When RADV_DEBUG=shaders is set, printing e.g. different NIR shaders from
different threads at the same time makes the output unreadable. Use a mutex
to synchronize shader dumping so that all shaders get printed in once piece.
Since we're writing everything to a file or terminal anyway, the
performance impact of forcing singlethreaded compilation is negligible.
Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25215>
out_args->scratch_offset and in_wg_id_x will alias on <gfx9.
To avoid the conversion code reading a garbage WG ID, move the
scratch/ring offset writing to the very end.
Fixes: 1e354172 ("radv,aco: Convert 1D ray launches to 2D")
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/30707>
This just adds support for allowing worst case image sizing with no
specified profiles and for using inline queries.
Reviewed-by: Lynne <dev@lynne.ee>
Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/30671>
This handles the new inlines queries for VK_KHR_video_maintenance1
(Decode doesn't do queries on AMD at the moment).
Reviewed-by: Lynne <dev@lynne.ee>
Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/30671>
This is a hint that the branch is worth removing. Assume that's the case,
regardless of the gfx level.
fossil-db (vega10):
Totals from 22 (0.03% of 63053) affected shaders:
Instrs: 23927 -> 23856 (-0.30%)
CodeSize: 125096 -> 124812 (-0.23%)
Latency: 138258 -> 137765 (-0.36%)
InvThroughput: 55900 -> 55884 (-0.03%)
Branches: 391 -> 320 (-18.16%)
Signed-off-by: Rhys Perry <pendingchaos02@gmail.com>
Reviewed-by: Daniel Schürmann <daniel@schuermann.dev>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/30321>
No fossil-db changes.
Signed-off-by: Rhys Perry <pendingchaos02@gmail.com>
Reviewed-by: Timur Kristóf <timur.kristof@gmail.com>
Reviewed-by: Daniel Schürmann <daniel@schuermann.dev>
Backport-to: 24.1
Backport-to: 24.2
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/30321>
DGC preprocessing is disabled for VBOs, so the dynamic vertex input
should always be initialized. Null VBO descriptors are slightly
different when used with dynamic VS inputs.
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/29831>
This feature is marginal but VKCTS recently got coverage and most of
the implementations were broken, including RADV.
Fixes recent tests
dEQP-VK.pipeline.pipeline_library.graphics_library.misc.other.view_index_from_device_index_.*
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/30642>
Decompressing HTILE should also reset the HTILE metadata to initial
state which means that re-initializing it for non-compressed to
compressed transitions is redundant.
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/30122>
Resolves:
[328/4125] Compiling C++ object src/amd/addrlib/libaddrlib.a.p/src_core_addrlib1.cpp.o
In static member function 'static VOID Addr::Object::ClientFree(VOID*, const Addr::Client*)',
inlined from 'static VOID Addr::Object::operator delete(VOID*)' at ../src/amd/addrlib/src/core/addrobject.cpp:190:15,
inlined from 'virtual Addr::Object::~Object()' at ../src/amd/addrlib/src/core/addrobject.cpp:71:1:
../src/amd/addrlib/src/core/addrobject.cpp:129:28: error: '*(const Addr::Client*)((char*)this + 8).Addr::Client::callbacks._ADDR_CALLBACKS::freeSysMem' is used uninitialized [-Werror=uninitialized]
129 | if (pClient->callbacks.freeSysMem != NULL)
| ~~~~~~~~~~~~~~~~~~~^~~~~~~~~~
Closes: https://gitlab.freedesktop.org/mesa/mesa/-/issues/11634
Suggested-by: Marek Olšák <marek.olsak@amd.com>
Signed-off-by: David Heidelberg <david@ixit.cz>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/30482>
If we can ignore the sign of the derivative, we can swap the lanes
instead of broadcasting per direction. abs(a - b) = abs(b - a).
Shamelessly copied from bifrost.
Foz-DB Navi31:
Totals from 5 (0.01% of 79206) affected shaders:
Instrs: 6191 -> 6184 (-0.11%)
CodeSize: 31960 -> 31920 (-0.13%)
Latency: 111961 -> 111926 (-0.03%)
InvThroughput: 10390 -> 10372 (-0.17%)
VALU: 3286 -> 3279 (-0.21%)
Reviewed-by: Rhys Perry <pendingchaos02@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/30557>
