
Attachment #114599|0 |1
is obsolete|
Correct error with commit 151fb1e where assert was renamed
to unreachable without removing ! from string argument.
Signed-off-by: Tapani Pälli
---
src/glsl/loop_controls.cpp | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/src/glsl/loop_controls.cpp b/src/glsl/loop_controls.cp
Reviewed-by: Ilia Mirkin
On Mon, Mar 30, 2015 at 1:05 AM, Tapani Pälli wrote:
> Correct error with commit 151fb1e where assert was renamed
> to unreachable without removing ! from string argument.
>
> Signed-off-by: Tapani Pälli
> ---
> src/glsl/loop_controls.cpp | 2 +-
> 1 file changed, 1 in
Hello mesa-dev,
I've created a changeset for the legacy swrast_dri.so driver which fixes
vertical lines at the GL_MAX_TEXTURE_SIZE boundaries due to miscomputed
column range computations. Please consider.
Basically, I was very meticulous about the mathematical formulas in
terms of rounding
Kenneth Graunke writes:
> On Friday, March 27, 2015 01:54:32 PM Eric Anholt wrote:
>> This will be used by the VC4 driver for doing device-independent
>> optimization, and hopefully eventually replacing its whole IR. It also
>> may be useful to other drivers for the same reason.
> Hi Eric!
>
>
From: Dave Airlie
At the moment to get an EGL image to a dma-buf file descriptor,
you have to use EGL_MESA_drm_image, and then use libdrm to
convert this to a file descriptor.
This extension just provides an API modelled on EGL_MESA_drm_image,
to return a dma-buf file descriptor.
v2: update spe
Kenneth Graunke writes:
> Mostly a proof of concept that it works; we free the memory shortly
> afterwards anyway, so it's kind of dumb to do this.
>
> The plan is to instead build nir_shaders at link time, rather than when
> compiling each shader specialization, and delete the GLSL IR.
This sou
51 matches
Mail list logo
