Chih-Wei Huang writes:
The commit message needs some explanation: Why are we doing this? What
does EGL_ANDROID_recordable get us, if we don't have any configs
exposing it? What tests do we have for it?
> Signed-off-by: Chih-Wei Huang
> ---
> src/egl/main/eglconfig.c | 5 -
> src/egl/main
Emil Velikov writes:
> Signed-off-by: Emil Velikov
s/confiigure/configure/ in the subject.
After that, and some cleanup of the commit message in 4, patches 1-8
will be:
Reviewed-by: Eric Anholt
I find 9/10 weird -- if the DRI drivers need libglapi, why aren't they
just linking
Noticed while trying to add GL_ARB_texture_multisample support to vc4.
---
src/gallium/auxiliary/nir/tgsi_to_nir.c | 12 ++--
1 file changed, 10 insertions(+), 2 deletions(-)
diff --git a/src/gallium/auxiliary/nir/tgsi_to_nir.c
b/src/gallium/auxiliary/nir/tgsi_to_nir.c
index 061f39a..065
Kenneth Graunke writes:
> On Wednesday, June 03, 2015 09:21:11 PM Topi Pohjolainen wrote:
>> Shader programs are cached per stage (FS, VS, GS) using the
>> corresponding shader source identifier and compile time choices
>> as key. However, one not only stores the program binary but
>> a pair cons
a "num_decls" and
"array_size[_per_decl]" would have clarified things for me), and I'm
kind of bothered that we're not just splitting non-variable-indexed
array declarations at the NIR level (which would also solve the problem
it seems), but this seems correct, so:
nny location and decode on the other end.
>
> Signed-off-by: Ilia Mirkin
Yeah, seems like an OK patch for now, though we really need to start
using common enums instead.
Reviewed-by: Eric Anholt
signature.asc
Description: PGP signature
___
m
Ilia Mirkin writes:
> On Thu, Jul 2, 2015 at 2:31 PM, Jose Fonseca wrote:
>> On 02/07/15 17:49, Ilia Mirkin wrote:
>>>
>>> On Thu, Jul 2, 2015 at 12:40 PM, Jose Fonseca wrote:
On 02/07/15 17:24, Ilia Mirkin wrote:
>
>
> On Thu, Jul 2, 2015 at 12:17 PM, Jose Fonseca
>
Brian Paul writes:
> The main issue is I need to do additional work for cube and/or cube
> array textures. This is still on my to-do list. Do you need this?
FWIW, I just wanted this extension in X yesterday.
signature.asc
Description: PGP signature
__
Emil Velikov writes:
> Hello all,
>
> A recent patch by Chris, fixing some libudev fun in our loader, made
> me think if we can clear it up a bit.
>
> Having three different ways of retrieving the vendor/device ID does
> feel a bit excessive. Plus as one gets fixed others are likely to
> break -
Rob Clark writes:
> From: Rob Clark
>
> Some, but not all, state trackers will explicitly unref (and set to
> NULL) the previous *fence before calling pipe->flush(). So driver
> should use fence_ref() which will unref the old fence if not NULL.
>
> Signed-off-by: Ro
Timothy Arceri writes:
> ---
>
> Some of the AoA tests currently get stuck in an optimisation
> loop for a long time and this was taking a large amount of time.
> Its a symptom not a cause but I don't see any issue with using the faster
> version.
>
> src/glsl/ir_validate.cpp | 17 --
Rob Clark writes:
> well, freedreno and vmware ddx still use XA.. they both could probably
> be ported to use glamor instead, but that hasn't been done yet..
>
> At least for freedreno with upstream drm/kms driver, one can just use
> -modesetting ddx instead. But that doesn't work w/ android fbd
These are really useful hints to the compiler in the absence of link-time
optimization, and I'm going to use them in VC4.
I've made the const attribute be ATTRIBUTE_CONST unlike other function
attributes, because we have other things in the tree #defining CONST for
their own unrelated purposes.
--
On a release build, this makes the rest of vc4_qpu_validate.c go away
(the compiler didn't know that our qpu helper function calls had no
side effects).
---
src/gallium/drivers/vc4/vc4_qpu.h| 30 +++---
src/gallium/drivers/vc4/vc4_tiling.h | 6 +++---
2 files changed,
y/util/u_math.h
> +++ b/src/gallium/auxiliary/util/u_math.h
> @@ -240,7 +240,7 @@ util_iround(float f)
> static INLINE boolean
> util_is_approx(float a, float b, float tol)
> {
> - return fabs(b - a) <= tol;
> + return fabsf(b - a) <= tol;
> }
Reviewed-by: Eri
Timothy Arceri writes:
> Hi guys,
>
> As I've mentioned a couple of times in previous patches some of the cts AoA
> tests are taking very long time to compile. This is due to excessive
> optimisation passes mainly in the glsl optimisations (there are some slowdowns
> in the intel backend too but
Matt Turner writes:
> On Fri, Jul 17, 2015 at 10:17 AM, Emil Velikov
> wrote:
>> Cc: Eric Anholt
>> Signed-off-by: Emil Velikov
>> ---
>> src/util/tests/hash_table/Makefile.am | 3 +--
>> 1 file changed, 1 insertion(+), 2 deletions(-)
>>
>> di
Emil Velikov writes:
> On 14/07/15 19:45, Eric Anholt wrote:
>> These are really useful hints to the compiler in the absence of link-time
>> optimization, and I'm going to use them in VC4.
>>
>> I've made the const attribute be ATTRIBUTE_CONST unlike oth
re_complete(t, sampler)) {
> - return &null_sample_func;
> + return null_sample_func;
Much prettier.
Reviewed-by: Eric Anholt
signature.asc
Description: PGP signature
___
mesa-dev mailing list
mesa-dev@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/mesa-dev
&& mode != GL_SMOOTH) {
>_mesa_error(ctx, GL_INVALID_ENUM, "glShadeModel");
>return;
> }
>
> - if (ctx->Light.ShadeModel == mode)
> - return;
> -
> FLUSH_VERTICES(ctx, _NEW_LIGHT);
> ctx->Light.Shade
Adam Jackson writes:
> Without this this extension basically can't work in indirect contexts,
> TexImage2D will compute the image size as 0 and we'll send no image data
> to the server.
>
> v2: Add EXT_texture_integer to the client extension list too (Ian)
We can't completely drop this #define because it's used in the xserver
generated code. Until someone gets around to putting xml generation
directly in the server, we're stuck #defining it for them.
---
src/mapi/glapi/gen/gl_XML.py | 10 ++
1 file changed, 6 insertions(+), 4 deletions(-)
Brian Paul writes:
> ---
> src/mesa/main/blend.c | 2 +-
> 1 file changed, 1 insertion(+), 1 deletion(-)
>
> diff --git a/src/mesa/main/blend.c b/src/mesa/main/blend.c
> index d225f3d..f14949f 100644
> --- a/src/mesa/main/blend.c
> +++ b/src/mesa/main/blend.c
> @@ -407,7 +407,7 @@ _mesa_BlendEqu
thing like:
GLbitfield computed_states = ~(_NEW_CURRENT_ATTRIB | _NEW_LINE);
if (!(new_state & computed_states))
goto out;
making the optimization slightly more general and more self-documenting?
Either way, other than the comment on #7,
R
Steve Lynch writes:
> Hi,
>
> I've been using afl (http://lcamtuf.coredump.cx/afl/) on the standalone
> glsl compiler.
>
> It found four different crashes in the latest code in master and I have
> minimised the test cases that cause the crashes. I spent a couple of hours
> poking around but haven
Boyan Ding writes:
> Signed-off-by: Boyan Ding
> ---
> src/gallium/drivers/vc4/vc4_nir_lower_blend.c | 2 +-
> src/gallium/drivers/vc4/vc4_nir_lower_io.c| 4 ++--
> src/gallium/drivers/vc4/vc4_program.c | 8
> 3 files changed, 7 insertions(+), 7 deletions(-)
Pushed the vc4
Martin Peres writes:
> The preserved swap behavior is currently untested in piglit and not supported
> on the GLX side. Before working on implementing it for EGL/DRI3, let's
> disable it until support comes.
>
> This patch is trivial enough and will likely be squashed in the commit
> egl/x11: Imp
ume you meant *-op-div-large-uint-uint.shader_test.
vc4 doesn't have uge yet, but I've got a patch to add it and it does
fix one subtest. What this lowering pass is actually doing has never
really made sense to me, but it works, so:
Acked-by: Eric Anholt
signature.as
This is the equivalent of brw_fs_channel_expressions.cpp, which I wanted
for vc4.
v2: Use the nir_src_for_ssa() helper, and another instance of
nir_alu_src_copy().
v3: Drop the non-SSA support. All intended callers will have SSA-only ALU
ops.
v4: Use insert_before, drop stale bcsel/fcsel
tiguous
>> components is enough.
>>
>> On Thu, Jan 22, 2015 at 4:51 PM, Eric Anholt wrote:
>>> +static unsigned
>>> +insert_movs(nir_alu_instr *vec, unsigned start_channel,
>>> +unsigned start_src_idx, void *mem_ctx)
>
> We need a comment
Jason Ekstrand writes:
> Also, Could we rename this to nir_lower_alu_to_scalar? That's more
> descriptive.
That's even better. Done.
signature.asc
Description: PGP signature
___
mesa-dev mailing list
mesa-dev@lists.freedesktop.org
http://lists.free
The code was exactly the same, except util/ has c++ guards and a struct
simple_node declaration.
---
For my NIR work, there were clashes between gallium's copy (from my driver)
and Mesa's (from NIR)
src/gallium/auxiliary/Makefile.sources| 1 -
src/gallium/auxiliary/draw/draw_llvm.c
We have two copies of it in the tree, I'm going to delete one.
---
src/gallium/targets/dri/Makefile.am| 1 +
src/mesa/Makefile.sources | 1 -
src/mesa/drivers/dri/i915/i830_texblend.c | 2 +-
src/mesa/drivers/dri/i915/intel_syncobj.c
---
src/glsl/nir/nir_print.c | 22 --
1 file changed, 16 insertions(+), 6 deletions(-)
diff --git a/src/glsl/nir/nir_print.c b/src/glsl/nir/nir_print.c
index 1a50ae9..2ef55ed 100644
--- a/src/glsl/nir/nir_print.c
+++ b/src/glsl/nir/nir_print.c
@@ -210,7 +210,9 @@ print_var_dec
It's nice to have this present in your default cases so you can see what
instruction is triggering an abort.
v2: Just pass a NULL state, now that it won't crash when you do.
---
src/glsl/nir/nir.h | 1 +
src/glsl/nir/nir_print.c | 9 +++--
2 files changed, 8 insertions(+), 2 deletions(-
t) {
> - ret->indirect = nir_src_copy(deref->indirect, mem_ctx);
> + nir_src_copy(&ret->indirect, &deref->indirect, mem_ctx);
> }
> ret->deref.type = deref->deref.type;
> if (deref->deref.child)
Unintentional whitespace change?
Other
Jason Ekstrand writes:
> Previously, we called the abs() function in math.h. However, this involves
> unnecessarily going through double. This commit changes it to use integers
> directly with a ternary.
Or just use fabs()?
signature.asc
Description: PGP signature
___
Jason Ekstrand writes:
> diff --git a/src/glsl/nir/nir_lower_phis_to_scalar.c
> b/src/glsl/nir/nir_lower_phis_to_scalar.c
> new file mode 100644
> index 000..9f901d6
> --- /dev/null
> +++ b/src/glsl/nir/nir_lower_phis_to_scalar.c
> @@ -0,0 +1,238 @@
> +/*
> + * Copyright © 2014 Intel Corporat
Jason Ekstrand writes:
> On Mon, Jan 26, 2015 at 11:15 AM, Eric Anholt wrote:
>
>> Jason Ekstrand writes:
>>
>> > Previously, we called the abs() function in math.h. However, this
>> involves
>> > unnecessarily going through double. This commit c
Connor Abbott writes:
> On Fri, Jan 23, 2015 at 5:34 PM, Eric Anholt wrote:
>> Connor Abbott writes:
>>
>>> Argh, nevermind, I was reading it wrong...
>>>
>>> On Thu, Jan 22, 2015 at 8:18 PM, Connor Abbott wrote:
>>>> What happens if you h
The idea is that after a remove_from_list(), you might want to be able to
do a remove_from_list() on it again or an is_empty_list(). This is
apparently relied on by r300g.
---
src/util/simple_list.h | 1 +
1 file changed, 1 insertion(+)
diff --git a/src/util/simple_list.h b/src/util/simple_list.
Jason Ekstrand writes:
> On Mon, Jan 26, 2015 at 11:21 AM, Eric Anholt wrote:
>
>> Jason Ekstrand writes:
>> > + case nir_instr_type_phi: {
>> > + nir_phi_instr *src_phi = nir_instr_as_phi(src_instr);
>> > +
>> > + /* Ins
Jason Ekstrand writes:
> On Mon, Jan 26, 2015 at 3:58 PM, Eric Anholt wrote:
>
>> Jason Ekstrand writes:
>>
>> > On Mon, Jan 26, 2015 at 11:21 AM, Eric Anholt wrote:
>> >
>> >> Jason Ekstrand writes:
>> >> > + case
Matt Turner writes:
> On Tue, Jan 27, 2015 at 5:31 PM, Jason Ekstrand wrote:
>> + emit(MUL(acc, op[0], op[1]));
>> + emit(MACH(reg_null_d, op[0], op[1]));
>> + emit(MOV(result, fs_reg(acc)));
>>break;
>> }
>>
>> @@ -773,72 +767,38 @@ fs_visitor::nir_emit_alu(nir_alu_in
---
src/glsl/nir/nir.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/src/glsl/nir/nir.c b/src/glsl/nir/nir.c
index 25cdac0..10e6ed3 100644
--- a/src/glsl/nir/nir.c
+++ b/src/glsl/nir/nir.c
@@ -1295,7 +1295,7 @@ nir_instr_insert_before_cf(nir_cf_node *node, nir_instr
*before)
This will be used by tgsi_to_nir, which needs to get vec4 types for
declaring shader input/output variables.
---
src/glsl/nir/nir_types.cpp | 12
src/glsl/nir/nir_types.h | 3 +++
2 files changed, 15 insertions(+)
diff --git a/src/glsl/nir/nir_types.cpp b/src/glsl/nir/nir_types.cp
v2: Rebase on the nir_opcodes.h python code generation support.
v3: Use SSA values, and set an appropriate writemask on dot products.
---
src/glsl/Makefile.am | 5 ++
src/glsl/nir/.gitignore | 1 +
src/glsl/nir/nir_builder.h| 95
---
The i965 driver apparently only has things all in ssa or all in non-ssa, while
for TGSI I ended up doing some things directly in SSA before the to-SSA pass.
src/glsl/nir/nir_to_ssa.c | 3 +++
1 file changed, 3 insertions(+)
diff --git a/src/glsl/nir/nir_to_ssa.c b/src/glsl/nir/nir_to_ssa.c
in
We end up with these from TGSI-to-NIR because the pass generating the
comparisons doesn't know if the arg is actually a bool input or not.
total instructions in shared programs: 41801 -> 41508 (-0.70%)
instructions in affected programs: 4253 -> 3960 (-6.89%)
---
src/glsl/nir/nir_opt_algebraic
Connor Abbott writes:
> Hi Eric,
>
> A general suggestion:
>
> I think that we should rename these functions to "nir_op_create()" so
> that we can make a more specific "nir_op()" function that takes and
> returns a nir_ssa_def *, so that we can do stuff like:
>
> nir_builder *b = ...
>
> nir_ssa_
Jason Ekstrand writes:
> I've got a patch I cooked up today that generalizes this a bit. I haven't
> sent it yet because it lacks shader-db statistics. That said, I'm fine
> with merging this in the mean time.
Not sure what generalization you're working on, but something that let
me have "any
Roland Scheidegger writes:
> Hi,
>
> the address reg in tgsi is quite a nuisance. glsl-to-tgsi code assumes
> that indirections can only be done through the address reg and has quite
> some extra code to deal with this. Even though hardware and apis which
> worked like that are definitely old by
v2: Rebase on the nir_opcodes.h python code generation support.
v3: Use SSA values, and set an appropriate writemask on dot products.
v4: Make the arguments be SSA references as well. This lets you stack up
expressions in the arguments of other expressions, at the cost of
having to insert
This could be done in a separate pass like we do in GLSL IR, but it seems
to me like having the definitions of the transformations in the two
directions next to each other makes a lot of sense.
---
src/glsl/nir/nir_opt_algebraic.py | 1 +
1 file changed, 1 insertion(+)
diff --git a/src/glsl/nir/n
vc4 cse/algebraic-disabled stats:
total instructions in shared programs: 44911 -> 44732 (-0.40%)
instructions in affected programs: 11371 -> 11192 (-1.57%)
---
src/gallium/drivers/vc4/vc4_program.c | 1 +
src/glsl/nir/nir.h| 2 ++
src/glsl/nir/nir_opt_algebraic.py | 1
Mesa has a shader compiler struct flagging whether GLSL IR's opt_algebraic
and other passes should try and generate certain types of opcodes or
patterns. Extend that to NIR by defining our own struct, which is
automatically generated from the Mesa struct in glsl_to_nir and provided
directly by the
I'm pretty happy with how it turned out. What I'm not sure on is that
I've got a couple of places where I replicated another transformation,
like fsat(fsat(a)) -> fsat(a) but with a lowered fsat. I could see
potentially having only the canonical form appear in nir_opt_algebraic,
and I'd have two
vc4 cse/algebraic-disabled stats:
total uniforms in shared programs: 13972 -> 13966 (-0.04%)
uniforms in affected programs: 408 -> 402 (-1.47%)
total instructions in shared programs: 44973 -> 44911 (-0.14%)
instructions in affected programs: 1551 -> 1489 (-4.00%)
---
src/gallium/drivers/vc
No effect on vc4 shader-db.
---
src/glsl/nir/nir_opt_algebraic.py | 2 ++
1 file changed, 2 insertions(+)
diff --git a/src/glsl/nir/nir_opt_algebraic.py
b/src/glsl/nir/nir_opt_algebraic.py
index 9542ef3..fc7e47c 100644
--- a/src/glsl/nir/nir_opt_algebraic.py
+++ b/src/glsl/nir/nir_opt_algebraic.
We've probably never seen this ridiculous pattern in the wild, so it
didn't matter.
---
src/glsl/nir/nir_opt_algebraic.py | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/src/glsl/nir/nir_opt_algebraic.py
b/src/glsl/nir/nir_opt_algebraic.py
index 8a14fe0..84c60ec 100644
--- a/s
---
src/glsl/nir/nir_algebraic.py | 9 +++--
1 file changed, 3 insertions(+), 6 deletions(-)
diff --git a/src/glsl/nir/nir_algebraic.py b/src/glsl/nir/nir_algebraic.py
index ea7f5fc..4929745 100644
--- a/src/glsl/nir/nir_algebraic.py
+++ b/src/glsl/nir/nir_algebraic.py
@@ -179,10 +179,7 @@ st
vc4 cse/algebraic-disabled stats:
total instructions in shared programs: 44356 -> 44354 (-0.00%)
instructions in affected programs: 55 -> 53 (-3.64%)
---
src/gallium/drivers/vc4/vc4_program.c | 1 +
src/glsl/nir/nir.h| 1 +
src/glsl/nir/nir_opt_algebraic.py | 3 ++-
3 f
vc4 cse/algebraic-disabled stats:
total uniforms in shared programs: 13966 -> 13791 (-1.25%)
uniforms in affected programs: 435 -> 260 (-40.23%)
total instructions in shared programs: 44732 -> 44356 (-0.84%)
instructions in affected programs: 9599 -> 9223 (-3.92%)
---
src/gallium/drivers/v
Connor Abbott writes:
> On Sun, Feb 1, 2015 at 4:17 PM, Eric Anholt wrote:
>> We've probably never seen this ridiculous pattern in the wild, so it
>> didn't matter.
>
> I don't think this pattern is 'ridiculous' at all -- GLSL doesn't have
>
Jason Ekstrand writes:
> Hi all,
> I wanted to send out a quick message about naming conventions for mesa
> entrypoints and dd table fallbacks. There has been some confusion and
> disagreement about this with some of the stuff Laura has been doing to
> implement DSA. Instead of side-track one o
I wanted to use it for nir-to-tgsi. The equivalent ureg_src_register() is
also located here.
---
src/gallium/auxiliary/tgsi/tgsi_ureg.c | 28
src/gallium/auxiliary/tgsi/tgsi_ureg.h | 25 +
2 files changed, 25 insertions(+), 28 deletions(-)
dif
I wanted all of these for NIR-to-TGSI.
---
src/gallium/auxiliary/tgsi/tgsi_opcode_tmp.h | 4
1 file changed, 4 insertions(+)
diff --git a/src/gallium/auxiliary/tgsi/tgsi_opcode_tmp.h
b/src/gallium/auxiliary/tgsi/tgsi_opcode_tmp.h
index 147d989..d8752ce 100644
--- a/src/gallium/auxiliary/tgs
Francisco Jerez writes:
> Eric Anholt writes:
>
>> Francisco Jerez writes:
>>
>>> This reverts commit 3fad0868f023f1d726e230968a4df3327de38823.
>>>
>>> This test doesn't make any sense to me, it begins quoting the GLSL
>>> 1.30 spec
Thomas Helland writes:
> 2015-02-02 19:18 GMT+01:00 Jason Ekstrand :
>> Hi all,
>> I wanted to send out a quick message about naming conventions for mesa
>> entrypoints and dd table fallbacks. There has been some confusion and
>> disagreement about this with some of the stuff Laura has been doin
Roland Scheidegger writes:
> Am 02.02.2015 um 20:39 schrieb Eric Anholt:
>> I wanted all of these for NIR-to-TGSI.
>> ---
>> src/gallium/auxiliary/tgsi/tgsi_opcode_tmp.h | 4
>> 1 file changed, 4 insertions(+)
>>
>> diff --git a/src/gallium/aux
Erik Faye-Lund writes:
> On Sun, Feb 1, 2015 at 10:17 PM, Eric Anholt wrote:
>> This could be done in a separate pass like we do in GLSL IR, but it seems
>> to me like having the definitions of the transformations in the two
>> directions next to each other makes a lot of
This will be used to give the optimization passes a chance to customize
behavior for the particular target device.
---
I had things as one giant patch before because I thought you couldn't
have a struct with no members, and having a temporary version with a
dummy member seemed ugly. Turns out you
This will be used so that we can customize the transforms for the target
GPU, so we don't un-lower expressions that had already been lowered (or
introduce new lowering transformations that not all GPUs want)
---
src/glsl/nir/nir_algebraic.py | 44 ---
1 file
Mesa has a shader compiler struct flagging whether GLSL IR's opt_algebraic
and other passes should try and generate certain types of opcodes or
patterns. Extend that to NIR by defining our own struct, which is
automatically generated from the Mesa struct in glsl_to_nir and provided
directly by the
Connor Abbott writes:
> I mentioned this on IRC, but it would be good to add the ability to
> append the sequence of instructions before/after an instruction as
> well as at the beginning or end of a basic block. We would need to
> store the current basic block, the current instruction, and an en
Connor Abbott writes:
> A few comments here...
>
> 1. We should really split this up into a patch that adds
> nir_shader_compiler_options (which would be empty at this point) and
> the nir_algebraic infrastructure for conditionalizing transforms
> (actually, maybe we should split these up into tw
Connor Abbott writes:
> I think this logic should be flipped around... you should have a flag
> 'lower_sub' which is false for vc4 but true for i965, and then if the
> flag is true we lower sub to add+neg and if false we try and
> reconstruct sub (and do other things like pulling neg's out of
> m
v2: Rebase on the nir_opcodes.h python code generation support.
v3: Use SSA values, and set an appropriate writemask on dot products.
v4: Make the arguments be SSA references as well. This lets you stack up
expressions in the arguments of other expressions, at the cost of
having to insert
ming to the
other repo. Series is:
Reviewed-by: Eric Anholt
signature.asc
Description: PGP signature
___
mesa-dev mailing list
mesa-dev@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/mesa-dev
No change on shader-db on i965.
---
src/glsl/nir/nir_opt_algebraic.py | 9 +
1 file changed, 9 insertions(+)
diff --git a/src/glsl/nir/nir_opt_algebraic.py
b/src/glsl/nir/nir_opt_algebraic.py
index a5fe19a..0512a8f 100644
--- a/src/glsl/nir/nir_opt_algebraic.py
+++ b/src/glsl/nir/nir_opt
NIR instruction count results on i965:
total instructions in shared programs: 1261954 -> 1261937 (-0.00%)
instructions in affected programs: 455 -> 438 (-3.74%)
One in yofrankie, two in tropics. Apparently i965 had also optimized all
of these out anyway.
---
src/glsl/nir/nir_opt_cse.c | 43 +
We want the size of a float per component, not the size of a whole vec4.
NIR instructions on i965:
total instructions in shared programs: 1261937 -> 1261929 (-0.00%)
instructions in affected programs: 114 -> 106 (-7.02%)
Looking at one of these examples (tesseract), it's from vec4 load_consts
Connor Abbott writes:
> On Sat, Feb 7, 2015 at 12:16 AM, Eric Anholt wrote:
>> NIR instruction count results on i965:
>> total instructions in shared programs: 1261954 -> 1261937 (-0.00%)
>> instructions in affected programs: 455 -> 438 (-3.74%)
>>
&g
Kenneth Graunke writes:
> On Tuesday, February 10, 2015 08:08:38 PM Matt Turner wrote:
>> Dead since
>>
>>commit 284ce20901b0c2cfab1d952cc129b8f3cd068f12
>>Author: Eric Anholt
>>Date: Fri Aug 20 10:52:14 2010 -0700
>>
>>
Connor Abbott writes:
> On Tue, Feb 10, 2015 at 1:32 PM, Eric Anholt wrote:
>> Connor Abbott writes:
>>
>>> On Sat, Feb 7, 2015 at 12:16 AM, Eric Anholt wrote:
>>>> NIR instruction count results on i965:
>>>> total instructions in shared program
Connor Abbott writes:
> On Wed, Feb 11, 2015 at 2:36 PM, Eric Anholt wrote:
>> Connor Abbott writes:
>>
>>> On Tue, Feb 10, 2015 at 1:32 PM, Eric Anholt wrote:
>>>> Connor Abbott writes:
>>>>
>>>>> On Sat, Feb 7, 2015 at 12:16
We were filling out almost all fields of almost all instructions, but
leaving out a couple of them. This simplifies the source code, cuts 700
bytes from the compiled binary, and prevents developer surprise when one
field of your otherwise-containing-defaults struct is actually
uninitialized.
---
Quiets compiler warning since e7f2f2dea5acdbd1a12ed88914e64a38a97432f0.
---
src/mesa/drivers/dri/r200/r200_ioctl.c | 1 -
1 file changed, 1 deletion(-)
diff --git a/src/mesa/drivers/dri/r200/r200_ioctl.c
b/src/mesa/drivers/dri/r200/r200_ioctl.c
index 515be92..d665c8b 100644
--- a/src/mesa/driver
If execution was supposed to be supported in this case, we'd run into
trouble from completely uninitialized sat_imm values. Shuts up compiler
warnings.
---
src/mesa/drivers/dri/i965/brw_shader.cpp | 4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)
diff --git a/src/mesa/drivers/dri/i965/br
The compiler doesn't see that buffers is set in the !image case and used
in the !image case.
---
src/gallium/state_trackers/dri/dri2.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/src/gallium/state_trackers/dri/dri2.c
b/src/gallium/state_trackers/dri/dri2.c
index aecc8eb..7
We always pass this argument, even if it won't be used by the particular
texture op.
---
src/mesa/drivers/dri/i965/brw_fs_nir.cpp | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/src/mesa/drivers/dri/i965/brw_fs_nir.cpp
b/src/mesa/drivers/dri/i965/brw_fs_nir.cpp
index da81b61..
The compiler can't tell that we're always going to hit the first if block
on the first time through the loop.
---
src/mesa/drivers/dri/i965/brw_vec4.cpp | 4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)
diff --git a/src/mesa/drivers/dri/i965/brw_vec4.cpp
b/src/mesa/drivers/dri/i965/brw_v
It introduces references to gallium util/ symbols which means we don't get
to include it from outside-of-gallium code.
---
src/gallium/auxiliary/util/u_math.h | 1 -
1 file changed, 1 deletion(-)
diff --git a/src/gallium/auxiliary/util/u_math.h
b/src/gallium/auxiliary/util/u_math.h
index 5db5b66
It's used in one of the methods, not in the structure definitions.
---
src/glsl/glsl_types.h | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/src/glsl/glsl_types.h b/src/glsl/glsl_types.h
index 441015c..74a038a 100644
--- a/src/glsl/glsl_types.h
+++ b/src/glsl/glsl_types.h
@@ -2
---
src/glsl/Makefile.am | 2 ++
src/mesa/drivers/dri/common/Makefile.am | 2 ++
src/mesa/drivers/dri/i915/Makefile.am| 2 ++
src/mesa/drivers/dri/i965/Makefile.am| 2 ++
src/mesa/drivers/dri/nouveau/Makefile.am | 2 ++
src/mesa/drivers/dri/r200/Makefile.am| 2 ++
OK, I think this is kind of the next step from the gallium-versus-mesa
includes discussion earlier: Accept that u_math.h is an OK name, and the
parent directory is OK, too, and just start sharing code. Once that's in
place, I can get bitset.h out of Mesa main and into util/, so that nir.h
doesn't
---
src/glsl/nir/nir.h | 2 +-
src/mesa/Makefile.sources | 1 -
.../drivers/dri/i965/brw_fs_copy_propagation.cpp | 2 +-
src/mesa/drivers/dri/i965/brw_fs_live_variables.h | 2 +-
.../drivers/dri/i965/brw_performance_monitor.c | 2
Note that we can't use u_math.h's align() because it's a function instead
of a macro, while BITSET_DECLARE needs a constant expression for nouveau's
usage in global declarations.
---
src/mesa/main/bitset.h | 7 +++
1 file changed, 3 insertions(+), 4 deletions(-)
diff --git a/src/mesa/main/bit
This avoids duplication of some macros and other definitions across the
tree.
Note that COPY_4FV switches from a memcpy-based implementation to an
assignment of 4 floats.
---
src/mesa/drivers/dri/i915/intel_tris.c | 6 --
src/mesa/drivers/dri/nouveau/nouveau_util.h | 6 --
src/mesa
One less new directory necessary for gallium code that wants to interact
with NIR.
---
src/glsl/nir/nir.h | 4 +++-
src/glsl/nir/nir_lower_atomics.c | 1 +
src/glsl/nir/nir_lower_system_values.c | 1 +
3 files changed, 5 insertions(+), 1 deletion(-)
diff --git a/src/glsl
1 - 100 of 5753 matches
Mail list logo
