This is a v2 of the series that I sent out a month or two ago to fix up the
LOAD_PAYLOAD instruction in the i965 FS backend compiler. This new version
incorporates comments from a variety of people.
The last patch in the series is one that I think at least 3 of us have
independantly written to enable copy-propagation from the ATTR register
file. This helps the old fs_visitor code some but it helps NIR a *lot*.
The end result is that NIR is now on par from a shader-db perspective with
the old fs_visitor code.
Shader-db results across all but the last patch:
total instructions in shared programs: 7440514 -> 7440378 (-0.00%)
instructions in affected programs: 11954 -> 11818 (-1.14%)
helped: 68
HURT: 0
GAINED: 3
LOST: 0
All 68 helped shaders are in unigen heaven ultra 4xmsaa. It appears that
it helps us to copy-propagate uses of omask. The omask is weird because
it's a single-register header that's a 16-wide UW type. The previous
guessing code was actually getting it wrong and copying it with an 8-wide
UD move that was not force_writemask all. Now it is getting copied with
force_writemask_all set correctly (because it's a header source). It seems
like copy propagation can now come through and clean it up. (At least
that's my theory; I didn't look at it that hard.)
Francisco Jerez (1):
i965/fs: Fix passing an immediate to half().
Jason Ekstrand (13):
i965/fs: Make half(fs_reg, unsigned) handle register files more
explicitly
i965/fs_cse: Factor out code to create copy instructions
i965: Change header_present to header_size in backend_instruction
i965/fs_inst: Add an is_copy_payload helper
i965/fs: Make emit_single_fb_write take an explicit exec_size
i965/fs: Make LOAD_PAYLOAD take a header size
i965/fs: Rework fs_visitor::LOAD_PAYLOAD
SQUASH: i965/fs: Make destinations of load_payload have the
appropreate width
SQUASH: i965/fs: Rework fs_visitor::lower_load_payload
SQUASH: i965/fs_cse: Support the new-style LOAD_PAYLOAD
SQUASH: i965/fs_inst::is_copy_payload: Support the new-style
LOAD_PAYLOAD
SQUASH: i965/fs: Simplify setup_color_payload
i965/fs_inst: Get rid of the effective_width field
Kenneth Graunke (1):
i965/fs: Allow copy propagation on ATTR file registers.
src/mesa/drivers/dri/i965/brw_blorp_blit_eu.cpp | 4 +-
src/mesa/drivers/dri/i965/brw_fs.cpp | 270 +++++++++++----------
src/mesa/drivers/dri/i965/brw_fs.h | 9 +-
.../drivers/dri/i965/brw_fs_copy_propagation.cpp | 12 +-
src/mesa/drivers/dri/i965/brw_fs_cse.cpp | 117 ++++-----
src/mesa/drivers/dri/i965/brw_fs_generator.cpp | 20 +-
.../drivers/dri/i965/brw_fs_register_coalesce.cpp | 17 +-
src/mesa/drivers/dri/i965/brw_fs_visitor.cpp | 237 +++++++-----------
src/mesa/drivers/dri/i965/brw_ir_fs.h | 30 ++-
src/mesa/drivers/dri/i965/brw_shader.h | 4 +-
src/mesa/drivers/dri/i965/brw_vec4.cpp | 2 +-
src/mesa/drivers/dri/i965/brw_vec4_generator.cpp | 10 +-
src/mesa/drivers/dri/i965/brw_vec4_visitor.cpp | 16 +-
13 files changed, 338 insertions(+), 410 deletions(-)
--
2.3.6
_______________________________________________
mesa-dev mailing list
mesa-dev@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/mesa-dev
