Stuff COMPUTE_WALKER_BODY in COMPUTER_WALKER in both iris and anv.
This also fixes the tracepoint for ray dispatches. Stuffing
COMPUTE_WALKER_BODY allow us to set the
cmd_buffer->state.last_compute_walker.
Signed-off-by: Sagar Ghuge <sagar.ghuge@intel.com>
Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/31822>
Should only impact Xe2+
Signed-off-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Fixes: 02294961ee ("anv: stop using a binding table entry for gl_NumWorkgroups")
Reviewed-by: Caio Oliveira <caio.oliveira@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/31799>
We're not using a binding table entry anymore for num_workgroups.
Signed-off-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Fixes: 02294961ee ("anv: stop using a binding table entry for gl_NumWorkgroups")
Reviewed-by: Caio Oliveira <caio.oliveira@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/31799>
This will make things easier in situations where we don't want to use
the binding table at all (indirect draws/dispatches).
The mechanism is simple, upload a vec3 either through push constants
(<= Gfx12.0) or through the inline parameter register (>= Gfx12.5).
In the shader, do this :
if vec.x == 0xffffffff:
addr = pack64_2x32 vec.y, vec.z
vec = load_global addr
This works because we limit the maximum number of workgroup size to
0xffff in all dimension :
maxComputeWorkGroupCount = { 65535, 65535, 65535 },
So we can use the large values to signal the need for indirect
loading.
Signed-off-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Reviewed-by: Caio Oliveira <caio.oliveira@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/31508>
GFX_VER block matches both workarounds and while these workarounds are
almost about the same cause, other one applies only for LNL and other
one for BMG, need to check for both.
Signed-off-by: Tapani Pälli <tapani.palli@intel.com>
Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/31571>
This is also encouraged by another wa, Wa_14018813551.
Both workarounds state that StackIDControlOverride_RTGlobals should
always be set to 0 (i.e. 2k).
Signed-off-by: Tapani Pälli <tapani.palli@intel.com>
Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/30937>
The PRMs suggest that certain classes of auxiliary surface operations
will automatically synchronize when performed back-to-back:
Any transition from any value in {Clear, Render, Resolve} to a
different value in {Clear, Render, Resolve} requires end of pipe
synchronization.
Make use of this functionality by batching CCS and MCS flushes when
compatible auxiliary surface operations are performed within a command
buffer.
Ref: https://gitlab.freedesktop.org/mesa/mesa/-/issues/11325
Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/29922>
This field encodes bits [27:6] of the scratch surface state offset
according to the hardware spec, already on XeHP platforms. However,
on previous platforms we were passing bits [25:4] instead, which was
apparently okay for two reasons:
1/ We never used more than 8 MB of scratch surface states apparently.
2/ A shift right by 2 was implicitly happening while copying the
value of r0.5 into the address register holding the extended
descriptor, which with the ExBSO addressing mode disabled
considered bits [31:12] as the surface state index within the
pool.
However on Xe2 ExBSO addressing mode is always enabled for the UGM
shared function, so we have to add an extra SHR instruction to format
the extended descriptor regardless, and there is no point in
disobeying the hardware spec passing a left-shifted offset.
Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/29543>
This is a performance tuning value, recommended value is 512 on DG2.
On DG2 this was in the privileged register RT_CTRL.
Minor CFE_STATE defintion fixes from Jose.
Reviewed-by: José Roberto de Souza <jose.souza@intel.com>
Signed-off-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/29616>
GFX versions older than GFX 20 have 'Thread Preemption disable' while
GFX 20 has 'Thread Preemption' with value flipped in compute walker
instruction.
So here by default enabling thread preemption, only disabling it
when BTD mode is enabled as instructed in Wa_14017794102.
Similar for 3DSTATE_BTD, enabling preemption by default and
only disabling when platform is affected by Wa_14017794102.
Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Signed-off-by: José Roberto de Souza <jose.souza@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/29616>
Up to now preferred SLM size was being set to maximum preferred SLM
size for GFX 12.5 platforms and to workgroup SLM size for Xe2 but
neither of those values are the optimal.
The optimal value is:
<number of workgroups that can run per subslice> * <workgroup SLM size>
Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Signed-off-by: José Roberto de Souza <jose.souza@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/28910>
Xe2 has 2 requirements for preferred SLM size:
- this value needs to be >= then SLM size
- this value must be less than shared SLM/L1$ RAM in the sub-slice of platform
Also Xe2 don't have the special '0' encode that sets preferred SLM
allocation size to the maximum supported.
So here setting a value that is equal or larger than SLM size.
It was always setting SLM_ENCODES_128K for LNL A0 stepping probably
because of Wa_16018610683 but this restriction applies to all Xe2
platforms, also because of the first restriction mentioned here
this workaround is not being properly implemented, will fix that
in the next patch.
We should have a formula to calculate a preferred SLM allocation size
for gfx125 and Xe2 platfoms but until that this is enough to fix at
least the applications and tests below on LNL:
- GFXBench Aztec Ruins VK
- GravityMark VK
- Wildlife Extreme VK
- 5 crucible tests
Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Signed-off-by: José Roberto de Souza <jose.souza@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/28910>
This functions were inlined in a header and duplicated between brw and
elk.
That would be enough reasons to move to a C file but next patches
will add more code to support Xe2 platforms, what would cause more
code to be inlined, duplicating even more code and increasing lib
size.
Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Signed-off-by: José Roberto de Souza <jose.souza@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/28910>
These 2 compute code paths were checking for
anv_cmd_buffer_is_render_queue() before calling
flush_pipeline_select_gpgpu() causing cmd_buffer->state.current_pipeline
to never to be set to GPGPU, trigerring
assert(cmd_buffer->state.current_pipeline == GPGPU) when running in
the compute engine.
So here just dropping the anv_cmd_buffer_is_render_queue() check.
Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Signed-off-by: José Roberto de Souza <jose.souza@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/28053>
A number of allocations during command buffer building are sourced
from the dynamic state heap. They're not actually access using an
offset in the dynamic state heap, it just happens to be a conveninent
place.
Use different helpers for thoses so we dynamically change the dynamic
state heap location in the next commits.
Signed-off-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Reviewed-by: Ivan Briano <ivan.briano@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/22151>
Take the opportunity to fix the flush of the descriptor buffer surface
when needed. Previously we would only flush it if the shader used one
of the push descriptor.
Signed-off-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Cc: mesa-stable
Reviewed-by: Ivan Briano <ivan.briano@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/27504>
This adds a few new fields in the brw_cs_prog_data struct and then
uses them to fill in the relevant COMPUTE_WALKER fields.
Although the Tile Layout field theoretically has different settings for
32/64/128bpe, it appears that the recommended programming is to always
pick either TileY 32bpe or Linear. It's not very practical to look at
the surface formats involved, anyway.
Reviewed-by: Ivan Briano <ivan.briano@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/27167>
We're missing the ISA code in renderdoc. You can reproduce with the
Sascha Willems graphics pipeline demo.
The change is large here because we have to fix a confusion between
anv_shader_bin & anv_pipeline_executable. anv_pipeline_executable is
there as a representation for the user and multiple
anv_pipeline_executable can point to a single anv_shader_bin.
In this change we split the anv_shader_bin related logic that was
added in anv_pipeline_add_executable*() and move it to a new
anv_pipeline_account_shader() function.
When importing RT libraries, we add all the anv_pipeline_executable
from the libraries.
When importing Gfx libraries, we add the anv_pipeline_executable only
if not doing link time optimization.
anv_shader_bin related properties are added whenever we're importing a
shader from a library, compiling or finding in the cache.
Signed-off-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Fixes: 3d49cdb71e ("anv: implement VK_EXT_graphics_pipeline_library")
Reviewed-by: Ivan Briano <ivan.briano@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/26594>
