Following on the previous commit, this commit adds support for selecting
and reporting constants to pull from the FAU instead of loading them
into registers first with ADD_IMM. This is beneficial because we can then
use them as a source directly and save ourselves one instruction to move
them into the register first.
Reviewed-by: Mary Guillemard <mary.guillemard@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/36872>
That pass tries to remove blocks that are not needed:
blocks with only 1 predecessor and 1 successor containing no
instruction or only 1 branch instruction.
Also remove unnecessary branch at the end of block with only 1
successor which is the next block
Run this pass at the end of the compiler flow once all optimisations
have been applied.
Acked-by: Erik Faye-Lund <erik.faye-lund@collabora.com>
Reviewed-by: Iago Toral Quiroga <itoral@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/36021>
In the C23 standard unreachable() is now a predefined function-like
macro in <stddef.h>
See https://android.googlesource.com/platform/bionic/+/HEAD/docs/c23.md#is-now-a-predefined-function_like-macro-in
And this causes build errors when building for C23:
-----------------------------------------------------------------------
In file included from ../src/util/log.h:30,
from ../src/util/log.c:30:
../src/util/macros.h:123:9: warning: "unreachable" redefined
123 | #define unreachable(str) \
| ^~~~~~~~~~~
In file included from ../src/util/macros.h:31:
/usr/lib/gcc/x86_64-linux-gnu/14/include/stddef.h:456:9: note: this is the location of the previous definition
456 | #define unreachable() (__builtin_unreachable ())
| ^~~~~~~~~~~
-----------------------------------------------------------------------
So don't redefine it with the same name, but use the name UNREACHABLE()
to also signify it's a macro.
Using a different name also makes sense because the behavior of the
macro was extending the one of __builtin_unreachable() anyway, and it
also had a different signature, accepting one argument, compared to the
standard unreachable() with no arguments.
This change improves the chances of building mesa with the C23 standard,
which for instance is the default in recent AOSP versions.
All the instances of the macro, including the definition, were updated
with the following command line:
git grep -l '[^_]unreachable(' -- "src/**" | sort | uniq | \
while read file; \
do \
sed -e 's/\([^_]\)unreachable(/\1UNREACHABLE(/g' -i "$file"; \
done && \
sed -e 's/#undef unreachable/#undef UNREACHABLE/g' -i src/intel/isl/isl_aux_info.c
Reviewed-by: Erik Faye-Lund <erik.faye-lund@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/36437>
We have a lot of pattern like this on Valhall:
FCMP_OR.f32.lt.m1 r28, ^r28, r27, 0x0
FCMP_OR.f32.lt.m1 r29, r27, r25, 0x0
LSHIFT_AND.i32 r28, ^r28, 0x0.b00, ^r29
That can be simplified into:
FCMP_OR.f32.lt.m1 r29, r27, r25, 0x0
FCMP_AND.f32.lt.m1 r28, ^r28, r27, ^r29
This pass merge those specific cases while setting the appropriate
logical variant of the CMP instruction.
We do not try to merge the srcs that do not originate from a matching CMP
instruction with matching result type as the logical operation is
performed before the result type transformation.
Now this is enough to optimise a lot of common cases anyway so it is
still a win.
Results on fossils/sascha-willems:
Totals:
Instrs: 42157 -> 42059 (-0.23%)
CodeSize: 582784 -> 581760 (-0.18%)
Estimated normalized SFU cycles: 159.9375 -> 153.75 (-3.87%)
Totals from 13 (2.07% of 627) affected shaders:
Instrs: 3490 -> 3392 (-2.81%)
CodeSize: 29696 -> 28672 (-3.45%)
Estimated normalized SFU cycles: 15.8125 -> 9.625 (-39.13%)
Signed-off-by: Mary Guillemard <mary.guillemard@collabora.com>
Reviewed-by: Eric R. Smith <eric.smith@collabora.com>
Reviewed-by: Olivia Lee <olivia.lee@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/36327>
Before doing register allocation, use information available from
the SSA representation to determine register pressure and to
spill registers. This spilling doesn't have to be perfect (the
register allocator is still allowed to spill) but it will be
much faster to do the SSA spilling than RA spilling. In general
this should vastly improve the performance of register allocation.
Reviewed-by: Mary Guillemard <mary.guillemard@collabora.com>
Acked-by: Boris Brezillon <boris.brezillon@collabora.com>
Acked-by: Erik Faye-Lund <erik.faye-lund@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/34446>
The MEMORY_BARRIER instruction has some issues, where we end up
dead-code eliminating it before it gets to do what it's supposed to do.
But even if we fix that, we have issues where we can end up inserting
flow control into it, which isn't going to work because we have nothing
to emit here either.
So let's rework this to a special-cased NOP instruction, which is marked
as a scheduling barrier. The beneft here is that NOPs are already properly
handled when it comes to flow control.
Note that this isn't perfect either; this only prevents memory operations
from crossing the scheduling barrier. We should really prevent any
operation with observable side effects from crossing the barrier. This
includes things like reading clocks etc.
But that's a larger change, and it's a step in the right direction to get
this to no longer be dead-code eliminated. So let's put this band-aid on
for now.
Fixes: f77a50e45e ("pan/bi: add a MEMORY_BARRIER pseudo-instruction")
Reviewed-by: Caterina Shablia <caterina.shablia@collabora.com>
Reviewed-by: Olivia Lee <olivia.lee@collabora.com>
Reviewed-by: Boris Brezillon <boris.brezillon@collabora.com>
Reviewed-by: Lars-Ivar Hesselberg Simonsen <lars-ivar.simonsen@arm.com>
Reviewed-by: Mary Guillemard <mary.guillemard@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/35502>
Primary reason to do this is to make codegen using the swizzle names in
bifrost/ISA.xml simpler. A secondary benefit is that dependent code can
now use the swizzle name that matches the context, making things a
little more readable.
We may want to consider giving widens separate values later, so that
va_lower_constants and bi_opt_constant_fold can fold them correctly, but
I don't know of current bugs caused by this.
Signed-off-by: Benjamin Lee <benjamin.lee@collabora.com>
Reviewed-by: Eric R. Smith <eric.smith@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/33416>
On Valhall we can optimize lower waits, which waits for both readers and
writers, into resource_waits which only wait for writers, allowing
threads accessing read-only resources to execute concurrently.
Let's use that on LD_TILE instructions so we can optmize the read-only
case.
Signed-off-by: Boris Brezillon <boris.brezillon@collabora.com>
Reviewed-by: Lars-Ivar Hesselberg Simonsen <lars-ivar.simonsen@arm.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/32540>
This is the compiler for both Bifrost and Valhall, and presumably future
Mali GPUs too. Give it a more generic name so we can use the bifrost/ path for
something a bit more specific.
For historical reasons the compiler's name is still "bifrost" and uses the
prefix `bi_`. I think that's ok in the same way that i915 in the kernel supports
way more than just i915.
Signed-off-by: Alyssa Rosenzweig <alyssa@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/20455>
Christoph Pillmayer