Not [originally] intended for upstream. Should cause a GPU hang if
some thread is executed with a non-contiguous dispatch mask breaking
assumptions of brw_stage_has_packed_dispatch(). Doesn't cause any
CTS, DEQP or Piglit regressions, while replacing
brw_stage_has_packed_dispatch() with a dummy implementation that
unconditionally returns true on top of this patch causes multiple GPU
hangs.
v2: Refactor into a separate function instead of emitting the test
code directly from emit_nir_code(), drop VEC4 test and clean up
slightly for upstream. (Jason)
Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>
This avoids emitting a few extra instructions required to take the
dispatch mask into account when it's known to be tightly packed.
Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>
The eliminate_find_live_channel optimization eliminates
FIND_LIVE_CHANNEL instructions in cases where control flow is known to
be uniform, and replaces them with 'MOV 0', which in turn unblocks
subsequent elimination of the BROADCAST instruction frequently used on
the result of FIND_LIVE_CHANNEL. This is however not correct in
per-sample fragment shader dispatch because the PSD can dispatch a
fully unlit sample under certain conditions. Disable the optimization
in that case.
Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>
v2: Add devinfo argument to brw_stage_has_packed_dispatch() to
implement hardware generation check.
On at least Sky Lake, ce0 does not contain the full story as far as enabled
channels goes. It is possible to have completely disabled channels where
the corresponding bits in ce0 are 1. In order to get the correct execution
mask, you have to mask off those channels which were disabled from the
beginning by taking the AND of ce0 with either sr0.2 or sr0.3 depending on
the shader stage. Failure to do so can result in FIND_LIVE_CHANNEL
returning a completely dead channel.
Signed-off-by: Jason Ekstrand <jason@jlekstrand.net>
Cc: Francisco Jerez <currojerez@riseup.net>
[ Francisco Jerez: Fix a couple of typos, add mask register type
assertion, clarify reason why ce0 can have bits set for disabled
channels, clarify that this may only be a problem when thread
dispatch doesn't pack channels tightly in the SIMD thread. Apply
same treatment to Align16 path. ]
Reviewed-by: Francisco Jerez <currojerez@riseup.net>
The state register sr0 is really a collection of dwords not a SIMD8
anything. It's much more convenient for brw_sr0_reg to return the
particular dword you're looking for rather than a giant blob you have to
massage into what you want.
Signed-off-by: Jason Ekstrand <jason@jlekstrand.net>
[ Francisco Jerez: Trivial simplification of brw_ud1_reg(). ]
Reviewed-by: Francisco Jerez <currojerez@riseup.net>
Reproduces this commit :
commit 0fb85ac08d
Author: Kenneth Graunke <kenneth@whitecape.org>
Date: Mon Jun 6 21:37:34 2016 -0700
i965: Use the correct number of threads for compute shaders.
Signed-off-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
This reproduces this commit :
commit 2213ffdb4b
Author: Kenneth Graunke <kenneth@whitecape.org>
Date: Mon Jun 6 21:37:34 2016 -0700
i965: Allocate scratch space for the maximum number of compute threads.
Signed-off-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
This should be analogous to 32-bit integers.
Reviewed-by: Edward O'Callaghan <funfunctor@folklore1984.net>
Signed-off-by: Nicolai Hähnle <nicolai.haehnle@amd.com>
v2:
- no PIPE_CAP_INT64 yet
- emit DIV/MOD without the divide-by-zero workaround
Reviewed-by: Marek Olšák <marek.olsak@amd.com> (v1)
Reviewed-by: Edward O'Callaghan <funfunctor@folklore1984.net>
Signed-off-by: Dave Airlie <airlied@redhat.com>
Signed-off-by: Nicolai Hähnle <nicolai.haehnle@amd.com>
This enables 64-bit integer support in gallivm and
llvmpipe.
v2: add conversion opcodes.
v3:
- PIPE_CAP_INT64 is not there yet
- restrict DIV/MOD defaults to the CPU, as for 32 bits
- TGSI_OPCODE_I2U64 becomes TGSI_OPCODE_U2I64
Reviewed-by: Roland Scheidegger <sroland@vmware.com>
Reviewed-by: Edward O'Callaghan <funfunctor@folklore1984.net>
Signed-off-by: Dave Airlie <airlied@redhat.com>
Signed-off-by: Nicolai Hähnle <nicolai.haehnle@amd.com>
This adds all the opcodes to tgsi_exec for softpipe to use.
v2: add conversion opcodes.
v3:
- no PIPE_CAP_INT64 yet
- change TGSI_OPCODE_I2U64 to TGSI_OPCODE_U2I64
Reviewed-by: Roland Scheidegger <sroland@vmware.com>
Reviewed-by: Edward O'Callaghan <funfunctor@folklore1984.net>
Signed-off-by: Dave Airlie <airlied@redhat.com>
Signed-off-by: Nicolai Hähnle <nicolai.haehnle@amd.com>
This adds support to TGSI for 64-bit integer immediates.
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Reviewed-by: Nicolai Hähnle <nicolai.haehnle@amd.com>
Reviewed-by: Roland Scheidegger <sroland@vmware.com>
Reviewed-by: Edward O'Callaghan <funfunctor@folklore1984.net>
Signed-off-by: Dave Airlie <airlied@redhat.com>
This just adds the basic support for 64-bit opcodes,
and the new types.
v2: add conversion opcodes.
add documentation.
v3:
- make docs more consistent
- change TGSI_OPCODE_I2U64 to TGSI_OPCODE_U2I64
Reviewed-by: Marek Olšák <marek.olsak@amd.com> (v2)
Reviewed-by: Roland Scheidegger <sroland@vmware.com>
Reviewed-by: Edward O'Callaghan <funfunctor@folklore1984.net>
Signed-off-by: Dave Airlie <airlied@redhat.com>
Signed-off-by: Nicolai Hähnle <nicolai.haehnle@amd.com>
"intelScreen" is wordy and also doesn't fit our style guidelines.
"screen" is shorter, which is nice, because we use it fairly often.
Signed-off-by: Kenneth Graunke <kenneth@whitecape.org>
Reviewed-by: Anuj Phogat <anuj.phogat@gmail.com>
I want to use "screen" as the variable name for a struct intel_screen
pointer. This means that we can't use it for __DRIscreen pointers.
Sometimes we called it "screen", sometimes "sPriv", sometimes
"driScrnPriv", and sometimes "psp" (Pointer to Screen Private?).
The last one is particularly confusing because we use "psp" to refer to
the Gen4 PIPELINED_STATE_POINTERS packet as well.
Let's be consistent. "dri_screen" is clear, and it's not used often
enough that I'm worried about the verbosity.
Signed-off-by: Kenneth Graunke <kenneth@whitecape.org>
Reviewed-by: Anuj Phogat <anuj.phogat@gmail.com>
This extension is a combination of AMD_vertex_shader_viewport_index and
AMD_vertex_shader_layer, making it rather trivial to implement.
For gallium I *think* this needs a new cap because of the addition of
support in tessellation evaluation shaders, and since I don't have any
hardware to test it on, I've left that for someone else to wire up.
Signed-off-by: Dylan Baker <dylanx.c.baker@intel.com>
Reviewed-by: Ilia Mirkin <imirkin@alum.mit.edu>
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
(derived from commit 3b6bda665a)
v2: fix the tabs(Leo)
Reviewed-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Nishanth Peethambaran <nishanth.peethambaran@amd.com>
Signed-off-by: Indrajit Das <indrajit-kumar.das@amd.com>
Signed-off-by: Leo Liu <leo.liu@amd.com>
so that the texture is rendered to back buffer before calling
flush_frontbuffer and can be copied to a different buffer in
the function
v2: change comment style
Signed-off-by: Nayan Deshmukh <nayan26deshmukh@gmail.com>
Reviewed-by: Michel Dänzer <michel.daenzer@amd.com>
Acked-by: Christian König <christian.koenig@amd.com>
so that the texture is rendered to back buffer before calling
flush_frontbuffer and can be copied to a different buffer in
the function
Signed-off-by: Nayan Deshmukh <nayan26deshmukh@gmail.com>
Reviewed-by: Michel Dänzer <michel.daenzer@amd.com>
Acked-by: Christian König <christian.koenig@amd.com>
In case of prime when rendering is done on GPU other then the
server GPU, use a seprate linear buffer for each back buffer
which will be displayed using present extension.
v2: Use a seprate linear buffer for each back buffer (Michel)
v3: Change variable names and fix coding style (Leo and Emil)
v4: Use PIPE_BIND_SAMPLER_VIEW for back buffer in case when
a seprate linear buffer is used (Michel)
v4.1: remove empty line
v4.2: destroy the context and handle the case when
create_context fails (Emil)
Signed-off-by: Nayan Deshmukh <nayan26deshmukh@gmail.com>
Reviewed-by: Leo Liu <leo.liu@amd.com>
Acked-by: Michel Dänzer <michel.daenzer@amd.com>
Acked-by: Christian König <christian.koenig@amd.com>
Same thing as nvc0_stage_set_sampler_views_range().
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Ilia Mirkin <imirkin@alum.mit.edu>
This function was quite similar to nvc0_stage_set_sampler_views()
and I don't see any reasons to not remove it.
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Ilia Mirkin <imirkin@alum.mit.edu>
This helps shaders in UE4 demos, especially with Elemental
(+1% perf). This optimization reduces spilling usage in one
shader which explains the little gain.
GF100/GK104:
total instructions in shared programs :2838551 -> 2838045 (-0.02%)
total gprs used in shared programs :396706 -> 396684 (-0.01%)
total local used in shared programs :34432 -> 34416 (-0.05%)
local gpr inst bytes
helped 1 19 112 112
hurt 0 0 0 0
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Ilia Mirkin <imirkin@alum.mit.edu>
This should emit src0 instead of src1.
Found by inspection.
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Ilia Mirkin <imirkin@alum.mit.edu>
Cc: mesa-stable@lists.freedesktop.org
A couple of forward-declarations were causing warnings in clang:
'value' defined as a class here but previously declared as a struct
[-Wmismatched-tags]
Signed-off-by: Martina Kollarova <martina.kollarova@intel.com>
Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
Given robust access, we should just be returning zeroes if the user gives
us a base pointer that's too big, which is what was happens on a release
build. This was caught by a webgl conformance test for out-of-bounds
draws on servo.
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Fixes the following piglit test (for softpipe):
/spec/glsl-1.10/execution/fs-loop-return
Signed-off-by: Lars Hamre <chemecse@gmail.com>
Reviewed-by: Brian Paul <brianp@vmware.com>
This patch relaxes the restriction of compressed formats for texture
upload buffer. For now, 3D texture with compressed format
is still not supported in the texture upload buffer path.
As Brian noted, ETQW does many texture updates with glCompressedTexSubImage.
This patch greatly improves the performance of the ETQW trace.
Tested with ETQW, MTT piglit, glretrace, conform, viewperf
v2: Per Brian's suggestion, removed the subregion boundary check.
Reviewed-by: Brian Paul <brianp@vmware.com>
This reduces the number of times we flush in some situations (the
arbocclude demo is one trivial example).
Tested with Piglit, ETQW, Sauerbraten, arbocclude.
Reviewed-by: Charmaine Lee <charmainel@vmware.com>
Since commit 99d8fe20ab we don't have to flush the command buffer when
we end a query.
Tested with Piglit, Sauerbraten, arbocclude, ETQW (noticably faster now).
Reviewed-by: José Fonseca <jfonseca@vmware.com>
With this patch, when running with vgpu10, instead of mapping directly to the
guest backed memory for texture update, we'll use the texture upload buffer
and use the transfer from buffer command to update the host side texture memory.
This optimization yields about 20% performance improvement with
Lightsmark2008 and about 40% with Tropics.
Tested with Lightsmark2008, Tropics, Heaven, MTT piglit, glretrace, conform.
Reviewed-by: Brian Paul <brianp@vmware.com>
Split the functions into separate functions for dma and direct map to make
the code more readable.
Tested with MTT piglit, glretrace, viewperf, conform, various OpenGL apps
Reviewed-by: Brian Paul <brianp@vmware.com>
With this patch, single sample surface will be created as non-multisamples
surface.
Tested with piglit, glretrace.
Reviewed-by: Brian Paul <brianp@vmware.com>
This patch fixes a memory leak with sampler state when piglit
is run with HW version 11. Sampler state clean up was incorrectly skipped
in svga_cleanup_sampler_state() for vgpu9.
Tested with piglit.
Reviewed-by: Neha Bhende <bhenden@vmware.com>
Reviewed-by: Brian Paul <brianp@vmware.com>
Left over test code spotted by Sinclair.
Tested with piglit, Google Earth, Lightsmark, Heaven4, glretraces, etc.
Reviewed-by: Sinclair Yeh <syeh@vmware.com>
Use SVGA3D_QUERYTYPE_MAX instead of SVGA_QUERY_MAX for
svga query type check.
Tested with various OpenGL apps with GALLIUM_HUD set.
Reviewed-by: Brian Paul <brianp@vmware.com>
Replace the num-resources-mapped hud with
num-textures-mapped and num-buffers-mapped, so we can
differentiate the map counts for these two different resources.
Reviewed-by: Brian Paul <brianp@vmware.com>
Some OpenGL apps, like Cinebench R15, have many glDrawElements(GL_QUADS)
calls. Since we don't directly support quads we have to convert these
calls into GL_TRIANGLES which involves generating a new index buffer.
This patch saves the new/translated index buffer in the hope that it
can be reused for a later draw call.
Cinebench R15 increases by about 20% with this change.
The NobelClinician Viewer app also hits this code.
Tested with full piglit run.
Reviewed-by: Charmaine Lee <charmainel@vmware.com>
If a consecutive sequence of drawing commands references the same
vertex/index buffers, there should be no need to rebind the surfaces
for the second and subsequent drawing commands.
Apps that use multiple display lists benefit from this since the vertex
data for several display lists is often stored in one buffer.
In the case of the legacy E&S Glaze demo, this reduces the size of our
command buffers from 91KB to 44KB. One WSI Fusion trace shows a 33%
reduction in command buffer sizes.
Tested with full piglit run.
Reviewed-by: Charmaine Lee <charmainel@vmware.com>
Previously, every time we put shader constants into the default constant
buffer we called u_upload_alloc(), which mapped the buffer, and
u_upload_unmap(). We had to unmap the buffer before calling
svga_buffer_handle() to get the winsys handle for the buffer. But we
really only need to do that the first time we reference the const buffer.
Now we try to keep the upload manager's buffer mapped until we fill it or
flush the command buffer.
v2: add additional comment on the buffer unmapping code in
svga_context_flush(), per Charmaine.
Reviewed-by: Charmaine Lee <charmainel@vmware.com>
