The per-element fetch has quite some calculations which are constant,
these can be moved outside both the per-element as well as the main
shader loop (llvm can figure out it's constant mostly on its own, however
this can have a significant compile time cost).
Similarly, it looks easier swapping the fetch loops (outer loop per attrib,
inner loop filling up the per vertex elements - this way the aos->soa
conversion also can be done per attrib and not just at the end though again
this doesn't really make much of a difference in the generated code). (This
would also make it possible to vectorize the calculations leading to the
fetches.)
There's also some minimal change simplifying the overflow math slightly.
All in all, the generated code seems to look slightly simpler (depending
on the actual vs), but more importantly I've seen a significant reduction
in compile times for some vs (albeit with old (3.3) llvm version, and the
time reduction is only really for the optimizations run on the IR).
v2: adapt to other draw change.
No changes with piglit.
Reviewed-by: Jose Fonseca <jfonseca@vmware.com>
Previous attempts to zero initialize all inputs were not really optimal
(though no performance impact was measurable). In fact this is not really
necessary, since we know the max number of inputs used.
Instead, just generate fetch for up to max inputs used by the shader,
directly replacing inputs for which there was no vertex element by zero.
This also cleans up key generation, which previously would have stored
some garbage for these elements.
And also drop the assertion which indicates such bogus usage by a
debug_printf (the whole point of initializing the undefined inputs was to
make this case safe to handle).
Reviewed-by: Jose Fonseca <jfonseca@vmware.com>
Compilation to actual machine code can easily take as much time as the
optimization passes on the IR if not more, so print this out too.
Reviewed-by: Brian Paul <brianp@vmware.com>
Reviewed-by: Jose Fonseca <jfonseca@vmware.com>
For the texturing packs, things looked pretty terrible. For every
lerp, we were repacking the values, and while those look sort of cheap
with 128bit, with 256bit we end up with 2 of them instead of just 1 but
worse, plus 2 extracts too (the unpack, however, works fine with a
single instruction, albeit only with llvm 3.8 - the vpmovzxbw).
Ideally we'd use more clever pack for llvmpipe backend conversion too
since we actually use the "wrong" shuffle (which is more work) when doing
the fs twiddle just so we end up with the wrong order for being able to
do native pack when converting from 2x8f -> 1x16b. But this requires some
refactoring, since the untwiddle is separate from conversion.
This is only used for avx2 256bit pack/unpack for now.
Improves openarena scores by 8% or so, though overall it's still pretty
disappointing how much faster 256bit vectors are even with avx2 (or
rather, aren't...). And, of course, eliminating the needless
packs/unpacks in the first place would eliminate most of that advantage
(not quite all) from this patch.
Reviewed-by: Jose Fonseca <jfonseca@vmware.com>
This moves to the shared vk_alloc inlines for vulkan
memory allocations.
Acked-by: Jason Ekstrand <jason@jlekstrand.net>
Signed-off-by: Dave Airlie <airlied@redhat.com>
This moves all the alloc/free in anv to the generic helpers.
Acked-by: Jason Ekstrand <jason@jlekstrand.net>
Signed-off-by: Dave Airlie <airlied@redhat.com>
vulkan allocation allows for overriding the allocator used,
add some macros for anv/radv to share for this.
Acked-by: Jason Ekstrand <jason@jlekstrand.net>
Signed-off-by: Dave Airlie <airlied@redhat.com>
Although the vulkan drivers include mesa macros.h, for
radv I'd like to move away from that.
Reviewed-by: Nicolai Hähnle <nicolai.haehnle@amd.com>
Acked-by: Jason Ekstrand <jason@jlekstrand.net>
Signed-off-by: Dave Airlie <airlied@redhat.com>
This removes the vector code from radv in favour of sharing
code with anv.
Acked-by: Jason Ekstrand <jason@jlekstrand.net>
Signed-off-by: Dave Airlie <airlied@redhat.com>
This just removes the anv vector code and uses the new helper.
Acked-by: Jason Ekstrand <jason@jlekstrand.net>
Signed-off-by: Dave Airlie <airlied@redhat.com>
This is ported from anv, both anv and radv can share this.
Reviewed-by: Nicolai Hähnle <nicolai.haehnle@amd.com>
Acked-by: Jason Ekstrand <jason@jlekstrand.net>
Signed-off-by: Dave Airlie <airlied@redhat.com>
During dual instance encoding submission, if the second encode task and first
encode task have no reference dependency, e.g. p following with idr-frame,
there is a chance the second task will use for its reconstructed picture
buffer the same buffer used by first task for its reference/reconstructed
picture. In this case, buffer corruption may occur depending on encoding
speed. Fix is to force flush these two tasks separately to avoid race condition
Fixes: https://bugs.freedesktop.org/show_bug.cgi?id=98005
Signed-off-by: Boyuan Zhang <boyuan.zhang@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Since commit 63c5d5c6c4, the surfaceless
platform has allowed creation of pbuffer surfaces. But the vtable entry
for eglSwapBuffers has remained NULL.
Discovered by running a little pbuffer test.
Cc: Gurchetan Singh <gurchetansingh@chromium.org>
Reviewed-by: Anuj Phogat <anuj.phogat@gmail.com>
Building the Mesa 12.0.3 distfile failed on a system without python
as generated files were not included in the distfile.
Cc: "12.0" <mesa-stable@lists.freedesktop.org>
Signed-off-by: Jonathan Gray <jsg@jsg.id.au>
Reviewed-by: Emil Velikov <emil.velikov@collabora.com>
Add mesa_glinterop.h to the list of headers that will get included
in the distfile as it is required to build Mesa itself.
Corrects a regression introduced in a89faa2022.
Signed-off-by: Jonathan Gray <jsg@jsg.id.au>
Reviewed-by: Emil Velikov <emil.velikov@collabora.com>
The egl docs directory no longer exists as of
88b5c36fe1.
Remove it from EXTRA_DIST to unbreak 'make dist'
Signed-off-by: Jonathan Gray <jsg@jsg.id.au>
Reviewed-by: Eric Engestrom <eric.engestrom@imgtec.com>
Tested-by: Eric Engestrom <eric.engestrom@imgtec.com>
Reviewed-by: Emil Velikov <emil.velikov@collabora.com>
% pattern rules are a GNU extension. Convert the use of one to a
inference rule to allow this to build on OpenBSD.
This is a related change to the one made in
e3d43dc5ea
Signed-off-by: Jonathan Gray <jsg@jsg.id.au>
Reviewed-by: Emil Velikov <emil.velikov@collabora.com>
Rather than having 4-5 places which do the explicit check/message just
polish the gallium helper and use it everywhere.
Signed-off-by: Emil Velikov <emil.velikov@collabora.com>
Reviewed-by: Axel Davy <axel.davy@ens.fr>
Reviewed-by: Nicolai Hähnle <nicolai.haehnle@amd.com>
Previously it was used to differentiate between the different codepaths
in the loader. Although strictly speaking the (core) of the loader is
only used when a hardware device is available. The latter of which in
itself requires libdrm (one of the codepaths available).
That said, all the configure toggles which relate to enabling/using hw
device should attribute and require libdrm, so there's no need to keep
this code around.
With this gallium_require_drm_loader becomes an empty stub, so nuke that
one as well.
Signed-off-by: Emil Velikov <emil.velikov@collabora.com>
Reviewed-by: Axel Davy <axel.davy@ens.fr>
Reviewed-by: Nicolai Hähnle <nicolai.haehnle@amd.com>
With previous patches nearly all the original code (as seen in the
various loaders) is gone.
Update the copyright/license section to reflect that.
Signed-off-by: Emil Velikov <emil.velikov@collabora.com>
Reviewed-by: Axel Davy <axel.davy@ens.fr>
Reviewed-by: Nicolai Hähnle <nicolai.haehnle@amd.com>
Reminiscent from the pre-loader days, were we had multiple instances of
the loader logic in separate places and one could build a "GALLIUM_ONLY"
version.
Since that is no longer the case and the loaders (glx/egl/gbm) do not
(and should not) require to know any classic/gallium specific we can
drop the argument and the related code.
Signed-off-by: Emil Velikov <emil.velikov@collabora.com>
Reviewed-by: Axel Davy <axel.davy@ens.fr>
Reviewed-by: Nicolai Hähnle <nicolai.haehnle@amd.com>
Effectively everyone with actual hardware and/or requesting the
"device_name" requires a working libdrm. Thus they could/should already
be using the (now only) codepath.
Apart from the code simplification, we can slim down our configure.ac
even further. But that will be done in separate patch(es).
Cc: Gary Wong <gtw@gnu.org>
Signed-off-by: Emil Velikov <emil.l.velikov@gmail.com>
Reviewed-by: Axel Davy <axel.davy@ens.fr>
Reviewed-by: Nicolai Hähnle <nicolai.haehnle@amd.com>
The gbm_device_get_backend_name() provides an (somewhat) internal name
of the implementation/backend used. Is has nothing to do with the udev,
one cannot and should not attempt to derive the name from it.
Signed-off-by: Emil Velikov <emil.velikov@collabora.com>
Reviewed-by: Axel Davy <axel.davy@ens.fr>
Reviewed-by: Nicolai Hähnle <nicolai.haehnle@amd.com>
As of last commit nothing in mesa depends on libudev.
Signed-off-by: Emil Velikov <emil.velikov@collabora.com>
Reviewed-by: Axel Davy <axel.davy@ens.fr>
Reviewed-by: Nicolai Hähnle <nicolai.haehnle@amd.com>
With this all the libudev related code is now gone.
Signed-off-by: Emil Velikov <emil.velikov@collabora.com>
Reviewed-by: Axel Davy <axel.davy@ens.fr>
Reviewed-by: Nicolai Hähnle <nicolai.haehnle@amd.com>
Currently not everyone has libudev and with follow-up patches we'll
completely remove the divergent codepaths.
Use the libdrm drm device API to construct the required ID_PATH_TAG-like
string, to preserve the current functionality for libudev users and
allow others to benefit from it as well.
v2: Drop ranty comments, pick the correct device
v3: \n -> \0 in PCI_ID_PATH_TAG_LENGTH comment (Axel).
v4: Use snprintf (Nicolai)
Signed-off-by: Emil Velikov <emil.velikov@collabora.com>
Reviewed-by: Axel Davy <axel.davy@ens.fr>
Improves readability and allows us to do further cleanups a lot easier.
Signed-off-by: Emil Velikov <emil.velikov@collabora.com>
Reviewed-by: Axel Davy <axel.davy@ens.fr>
Reviewed-by: Nicolai Hähnle <nicolai.haehnle@amd.com>
Currently mesa has three code paths in the loader - libudev, manual
sysfs and drm ioctl one.
Considering the issues we had with libudev - strip those down in favour
of the libdrm drm device API. The latter can be implemented in any way
depending on the platform and can be reused by others.
v2: Use correct message on drmGetDevice failure. (Nicolai)
Cc: Jonathan Gray <jsg@jsg.id.au>
Cc: Jean-Sébastien Pédron <dumbbell@FreeBSD.org>
Signed-off-by: Emil Velikov <emil.velikov@collabora.com>
Reviewed-by: Axel Davy <axel.davy@ens.fr>
With follow on work, we'll untangle and simplify all the different
codepaths in loader. Then again, we forget to set have_pci_id when
libdrm is present (one of the codepaths available).
Signed-off-by: Emil Velikov <emil.velikov@collabora.com>
Reviewed-by: Axel Davy <axel.davy@ens.fr>
Reviewed-by: Nicolai Hähnle <nicolai.haehnle@amd.com>
The indirect handle has to come right after the coordinates, so if there
was a sample/bias/depth compare/offset, everything would end up being
shifted by one argument position.
Signed-off-by: Ilia Mirkin <imirkin@alum.mit.edu>
Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Cc: mesa-stable@lists.freedesktop.org
As specified in va.h, default value should be set on attributes
not present in the input list.
Signed-off-by: Julien Isorce <j.isorce@samsung.com>
Reviewed-by: Mark Thompson <sw@jkqxz.net>
dEQP-GLES31.functional.geometry_shading.instanced.geometry_1_invocations
draws using a geometry shader that specifies
layout(points, invocations = 1) in;
and then uses gl_InvocationID. According to the Haswell PRM, the
"GS Instance ID 0" (and 1) thread payload fields are undefined in
dual object mode:
"If 'dispatch mode' is DUAL_OBJECT this field is not valid."
But there's no point in using them - if there's only one invocation,
the ID will be 0. So just load a constant.
Cc: mesa-stable@lists.freedesktop.org
Signed-off-by: Kenneth Graunke <kenneth@whitecape.org>
Reviewed-by: Anuj Phogat <anuj.phogat@gmail.com>