This should bring the overall execution time at around 15 minutes \o/,
but it exposes more issues in the driver:
- [...].linear_interpolation.no_offset_8_samples (#7220)
- Timeouts on some dedicated_alloc.image.full.1048577.* tests
- Timeout on memory.pipeline_barrier.host_read_host_write
More performance improvements should come later, as the load average
remains in the [12, 16] range when running on my machine on a navi22
gets me a 31 load average and a lower execution time by a couple of
minutes.
References: #7220
Signed-off-by: Martin Roukala (né Peres) <martin.roukala@mupuf.org>
Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19111>
The macro guard of #if defined(_WIN32) && !defined(__CYGWIN__) is comes from yohhoy's implementation
and that's not consistence with util/u_uthread.h, this caused it's hard to understand.
Now we change the behavior that's always rely on how meson detecting HAVE_PTHREAD.
So we always disable detecting of threads on Win32 as it's always included in the kernel32 library to
avoid detecting pthreads in mingw.
Signed-off-by: Yonggang Luo <luoyonggang@gmail.com>
Reviewed-by: Jesse Natalie <jenatali@microsoft.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18425>
That was previously listed on the getopt_long struct but not actually
being used. This makes intel_clc argument processing easier as now
all of its arguments are handled with getopt and anything after the
special argument '--' is passed along to clang to form the final build
command.
Thanks to Dylan Baker for help with changes to the meson file.
Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19153>
The error when include simple_mtx.h in glsl_types.h:
../src/microsoft/compiler/dxil_function.c
In file included from ../src/compiler/glsl_types.h:37,
from ../src/compiler/nir_types.h:36,
from ../src/compiler/nir/nir.h:44,
from ../src/microsoft/compiler/dxil_signature.h:28,
from ../src/microsoft/compiler/dxil_module.h:37,
from ../src/microsoft/compiler/dxil_function.h:48,
from ../src/microsoft/compiler/dxil_function.c:24:
../src/util/simple_mtx.h:37:12: fatal error: valgrind.h: No such file or directory
37 | # include <valgrind.h>
Signed-off-by: Yonggang Luo <luoyonggang@gmail.com>
Acked-by: Jesse Natalie <jenatali@microsoft.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19154>
The ideal would be to use a D3D12Core.dll next to the GL DLL, via
ID3D12SDKConfiguration1, but this requires a new D3D12.dll to be
present on the machine, which only very new OSes will currently have.
For older OSes, we can at least try to use ID3D12SDKConfiguration to
set a global Agility SDK override, but that only works in developer
mode and requires the DLL to be at a relative path from the app EXE,
so only try that if we have environment variables telling us to do so.
Reviewed-by: Sil Vilerino <sivileri@microsoft.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18305>
As explained in a comment, a device factory allows modification of state
like debug layer enablement without touching global state / triggering
device removal on a singleton device, and can even create devices that are
not singletons if the driver supports that.
Reviewed-by: Sil Vilerino <sivileri@microsoft.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18305>
The GLSL singleton ref was too late, and we could end up with an un-paired
dec-ref if screen init failed. Also the transfer slab init was in the wrong
init function, so we'd re-init that struct if the screen was re-initialized.
Reviewed-by: Sil Vilerino <sivileri@microsoft.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18305>
VK_PIPELINE_CREATE_RAY_TRACING_SKIP_AABBS_BIT_KHR and
VK_PIPELINE_CREATE_RAY_TRACING_SKIP_TRIANGLES_BIT_KHR, when specified,
make TraceRay behave as if the corresponding shader flags were set, but
without affecting the value of IncomingRayFlags in shaders.
v2 (Lionel): Improve comments
Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19152>
So far `lengths` as been interpreted as a slice of usize. However, 0 is specified as a sentinel
value signalling that the corresponding string is nul terminated. Since checking for sentinel
values is frequently forgotten, Option types should be preferred if possible.
Option<NonZeroUsize> is layout compatible with usize. The None variant is then represented as 0,
which is exaclty what we need here.
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18978>
The code assumed that if the length of a string was specified and greater than zero, the string
would not contain a nul byte.
Legal or not, there are apparently applications which violate that assumption. Since the spec
doesn't say anything about this case, take the likely most compatible route and treat a nul byte as
terminating the string early.
Closes: https://gitlab.freedesktop.org/mesa/mesa/-/issues/7408
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18978>
Tomeu Vizoso