brw_hw_type_to_reg_type() needs to know only whether the file is
BRW_IMMEDIATE_VALUE or not, which is not a valid file for the
destination. gcc and clang will evaluate __builtin_strcmp() at compile
time, so we can use it to pass a constant file for the destination.
text data bss dec hex filename
7816214 346248 420496 8582958 82f72e i965_dri.so before
7816070 346248 420496 8582814 82f69e i965_dri.so after
Reviewed-by: Scott D Phillips <scott.d.phillips@intel.com>
text data bss dec hex filename
7816886 346248 420496 8583630 82f9ce i965_dri.so before
7816214 346248 420496 8582958 82f72e i965_dri.so after
Reviewed-by: Scott D Phillips <scott.d.phillips@intel.com>
Previously the brw_inst{,_set}_{dst,src0,src1}_reg_type() functions
provided access to the hardware encodings for the register types. We
often mixed these with the logical BRW_REGISTER_TYPE_* enums (which
themselves used to be the hardware format!) with bad results.
With that functionality now available with the hw_ versions (see
previous commit), we now add functions that take the logical
BRW_REGISTER_TYPE_* enums and convert into the hardware format and vice
versa. To do the conversion we also have to provide the file.
Note the asymmetry between the two functions: the new getter reads the
file from the instruction word, and to ensure that is always set the
setter writes both the file and the type.
Reviewed-by: Scott D Phillips <scott.d.phillips@intel.com>
I'm going to encapsulate all of the logic dealing with register types in
this file.
Rename the parameters for the hardware encodings from type -> hw_type at
the same time.
Reviewed-by: Scott D Phillips <scott.d.phillips@intel.com>
After the last patch converted things into enums, I helpfully got a
compiler warning about these missing from the switch statement.
Reviewed-by: Scott D Phillips <scott.d.phillips@intel.com>
The hardware encodings often mean different things depending on whether
the source is an immediate.
Reviewed-by: Scott D Phillips <scott.d.phillips@intel.com>
These vaguely corresponded to the hardware encodings, but that is purely
historical at this point. Reorder them so we stop making things "almost
work" when mixing enums.
The ordering has been closen so that no enum value is the same as a
compatible hardware encoding.
Reviewed-by: Scott D Phillips <scott.d.phillips@intel.com>
UB and B type encodings are the same as UV and VF. Noticed when writing
the following patch.
Reviewed-by: Scott D Phillips <scott.d.phillips@intel.com>
The destination stride must be equivalent to a dword if VF is used.
Also, since the only compaction table entires with "i:vf" have the
destination as "r:f" specifically check that the destination is of type
float.
Reviewed-by: Scott D Phillips <scott.d.phillips@intel.com>
Note that there's no point in testing on G45, since its compaction is
the same as Gen5. Same logic applies to Gen7 variants and low-power
parts.
Reviewed-by: Scott D Phillips <scott.d.phillips@intel.com>
Mesa will map user defined vertex input attributes to slots
starting at VERT_ATTRIB_GENERIC0 which gives us room for only 16
slots (up to GL_VERT_ATTRIB_MAX). This sufficient for GL, where
we expose exactly 16 vertex attributes for user defined inputs, but
in Vulkan we can expose up to 28 (which are also mapped from
VERT_ATTRIB_GENERIC0 onwards) so we need to account for this when
we scope the size of the array of attribute workaround flags
that is used during the brw_vertex_workarounds NIR pass. This
prevents out-of-bounds accesses in that array for NIR shaders
that use more than 16 vertex input attributes.
Fixes:
dEQP-VK.pipeline.vertex_input.max_attributes.*
Acked-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
If dual object compile fails (as seems to happen with virgl a
fair bit, and does piglit even have any tests for it?), we end up
not restarting the pull params, so we call
vec4_visitor::move_uniform_array_access_to_pull_constant
a second time and it runs over the ends of the alloc.
Fixes: tests/spec/glsl-1.50/execution/geometry/max-input-components.shader_test
running inside virgl on ivybridge.
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Cc: <mesa-stable@lists.freedesktop.org>
Signed-off-by: Dave Airlie <airlied@redhat.com>
Some hardware, like i965, doesn't support group sizes greater than 32.
In that case, we can reduce the destination size of the ballot
intrinsic, which will simplify our code generation.
Reviewed-by: Connor Abbott <cwabbott0@gmail.com>
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
The implementation of ballotARB() will start by zeroing the flags
register. So, a doing something like
if (gl_SubGroupInvocationARB % 2u == 0u) {
... = ballotARB(true);
[...]
} else {
... = ballotARB(true);
[...]
}
(like fs-ballot-if-else.shader_test does) would generate identical MOVs
to the same destination (the flag register!), and we definitely do not
want to pull that out of the control flow.
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
The implementations of the ARB_shader_ballot intrinsics will explicitly
read the flag as a source register.
Reviewed-by: Matt Turner <mattst88@gmail.com>
We already had a channel_num system value, which I'm renaming to
subgroup_invocation to match the rest of the new system values.
Note that while ballotARB(true) will return zeros in the high 32-bits on
systems where gl_SubGroupSizeARB <= 32, the gl_SubGroup??MaskARB
variables do not consider whether channels are enabled. See issue (1) of
ARB_shader_ballot.
Reviewed-by: Connor Abbott <cwabbott0@gmail.com>
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
The implementations of the ARB_shader_group_vote intrinsics will
explicitly write the flag as the destination register.
Reviewed-by: Matt Turner <mattst88@gmail.com>
I don't expect anyone is going to care about using this in vec4 programs
(vertex/tessellation/geometry on Gen6/7), no one has come up with a good
way to implement it much less test it.
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Specifically, constant fold intrinsics from ARB_shader_group_vote, but I
suspect it'll be useful for other things in the future.
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
This actually takes advantage of the newly pushed UBO data, avoiding
pull loads.
Improves performance in GLBenchmark Manhattan 3.1 by:
HSW: ~1%, BDW/SKL/KBL GT2: 3-4%, SKL GT4: 7-8%, APL: 4-5%.
(thanks to Eero Tamminen for these numbers)
shader-db results on Skylake, ignoring programs with spill/fill changes:
total instructions in shared programs: 13963994 -> 13651893 (-2.24%)
instructions in affected programs: 4250328 -> 3938227 (-7.34%)
helped: 28527
HURT: 0
total cycles in shared programs: 179808608 -> 172535170 (-4.05%)
cycles in affected programs: 79720410 -> 72446972 (-9.12%)
helped: 26951
HURT: 1248
LOST: 46
GAINED: 21
Many "Deus Ex: Mankind Divided" shaders which already spilled end up
spill a lot more (about 240 programs hurt, 9 helped). The cycle
estimator suggests this is still overall a win (-0.23% in cycle counts)
presumably because we trade pull loads for fills.
v2: Drop "PULL" environment variable left in for initial debugging
(caught by Matt).
Reviewed-by: Matt Turner <mattst88@gmail.com>
With UBOs, the answer of "have we decided to push this uniform" gets
a bit more complicated - for one, we have multiple surfaces. This
patch refactors things so we can add the new code in a single place.
Reviewed-by: Matt Turner <mattst88@gmail.com>
This patch starts uploading UBO data via 3DSTATE_CONSTANT_* packets,
and updates the compiler to know that there's extra payload data, so
things continue working. However, it still issues pull loads for all
data. I wanted to separate the two aspects for greater bisectability.
v2: Update for new intel_bufferobj_buffer parameter.
Reviewed-by: Matt Turner <mattst88@gmail.com>
This adds a NIR pass that decides which portions of UBOS we should
upload as push constants, rather than pull constants.
v2: Switch to uint16_t for the UBO block number, because we may
have a lot of them in Vulkan (suggested by Jason). Add more
comments about bitfield trickery (requested by Matt).
v3: Skip vec4 stages for now...I haven't finished wiring up support
in the vec4 backend, and so pushing the data but not using it
will just be wasteful.
Reviewed-by: Matt Turner <mattst88@gmail.com>
By default, 3DSTATE_CONSTANT_* Constant Buffer 0 is relative to dynamic
state base address. This makes it unusable for pushing UBOs. I'd like
to be able to use all four push buffers.
There is a bit in the INSTPM register (or CS_DEBUG_MODE2 on Skylake)
which controls whether buffer 0 is relative to dynamic state base
address, or simply a normal pointer. Setting that gives us full
flexibility.
We can't currently write this on Haswell and earlier, and will need
to update the kernel command parser, and then do the whole version
checking song and dance.
Reviewed-by: Matt Turner <mattst88@gmail.com>
