Annotate the according parameters accordingly.
---
src/gallium/auxiliary/util/u_debug_refcnt.h | 4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)
diff --git a/src/gallium/auxiliary/util/u_debug_refcnt.h
b/src/gallium/auxiliary/util/u_debug_refcnt.h
index cf04777666..daccdb2cc1 100644
---
Use size_t instread of unsigned for new_max. realloc later expects
size_t as parameter anyway.
---
src/gallium/auxiliary/util/u_async_debug.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/src/gallium/auxiliary/util/u_async_debug.c
b/src/gallium/auxiliary/util/u_async_debug.c
Explicitely convert one value to compare.
---
src/gallium/auxiliary/util/u_surface.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/src/gallium/auxiliary/util/u_surface.c
b/src/gallium/auxiliary/util/u_surface.c
index 0a79a25a43..c66f279dc2 100644
--- a/src/gallium/auxiliary/
Decorate the params with "UNUSED" accordingly.
---
src/gallium/auxiliary/util/u_transfer.c | 20 ++--
1 file changed, 10 insertions(+), 10 deletions(-)
diff --git a/src/gallium/auxiliary/util/u_transfer.c
b/src/gallium/auxiliary/util/u_transfer.c
index ba4b9dcc6b..10dc59a926 1006
* Decorate the params accordingly with UNUSED or MAYBE_UNUSED (for params
that are used in debug mode) to fix -Wunused-param warnings.
* Enclose if body with braces, since it is empty in release mode to fix
-Wparentheses warning.
---
src/gallium/auxiliary/util/u_threaded_context.c | 19 ++
Decorate the unused param accordingly with "UNUSED".
---
src/gallium/auxiliary/util/u_mm.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/src/gallium/auxiliary/util/u_mm.c
b/src/gallium/auxiliary/util/u_mm.c
index 7a45e29192..71c079f8fe 100644
--- a/src/gallium/auxiliary/util
Decorate the params accordingly with "UNUSED".
---
src/mesa/main/texcompress_s3tc_tmp.h | 4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)
diff --git a/src/mesa/main/texcompress_s3tc_tmp.h
b/src/mesa/main/texcompress_s3tc_tmp.h
index 61630f2475..c02f6242fc 100644
--- a/src/mesa/main/texco
Annotate the parameters accordingly.
---
src/gallium/auxiliary/util/u_blit.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/src/gallium/auxiliary/util/u_blit.c
b/src/gallium/auxiliary/util/u_blit.c
index 3e49667029..ee8d89d9ff 100644
--- a/src/gallium/auxiliary/util/u_blit.c
With --disable-debug a parameter is not used. Silence this
warning by fake-using it.
---
src/gallium/auxiliary/os/os_thread.h | 2 ++
1 file changed, 2 insertions(+)
diff --git a/src/gallium/auxiliary/os/os_thread.h
b/src/gallium/auxiliary/os/os_thread.h
index d9c685922a..f2629c5ffe 100644
--- a
For the series:
Reviewed-By: Gert Wollny
On 11.11.2017 03:40, Brian Paul wrote:
> To get rid of forward declaration.
> ---
> src/mesa/state_tracker/st_context.c | 123
> ++--
> 1 file changed, 62 insertions(+), 61 deletions(-)
>
>
Hi Dave,
since I'm currently experimenting with improving the performance of the
tesselation shaders on r600, I notes that this series has a negative
impact on performance, i.e. with the Unigine_Heaven benchmark I get a
10% lower frame rate that can be attributed to the newly added fence
that is
Obviously, you can now disregard my other email about the performance
regression, this fixes it.
Tested-By: Gert Wollny
Best wishes,
Am Freitag, den 10.11.2017, 13:47 +1000 schrieb Dave Airlie:
> From: Dave Airlie
>
> Otherwise we end up emitting the fence.
>
> Signed-off-
Hello Brian,
thanks for the review, one thing though ...
Am Montag, den 13.11.2017, 09:31 -0700 schrieb Brian Paul:
> Also, can you please wrap long lines at 78 columns?
I'm kind of in doubt whether the hard limit of 78 makes sense here,
i.e. when I do this then I will have to put every parame
Hi,
I'd like to send out a ping for this series,
many thanks,
Gert
Am Mittwoch, den 25.10.2017, 11:51 +0200 schrieb Gert Wollny:
> Dear all,
>
> this is a minor update to the patch set that adds enhanced tracking
> of IF/ELSE branches and tracking of reladdr*
Am Dienstag, den 14.11.2017, 14:53 +0100 schrieb Nicolai Hähnle:
> On 14.11.2017 12:55, Emil Velikov wrote:
> > On 12 November 2017 at 13:55, Gert Wollny
> > wrote:
> > > Hello all,
> > >
> > > this might be a bit over-zealos, but I usually prefer to
forces an additional CF_OP_NOP group like it is already done for
other final CF_OP groups.
Signed-off-by: Gert Wollny
---
src/gallium/drivers/r600/r600_shader.c | 4 +++-
1 file changed, 3 insertions(+), 1 deletion(-)
diff --git a/src/gallium/drivers/r600/r600_shader.c
b/src/gallium/drivers
Hi Dave,
tested on BARTS: the patch changes TESS_EVAL to end with a proper EOP
in
trivial-tess-gs_no-gs-inputs
(it doesn't fix the piglit though). However, it introduces a regression
in
vs-tcs-tes-tessinner-tessouter-inputs-quads
vs-tes-tessinner-tessouter-inputs-quads
because in th
Am Mittwoch, den 15.11.2017, 11:28 +1000 schrieb Dave Airlie:
>
> I'll likely apply this as well at some point, that test however hangs
> my cayman with or without my patch (and your patch doesn't do
> anything on cayman).
BTW: valgrind showed that there was some problem in sb in the parser
(bran
Pre-load all the LDS values who's range is accessed more than once.
Signed-off-by: Gert Wollny
---
src/gallium/drivers/r600/r600_shader.c | 33 +
1 file changed, 33 insertions(+)
diff --git a/src/gallium/drivers/r600/r600_shader.c
b/src/gallium/drivers
Cache values that are loaded more then once, or where various components
are loaded at separate places. This saves repeated calculation of the offsets.
Signed-off-by: Gert Wollny
---
src/gallium/drivers/r600/r600_shader.c | 211 +
1 file changed, 190 insertions
Instead of calculating creating the code for calculating a base offset
and then to caclucate the component offfsets, calculate this offset
for all components directly. This saves one instruction group.
Signed-off-by: Gert Wollny
---
src/gallium/drivers/r600/r600_shader.c | 113
4x640
---
Points:635 700
FPS:10 11
A github repo inclusing these patches can be found at
https://github.com/gerddie/mesa/tree/r600-tess-speedup
many thanks for any comments,
Gert
Gert Wollny (6):
r600:shader: Fi
Some offsets used for the LDS access are recalculated quite regularly.
Since tesselation shaders are not optimized by the SB manually pre-evaluate
some offsets to speed up this type of shader.
Signed-off-by: Gert Wollny
---
src/gallium/drivers/r600/r600_shader.c | 253
- fix a number of -Wsign-compare warnings
- fix two warnings for -Woverride-init because TGSI_OPCODE_CEIL == 83, and
the according field was defined two times.
Signed-off-by: Gert Wollny
---
src/gallium/drivers/r600/r600_shader.c | 67 ++
1 file changed, 36
Use the destination write mask to determine which values are really to be
read from LDS and load only these.
Signed-off-by: Gert Wollny
---
src/gallium/drivers/r600/r600_shader.c | 33 ++---
1 file changed, 26 insertions(+), 7 deletions(-)
diff --git a/src/gallium
sic-imagestore-from-uniform
disable_early_z
image_checkerboard
load-from-cleared-image
write-to-rendered-image
arb_shading_language_420pack
different-bindings-image2d
For the series:
Tested-By: Gert Wollny
Best,
Gert
_
Decorate the unused params with "UNUSED".
v2: move UNUSED decoration in front of parameter declaration
Signed-off-by: Gert Wollny
Reviewed-by: Brian Paul (v1)
---
src/gallium/auxiliary/util/u_format_latc.c | 24 +---
1 file changed, 17 insertions(+), 7 deletion
u_bit_scan may return -1 that then may be interpreted as (unsigned)-1 in
the following comparison, since num_names is unsigned. Convert the latter to
be int as well.
Signed-off-by: Gert Wollny
Reviewed-by: Brian Paul
---
src/gallium/auxiliary/util/u_dump_defines.c | 2 +-
1 file changed, 1
the according
UNUSED decoration.
- additionally -Wmissing-field-initialzers warnings were addressed.
Best,
Gert
PS: I don't have git write access.
Gert Wollny (29):
gallium/aux/util/u_blitter.c: Silence some warnings
gallium/aux/util/u_debug_describe.c: Silence an -Wunused-param warning
g
v2: move UNUSED decoration in front of parameter declaration
Signed-off-by: Gert Wollny
Reviewed-by: Brian Paul (v1)
---
src/gallium/auxiliary/util/u_dump_state.c | 4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)
diff --git a/src/gallium/auxiliary/util/u_dump_state.c
b/src/gallium
Decorate the params accordingly with "UNUSED".
v2: move UNUSED decoration in front of parameter declaration
Signed-off-by: Gert Wollny
Reviewed-by: Brian Paul (v1)
---
src/gallium/auxiliary/util/u_format_rgtc.c | 68 +-
1 file changed, 39 inserti
Decorate the params accordingly with "UNUSED".
v2: move UNUSED decoration in front of parameter declaration
Signed-off-by: Gert Wollny
Reviewed-by: Brian Paul (v1)
---
src/gallium/auxiliary/util/u_format_yuv.c | 232 +++---
1 file changed, 116 insertio
Decorate the parameters accordingly with "UNUSED".
v2: move UNUSED decoration in front of parameter declaration
Signed-off-by: Gert Wollny
Reviewed-by: Brian Paul (v1)
---
src/gallium/auxiliary/util/u_format_etc.c | 8 ++--
1 file changed, 6 insertions(+), 2 deletions(-)
di
This warning was issued only in release mode. Fix it by fake-using the
parameter.
Signed-off-by: Gert Wollny
Reviewed-by: Brian Paul (v1)
---
src/gallium/auxiliary/util/u_format.c | 2 ++
1 file changed, 2 insertions(+)
diff --git a/src/gallium/auxiliary/util/u_format.c
b/src/gallium
* Annotate three parameters that are not used in release mode.
* explicitely convert an int to unsigned in an ?: construct.
v2: move MAYBE_UNUSED decoration in front of parameter declaration
Signed-off-by: Gert Wollny
Reviewed-by: Brian Paul (v1)
---
src/gallium/auxiliary/util/u_blitter.c
Decorate the unused params with "UNUSED".
v2: move UNUSED decoration in front of parameter declaration
Signed-off-by: Gert Wollny
Reviewed-by: Brian Paul (v1)
---
src/gallium/auxiliary/util/u_format_other.c | 34 ++---
1 file changed, 17 insertions(+), 17
Decorate the unused param accordingly with "UNUSED".
v2: move UNUSED decoration in front of parameter declaration
Signed-off-by: Gert Wollny
Reviewed-by: Brian Paul (v1)
---
src/gallium/auxiliary/util/u_mm.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/s
Annotate the unused parameter.
v2: move UNUSED decoration in front of parameter declaration
Signed-off-by: Gert Wollny
Reviewed-by: Brian Paul (v1)
---
src/gallium/auxiliary/util/u_debug_describe.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/src/gallium/auxiliary
he call failed.
Signed-off-by: Gert Wollny
Reviewed-by: Brian Paul (v1)
---
src/gallium/auxiliary/util/u_debug_stack.c | 6 +++---
1 file changed, 3 insertions(+), 3 deletions(-)
diff --git a/src/gallium/auxiliary/util/u_debug_stack.c
b/src/gallium/auxiliary/util/u_debug_stack.c
index 7013807b6
Silence the warning by making the conversion to int explicit.
Signed-off-by: Gert Wollny
Reviewed-by: Brian Paul
---
src/gallium/auxiliary/util/u_pstipple.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/src/gallium/auxiliary/util/u_pstipple.c
b/src/gallium/auxiliary/util
Explicitely convert one value to compare.
Signed-off-by: Gert Wollny
Reviewed-by: Brian Paul
---
src/gallium/auxiliary/util/u_surface.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/src/gallium/auxiliary/util/u_surface.c
b/src/gallium/auxiliary/util/u_surface.c
index
Annotate the according parameters accordingly.
v2: move UNUSED decoration in front of parameter declaration
Signed-off-by: Gert Wollny
Reviewed-by: Brian Paul (v1)
---
src/gallium/auxiliary/util/u_debug_refcnt.h | 4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)
diff --git a/src
Annotate the parameters accordingly.
v2: move UNUSED decoration in front of parameter declaration
Signed-off-by: Gert Wollny
v1: Reviewed-by: Brian Paul (v1)
---
src/gallium/auxiliary/util/u_blit.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/src/gallium/auxiliary/util
Decorate the params accordingly with UNUSED or MAYBE_UNUSED (for params
that are used in debug mode).
v2: move *UNUSED decoration in front of parameter declaration
Signed-off-by: Gert Wollny
Reviewed-by: Brian Paul (v1)
---
src/gallium/auxiliary/util/u_threaded_context.c | 21
Use size_t instread of unsigned for new_max. realloc later expects
size_t as parameter anyway.
Signed-off-by: Gert Wollny
Reviewed-by: Brian Paul
---
src/gallium/auxiliary/util/u_async_debug.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/src/gallium/auxiliary/util
Decorate the params accordingly with "UNUSED".
v2: move UNUSED decoration in front of parameter declaration
Signed-off-by: Gert Wollny
Reviewed-by: Brian Paul (v1)
---
src/mesa/main/texcompress_s3tc_tmp.h | 4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)
diff --git a/src
Decorate the parameters accordingly with "UNUSED" or "MAYBE_UNUSED" (for
the param that is used in debug mode, but not in release mode).
v2: move UNUSED decoration in front of parameter declaration
Signed-off-by: Gert Wollny
Reviewed-by: Brian Paul (v1)
---
src/util/simp
Signed-off-by: Gert Wollny
---
src/gallium/auxiliary/util/u_format_pack.py | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/src/gallium/auxiliary/util/u_format_pack.py
b/src/gallium/auxiliary/util/u_format_pack.py
index a2327196e6..c9b8cd7abc 100644
--- a/src/gallium
* Use a designated initializer to silence the warning.
* fix one intention that was using tabs instead of spaces
Signed-off-by: Gert Wollny
---
src/gallium/auxiliary/util/u_tests.c | 8
1 file changed, 4 insertions(+), 4 deletions(-)
diff --git a/src/gallium/auxiliary/util/u_tests.c
Decorate the according parameters with UNUSED.
Signed-off-by: Gert Wollny
---
src/gallium/auxiliary/util/u_debug_image.c | 4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)
diff --git a/src/gallium/auxiliary/util/u_debug_image.c
b/src/gallium/auxiliary/util/u_debug_image.c
index
Decorate the params with "UNUSED" accordingly.
v2: move UNUSED decoration in front of parameter declaration
Signed-off-by: Gert Wollny
Reviewed-by: Brian Paul (v1)
---
src/gallium/auxiliary/util/u_transfer.c | 20 ++--
1 file changed, 10 insertions(+), 10 deletion
With --disable-debug a parameter is not used. Silence this
warning by fake-using it.
Signed-off-by: Gert Wollny
Reviewed-by: Brian Paul
---
src/gallium/auxiliary/os/os_thread.h | 2 ++
1 file changed, 2 insertions(+)
diff --git a/src/gallium/auxiliary/os/os_thread.h
b/src/gallium/auxiliary
Silence warnings by decoration the parameters with UNUSED.
Signed-off-by: Gert Wollny
---
src/gallium/auxiliary/util/u_debug.c | 35 +++
1 file changed, 19 insertions(+), 16 deletions(-)
diff --git a/src/gallium/auxiliary/util/u_debug.c
b/src/gallium/auxiliary
Decorate the unused parameters with UNUSED.
Signed-off-by: Gert Wollny
---
src/gallium/auxiliary/util/u_debug_flush.c | 7 ---
1 file changed, 4 insertions(+), 3 deletions(-)
diff --git a/src/gallium/auxiliary/util/u_debug_flush.c
b/src/gallium/auxiliary/util/u_debug_flush.c
index
* Explicitely convert values to int in comparison.
* Remove one MAYBE_UNUSED that is actually not needed.
Signed-off-by: Gert Wollny
---
src/gallium/auxiliary/util/u_surface.c | 10 +-
1 file changed, 5 insertions(+), 5 deletions(-)
diff --git a/src/gallium/auxiliary/util/u_surface.c
Explicitely convert on value to the target type.
Signed-off-by: Gert Wollny
Reviewed-by: Brian Paul
---
src/mesa/main/framebuffer.h | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/src/mesa/main/framebuffer.h b/src/mesa/main/framebuffer.h
index bc6e7bc31a..bffa21dcea 100644
Am Donnerstag, den 16.11.2017, 16:34 + schrieb Emil Velikov:
> On 16 November 2017 at 15:10, Gert Wollny
> wrote:
> > * Use a designated initializer to silence the warning.
> > * fix one intention that was using tabs instead of spaces
> >
>
> Sigh, -Wmiss
Am Donnerstag, den 16.11.2017, 17:13 + schrieb Emil Velikov:
> On 16 November 2017 at 15:09, Gert Wollny
> wrote:
> > Explicitely convert on value to the target type.
> >
> > Signed-off-by: Gert Wollny
> > Reviewed-by: Brian Paul
> > ---
> > src/
Am Mittwoch, den 15.11.2017, 11:28 +1000 schrieb Dave Airlie:
> On 15 November 2017 at 04:50, Gert Wollny
> wrote:
> > Although the EOP bit is documented for the vertex fetch clause, it
> > is not
> > properly interpreted. As a result the piglit
> > spec/arb
6:09 +0100 schrieb Gert Wollny:
> Dear all,
>
> I've cleaned up the patch according to Brian's and Emil's pointers.
>
> I tried to keep close to the original formatting, unless the original
> lines
> were overly long (+90 chars on a line). I think keeping the
ction, which might have resulted
in GPU lockups.
Signed-off-by: Gert Wollny
---
I put an an RFC because the patch was only tested on BARTS (no piglit
changes w.r.t. master), and hence, before it is pushed it would probably
be good if it was tested on pre-evergreen cards, since for these cards
Am Donnerstag, den 16.11.2017, 22:52 +0100 schrieb Gert Wollny:
> Am Donnerstag, den 16.11.2017, 17:13 + schrieb Emil Velikov:
> >
> > According to the spec both of these are 'Z + Non-negative integer
> > or enumerated value' Worth doing that instead?
>
>
ssure this patch makes
the piglits
spec/glsl-1.50/execution -
variable-indexing/vs-output-array-vec3-index-wr-before-gs
geometry/max-input-components
pass on r600 (barts) where they would fail before with a "GPR limit exceeded"
error.
Signed-off-by: Gert Wollny
---
No furt
Am Freitag, den 24.11.2017, 18:07 + schrieb Eric Engestrom:
> Signed-off-by: Eric Engestrom
> ---
> src/util/ralloc.c | 18 +-
> src/util/slab.c | 4 ++--
> 2 files changed, 11 insertions(+), 11 deletions(-)
>
> diff --git a/src/util/ralloc.c b/src/util/ralloc.c
> index 42
Am Sonntag, den 26.11.2017, 18:01 -0800 schrieb Eric Anholt:
> Gert Wollny writes:
>
> > Array who's elements are only accessed directly are replaced by the
> > according number of temporary registers. By doing so the otherwise
> > > reserved register
Am Montag, den 27.11.2017, 12:14 +0100 schrieb Gert Wollny:
> Am Sonntag, den 26.11.2017, 18:01 -0800 schrieb Eric Anholt:
> >
> >
> > This looks like something that should be done with (at most) a
> > small change to opt_array_splitting.cpp, rather than reimplementing
Am Dienstag, den 28.11.2017, 08:45 +1100 schrieb Timothy Arceri:
>
>
> >
> > If I understand the comments in opt_array_splitting, arrays of
> > arrays are not properly handled there. Curretly, I don't see how I
> > can fix this, because I simply didn't look at any code there
> > before.
>
> The
Am Dienstag, den 28.11.2017, 10:53 + schrieb Eric Engestrom:
>
> > The parameters to a preprocessor macro are (basically) just
> > text. They are not evaluated until after the macro is
> > expanded. I'm 93.2% sure removing the guards around the assert()
> > should be fine in this case.
>
>
Am Mittwoch, den 29.11.2017, 14:36 +1000 schrieb Dave Airlie:
> This set of patches enables compute shaders on r600 and exposes GLSL
> 4.30 support. They are pretty alpha level, but I'd like to land some
> of them (maybe disabled) so I can avoid the rebasing fun with the
> more intrusive ones.
>
>
Am Donnerstag, den 30.11.2017, 09:30 +1000 schrieb Dave Airlie:
> On 29 November 2017 at 22:46, Gert Wollny
> wrote:
> >
> >
> > I run the arb_compute_shader piglits on BARTS, the piglits
> >
> > basic-texelfetch
> > border-color
> >
Am Donnerstag, den 30.11.2017, 17:56 +1000 schrieb Dave Airlie:
> On 30 November 2017 at 17:20, Gert Wollny
> wrote:
> > Am Donnerstag, den 30.11.2017, 09:30 +1000 schrieb Dave Airlie:
> > > On 29 November 2017 at 22:46, Gert Wollny
> > > wrote:
>
iewed-By: Gert Wollny
Am Donnerstag, den 30.11.2017, 12:16 + schrieb Eric Engestrom:
> Commit f0ba7d897d1c22202531a added this code to expose asserts to the
> compiler in an attempt to hide 'unused variable' warnings,
> incorrectly
> claiming it was a no-op. This has two
@@ -0,0 +1,202 @@
+/*
+ * Copyright © 2010 Intel Corporation
+ * Copyright © 2011 Bryan Cain
+ * Copyright © 2017 Gert Wollny
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ *
idates.
Many thanks for any comments,
Gert
Gert Wollny (3):
mesa/st: glsl_to_tgsi move some helper classes to extra files
mesa/st: glsl_to_tgsi Implement a new lifetime tracker for temporaries
mesa/st: glsl_to_tgsi: tie in the new register renaming app
This patch replaces the old register livetime estimation with the
new approach.
---
src/mesa/state_tracker/st_glsl_to_tgsi.cpp | 16 ++--
1 file changed, 14 insertions(+), 2 deletions(-)
diff --git a/src/mesa/state_tracker/st_glsl_to_tgsi.cpp
b/src/mesa/state_tracker/st_glsl_to_tgsi.
To prepare the implementation of a temp register lifetime tracker
some of the classes and functions are moved into seperate header/
implementation files to make them accessible from other files.
Specifically these are:
class st_src_reg;
class st_dst_reg;
class glsl_to_tgsi_instruction;
Hello Emil,
Am Freitag, den 16.06.2017, 15:21 +0100 schrieb Emil Velikov:
>
> Please don't use STL within core mesa code.
May I ask why? I always try to not re-implement already available
functionality and since mesa already uses C++ it seems kind of natural
to use the STL because it provides a
Hello Nicolai,
Am Sonntag, den 18.06.2017, 12:05 +0200 schrieb Nicolai Hähnle:
>
> > if HAVE_XLIB_GLX
> > SUBDIRS += drivers/x11
> > @@ -101,7 +101,7 @@ AM_CFLAGS = \
> > $(VISIBILITY_CFLAGS) \
> > $(MSVC2013_COMPAT_CFLAGS)
> > AM_CXXFLAGS = \
> > - $(LLVM_CFLAGS) \
> > +
Hello Dieter,
> W'll start with your attachment
> https://bugs.freedesktop.org/attachment.cgi?id=131683
This one is a dirty hack that works around the use of too many
registers mostly by allowing to allocate more than the allowed limit
and let the bytecode optimizer sort it out (there is also
e by better register-renaming.
The performance numbers estimated by running the shader-db are given in the
commit message of the last patch, the trend is the same like reported before.
Many thanks for any commenst,
Gert
Gert Wollny (7):
mesa/st: glsl_to_tgsi move some helper classes to extra fil
/src/mesa/state_tracker/tests/test_glsl_to_tgsi_lifetime.cpp
new file mode 100644
index 00..7e07f8868f
--- /dev/null
+++ b/src/mesa/state_tracker/tests/test_glsl_to_tgsi_lifetime.cpp
@@ -0,0 +1,969 @@
+/*
+ * Copyright © 2017 Gert Wollny
+ *
+ * Permission is hereby granted, free of charge
glsl_to_tgsi_instruction *op)
-{
- return op->info->is_tex || is_resource_instruction(op->op) ?
- op->info->num_src - 1 : op->info->num_src;
-}
glsl_to_tgsi_instruction *
glsl_to_tgsi_visitor::emit_asm(ir_instruction *ir, unsigned op,
diff --git a/src/mesa/state
The patch adds tests for the register rename mapping evaluation.
---
.../tests/test_glsl_to_tgsi_lifetime.cpp | 71 --
1 file changed, 66 insertions(+), 5 deletions(-)
diff --git a/src/mesa/state_tracker/tests/test_glsl_to_tgsi_lifetime.cpp
b/src/mesa/state_tracker/
The remapping evaluator first sorts the temporary registers ascending
based on their first life time instruction, and then uses a binary search
to find merge canidates.
For the initial sorting it uses std::sort because qsort is quite slow in
comparison. By removing the define USE_STL_SORT in
src/
For the new register renaming approach C++11 is wanted.
---
src/mesa/Makefile.am | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/src/mesa/Makefile.am b/src/mesa/Makefile.am
index 53f311d2a9..3339926d93 100644
--- a/src/mesa/Makefile.am
+++ b/src/mesa/Makefile.am
@@ -101,7 +101,
This patch ties in the new temporary register lifetime estiamtion and
rename mapping evaluation. In order to enable it, the evironment
variable MESA_GLSL_TO_TGSI_NEW_MERGE must be set.
Performance to compare between the current and the new implementation
were measured by running the shader-db in o
@@
+/*
+ * Copyright © 2017 Gert Wollny
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, co
+0200 schrieb Gert Wollny:
> Dear all,
>
> following the comments of Emil and Nicolai I've updated the patch
> set.
>
> Changes with respect to the old version are:
>
> - split the changes into more patches
> - correct formatting errors
> - remove the use o
Hi,
> > >+typedef int scope_idx;
> > >
> > > Please remove this, it's an unnecessary and distracting
> > > abstraction that doesn't gain you anything.
> >
> > Actually, with the refactoring it did the last two days it helped a
> > lot.
>
> How? Perhaps your variable names stand to be improved
any visually indications that registers would be megrged
wrongly.
Many thanks for any commenst,
Gert
** BLURB HERE ***
Gert Wollny (6):
mesa/st: glsl_to_tgsi move some helper classes to extra files
mesa/st: glsl_to_tgsi: implement new temporary register lifetime
tracker
mesa/st
The patch adds tests for the register rename mapping evaluation.
---
.../tests/test_glsl_to_tgsi_lifetime.cpp | 94 ++
1 file changed, 94 insertions(+)
diff --git a/src/mesa/state_tracker/tests/test_glsl_to_tgsi_lifetime.cpp
b/src/mesa/state_tracker/tests/test_glsl_
@@
+/*
+ * Copyright © 2017 Gert Wollny
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, co
glsl_to_tgsi_instruction *op)
-{
- return op->info->is_tex || is_resource_instruction(op->op) ?
- op->info->num_src - 1 : op->info->num_src;
-}
glsl_to_tgsi_instruction *
glsl_to_tgsi_visitor::emit_asm(ir_instruction *ir, unsigned op,
diff --git a/src/mesa/state
This patch ties in the new temporary register lifetime estiamtion and
rename mapping evaluation. In order to enable it, the evironment
variable MESA_GLSL_TO_TGSI_NEW_MERGE must be set.
Performance to compare between the current and the new implementation
were measured by running the shader-db in o
The remapping evaluator first sorts the temporary registers ascending
based on their first life time instruction, and then uses a binary search
to find merge canidates.
For the initial sorting it uses std::sort because qsort is quite slow in
comparison. By removing the define USE_STL_SORT in
src/
/test_glsl_to_tgsi_lifetime.cpp
new file mode 100644
index 00..5f3378637a
--- /dev/null
+++ b/src/mesa/state_tracker/tests/test_glsl_to_tgsi_lifetime.cpp
@@ -0,0 +1,976 @@
+/*
+ * Copyright © 2017 Gert Wollny
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
Thanks for the comments,
I've fixed these little issues locally, but I think in order to not to
spam the list, I'll send the changes later. I kind of suspect that
Nicolai might have one or the other additional comment :)
best,
Gert
___
mesa-dev mail
a few days now without noting any regressions.
As noted before, I don't have write access to mesa-git, so I'll need someone
who sponsors this patch.
Many thanks for any additional comments,
Gert
Gert Wollny (6):
mesa/st: glsl_to_tgsi move some helper classes to extra file
@@
+/*
+ * Copyright © 2017 Gert Wollny
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, co
The patch adds tests for the register rename mapping evaluation.
---
.../tests/test_glsl_to_tgsi_lifetime.cpp | 94 ++
1 file changed, 94 insertions(+)
diff --git a/src/mesa/state_tracker/tests/test_glsl_to_tgsi_lifetime.cpp
b/src/mesa/state_tracker/tests/test_glsl_
101 - 200 of 827 matches
Mail list logo
