[Mesa-dev] [PATCH 112/133] nir: Add an algebraic optimization pass

Matt Turner mattst88 at gmail.com
Mon Jan 5 23:38:33 PST 2015


On Mon, Dec 15, 2014 at 10:12 PM, Jason Ekstrand <jason at jlekstrand.net> wrote:
> This pass uses the previously built algebraic transformations framework and
> should act as an example for anyone else wanting to make an algebraic
> transformation pass for NIR.
> ---
>  src/glsl/Makefile.am                     | 10 ++++-
>  src/glsl/Makefile.sources                |  6 ++-
>  src/glsl/nir/nir.h                       |  2 +
>  src/glsl/nir/nir_opt_algebraic.py        | 67 ++++++++++++++++++++++++++++++++
>  src/mesa/drivers/dri/i965/brw_fs_nir.cpp |  2 +-
>  5 files changed, 84 insertions(+), 3 deletions(-)
>  create mode 100644 src/glsl/nir/nir_opt_algebraic.py
>
> diff --git a/src/glsl/Makefile.am b/src/glsl/Makefile.am
> index 7f62573..bddec9c 100644
> --- a/src/glsl/Makefile.am
> +++ b/src/glsl/Makefile.am
> @@ -25,6 +25,7 @@ AM_CPPFLAGS = \
>         -I$(top_srcdir)/src/mapi \
>         -I$(top_srcdir)/src/mesa/ \
>         -I$(top_srcdir)/src/glsl/glcpp \
> +       -I$(top_srcdir)/src/glsl/nir \
>         -I$(top_srcdir)/src/gtest/include \
>         $(DEFINES)
>  AM_CFLAGS = $(VISIBILITY_CFLAGS)
> @@ -205,7 +206,8 @@ BUILT_SOURCES =                                             \
>         glsl_parser.cpp                                 \
>         glsl_lexer.cpp                                  \
>         glcpp/glcpp-parse.c                             \
> -       glcpp/glcpp-lex.c
> +       glcpp/glcpp-lex.c                               \
> +        nir/nir_opt_algebraic.c

Tabs in Makefile.am.

>  CLEANFILES =                                           \
>         glcpp/glcpp-parse.h                             \
>         glsl_parser.h                                   \
> @@ -217,3 +219,9 @@ clean-local:
>  dist-hook:
>         $(RM) glcpp/tests/*.out
>         $(RM) glcpp/tests/subtest*/*.out
> +
> +nir/nir_opt_algebraic.c: nir/nir_opt_algebraic.py nir/nir_algebraic.py
> +       $(AM_V_GEN)set -e;                                                      \

I don't think you need this?

> +       $(MKDIR_P) `dirname $@`;                                                \

Just "$(MKDIR_P) nir" I think? I don't think you need the line
continuations either?

> +       $(PYTHON2) $(PYTHON_FLAGS) $(srcdir)/nir/nir_opt_algebraic.py > $@.tmp; \
> +       mv $@.tmp $@;

Just generate the file directly?

> diff --git a/src/glsl/Makefile.sources b/src/glsl/Makefile.sources
> index 201330d..d243f12 100644
> --- a/src/glsl/Makefile.sources
> +++ b/src/glsl/Makefile.sources
> @@ -13,6 +13,9 @@ LIBGLCPP_GENERATED_FILES = \
>         $(GLSL_BUILDDIR)/glcpp/glcpp-lex.c \
>         $(GLSL_BUILDDIR)/glcpp/glcpp-parse.c
>
> +NIR_GENERATED_FILES = \
> +       $(GLSL_BUILDDIR)/nir/nir_opt_algebraic.c
> +
>  NIR_FILES = \
>         $(GLSL_SRCDIR)/nir/nir.c \
>         $(GLSL_SRCDIR)/nir/nir.h \
> @@ -47,7 +50,8 @@ NIR_FILES = \
>         $(GLSL_SRCDIR)/nir/nir_to_ssa.c \
>         $(GLSL_SRCDIR)/nir/nir_validate.c \
>         $(GLSL_SRCDIR)/nir/nir_types.cpp \
> -       $(GLSL_SRCDIR)/nir/glsl_to_nir.cpp
> +       $(GLSL_SRCDIR)/nir/glsl_to_nir.cpp \
> +        $(NIR_GENERATED_FILES)

Use tabs

>
>  # libglsl
>
> diff --git a/src/glsl/nir/nir.h b/src/glsl/nir/nir.h
> index b3c61ca..630e841 100644
> --- a/src/glsl/nir/nir.h
> +++ b/src/glsl/nir/nir.h
> @@ -1402,6 +1402,8 @@ void nir_convert_to_ssa_impl(nir_function_impl *impl);
>  void nir_convert_to_ssa(nir_shader *shader);
>  void nir_convert_from_ssa(nir_shader *shader);
>
> +bool nir_opt_algebraic(nir_shader *shader);
> +
>  bool nir_opt_global_to_local(nir_shader *shader);
>
>  bool nir_copy_prop_impl(nir_function_impl *impl);
> diff --git a/src/glsl/nir/nir_opt_algebraic.py b/src/glsl/nir/nir_opt_algebraic.py
> new file mode 100644
> index 0000000..e829f37
> --- /dev/null
> +++ b/src/glsl/nir/nir_opt_algebraic.py
> @@ -0,0 +1,67 @@
> +#! /usr/bin/env python
> +#
> +# Copyright (C) 2014 Intel Corporation
> +#
> +# 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, copy, modify, merge, publish, distribute, sublicense,
> +# and/or sell copies of the Software, and to permit persons to whom the
> +# Software is furnished to do so, subject to the following conditions:
> +#
> +# The above copyright notice and this permission notice (including the next
> +# paragraph) shall be included in all copies or substantial portions of the
> +# Software.
> +#
> +# THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
> +# IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
> +# FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
> +# THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
> +# LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
> +# FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
> +# IN THE SOFTWARE.
> +#
> +# Authors:
> +#    Jason Ekstrand (jason at jlekstrand.net)
> +
> +import nir_algebraic
> +
> +# Convenience variables
> +a = 'a'
> +b = 'b'
> +c = 'c'
> +d = 'd'
> +
> +# Written in the form (<search>, <replace>) where <search> is an expression
> +# and <replace> is either an expression or a value.  An expression is
> +# defined as a tuple of the form (<op>, <src0>, <src1>, <src2>, <src3>)
> +# where each source is either an expression or a value.  A value can be
> +# either a numeric constant or a string representing a variable name.  For
> +# constants, you have to be careful to make sure that it is the right type
> +# because python is unaware of the source and destination types of the
> +# opcodes.
> +
> +optimizations = [
> +   (('fadd', a, 0.0), a),
> +   (('iadd', a, 0), a),
> +   (('fmul', a, 0.0), 0.0),
> +   (('imul', a, 0), 0),
> +   (('fmul', a, 1.0), a),
> +   (('imul', a, 1), a),
> +   (('fmul', a, -1.0), ('fneg', a)),
> +   (('imul', a, -1), ('ineg', a)),
> +   (('ffma', 0.0, a, b), b),
> +   (('ffma', a, 0.0, b), b),
> +   (('ffma', a, b, 0.0), ('fmul', a, b)),
> +   (('flrp', a, b, 0.0), a),
> +   (('flrp', a, b, 1.0), b),
> +   (('flrp', a, a, b), a),
> +   (('flrp', 0.0, a, b), ('fmul', a, b)),
> +   (('fadd', ('fmul', a, b), c), ('ffma', a, b, c)),
> +   (('fge', ('fneg', ('fabs', a)), 0.0), ('feq', a, 0.0)),
> +   (('fmin', ('fmax', a, 1.0), 0.0), ('fsat', a)),
> +# This one may not be exact
> +   (('feq', ('fadd', a, b), 0.0), ('feq', a, ('fneg', b))),
> +]
> +
> +print nir_algebraic.AlgebraicPass("nir_opt_algebraic", optimizations).render()
> diff --git a/src/mesa/drivers/dri/i965/brw_fs_nir.cpp b/src/mesa/drivers/dri/i965/brw_fs_nir.cpp
> index ec1ef28..039dd4d 100644
> --- a/src/mesa/drivers/dri/i965/brw_fs_nir.cpp
> +++ b/src/mesa/drivers/dri/i965/brw_fs_nir.cpp
> @@ -53,7 +53,7 @@ fs_visitor::emit_nir_code()
>        nir_validate_shader(nir);
>        progress |= nir_opt_peephole_select(nir);
>        nir_validate_shader(nir);
> -      progress |= nir_opt_peephole_ffma(nir);
> +      progress |= nir_opt_algebraic(nir);
>        nir_validate_shader(nir);
>     } while (progress);
>
> --
> 2.2.0

With those fixed,

Reviewed-by: Matt Turner <mattst88 at gmail.com>


More information about the mesa-dev mailing list