Set the per-pixel mask based on the value of skip_helpers.
This slightly increase the performance on several traces.
fps_avg helped: gl_gfxbench_trex.trace: 22.30 -> 22.79 (2.20%)
total fps_avg in all runs: 55.18 -> 55.71 (0.97%)
total fps_avg in affected (through threshold) runs: 22.30 -> 22.79 (2.20%)
helped: 1
HURT: 0
Signed-off-by: Juan A. Suarez Romero <jasuarez@igalia.com>
Reviewed-by: Georg Lehmann <dadschoorse@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/38759>
Just changing the intrinsic for load_push_constant is wrong, as nothing
guarantees they will have the same indices in the future.
Signed-off-by: Juan A. Suarez Romero <jasuarez@igalia.com>
Reviewed-by: Georg Lehmann <dadschoorse@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/38759>
It will be used with image loads to enable or disable helper invocations.
This fixes a Vulkan CTS test that perform an imageLoad() inside a
fwidth() operation.
Signed-off-by: Juan A. Suarez Romero <jasuarez@igalia.com>
Reviewed-by: Georg Lehmann <dadschoorse@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/38759>
When a time delta is a float, the minutes and seconds can produce a weird
output between 0.5 and 0.9 with strings like 1m60s. Just forcing a cast
to an integer, the bug is solved.
Signed-off-by: Sergi Blanch Torne <sergi.blanch.torne@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/39134>
Added in 411110f7 as part of !39105 an argument to define the polling period
to monitor a pipeline and check if there are jobs to be enabled. Part of this
MR, 8cf2c50e, also includes changes to improve the experience when using this
tool within a GitLab job. But the pretty_wait method, meant to show a
heartbeat to the user, is disturbing the job traces as '\r' is useless in a
non-terminal console.
Signed-off-by: Sergi Blanch Torne <sergi.blanch.torne@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/39134>
Move the enabled storage_8bit property toggle into the base a7xx GPUProps
class. This enables storageBuffer8BitAccess Vulkan feature on all a7xx
hardware, much like the proprietary driver does. It's also a required
feature with Vulkan 1.4.
Fixes: dEQP-VK.info.device_mandatory_features on pre-a750 a7xx hardware.
Signed-off-by: Zan Dobersek <zdobersek@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/39124>
The header should be 0 for older sdma as well. This fixes
DRI_PRIME support for radeonsi.
Fixes: f5ecc5ffd5 ("ac,radv,radeonsi: add ac_emit_sdma_copy_tiled_sub_window()")
Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/39019>
We already optimize the case where the destination format does not
contain alpha. However, there are a few more cases around formats and
blend constants which we can optimize. In particular, float blending
doesn't support constants so we really want to check if the client hands
us a 0/1 constant.
Reviewed-by: Lars-Ivar Hesselberg Simonsen <lars-ivar.simonsen@arm.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/39171>
This actually enables blending for 4 of the supported float formats.
Technically, RGB16F blending is possible as well, using RGBA16F
internally but we only support FORMAT_R16G16B16_SFLOAT for vertex
buffers so there's really no point. This elimiates a lot of blend
shaders and improves the performance of the 3DMark Wild Life benchmark
by about 5 FPS (7-8%) on my MediaTek Chromebook.
Reviewed-by: Aksel Hjerpbakk <aksel.hjerpbakk@arm.com>
Reviewed-by: Lars-Ivar Hesselberg Simonsen <lars-ivar.simonsen@arm.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/39171>
Blend equations that work on float are treated a bit differently, hence
the new is_float on pan_blend_equation.
Reviewed-by: Lars-Ivar Hesselberg Simonsen <lars-ivar.simonsen@arm.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/39171>
This is duplicated between the two drivers and about to get more
complicated.
Reviewed-by: Aksel Hjerpbakk <aksel.hjerpbakk@arm.com>
Reviewed-by: Lars-Ivar Hesselberg Simonsen <lars-ivar.simonsen@arm.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/39171>
Valhall adds float color target support in hardware, including hardware
blending. This commit just adds the XML and doesn't enable it in the C
code. Annoyingly, even though there's enough bits to do otherwise, the
hardware re-interprets the color (writeback) format field in the render
target descriptor based on the internal format. The easiest way to
handle this in the XML is to just have two different enums and fields in
the Render Target structs which alias. This seems to be the least
duplication while still encoding the necessary information.
Reviewed-by: Aksel Hjerpbakk <aksel.hjerpbakk@arm.com>
Reviewed-by: Lars-Ivar Hesselberg Simonsen <lars-ivar.simonsen@arm.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/39171>
Add 'force_robustness' to 'MESA_DEBUG_KK' to force robustness in all
shaders.
Reviewed-by: Arcady Goldmints-Orlov <arcady@lunarg.com>
Signed-off-by: Aitor Camacho <aitor@lunarg.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/38881>
Replaying a dump file requires the VM state in order to feed the
replay tool with the necessary VMA properties that described the hang,
however, these properties are not necessarily useful once the replay
tool re-runs said traces, however, this patch makes this optional.
Signed-off-by: Carlos Santa <carlos.santa@intel.com>
Reviewed-by: José Roberto de Souza <jose.souza@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/34829>
The tool now can seamlessly support GPU hang dump files from
either the i915 or the Xe drivers.
Signed-off-by: Carlos Santa <carlos.santa@intel.com>
Reviewed-by: José Roberto de Souza <jose.souza@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/34829>
The changes as part of the Contexts state now include:
**** Contexts ****
[HWCTX].replay_offset: 0x0
[HWCTX].replay_length: 0xd000
and the changes as part of the VM state now include:
**** VM state ****
VM.uapi_flags: 0x1
[40000].length: 0x2000
[40000].properties: read_write|bo|mem_region=0x1|pat_index=2|cpu_caching=1
[40000].data: &-)\3!!E9mzzzzzzzzzz
In order to be able to replay a GPU hang from a devcore dump file
new properties have been added describing the offset and the length
of the affected hw context as well as a global VM flag and
several VMA property types: memory region, bo caching, pat index,
memory permission and memory type.
Signed-off-by: Carlos Santa <carlos.santa@intel.com>
Reviewed-by: José Roberto de Souza <jose.souza@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/34829>
Before bringing support for Xe let's create a lib so that
the common code can live there.
Signed-off-by: Carlos Santa <carlos.santa@intel.com>
Reviewed-by: José Roberto de Souza <jose.souza@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/34829>
initial refactoring of the i915 code in preparation
for Xe. No functional changes.
Signed-off-by: Carlos Santa <carlos.santa@intel.com>
Reviewed-by: José Roberto de Souza <jose.souza@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/34829>
SHA1_DIGEST_LENGTH was changed to refect BLAKE3 exposed with SHA1 functions - switch
to BUILD_ID_EXPECTED_HASH_LENGTH.
Fixes: 492a176cbb ("util: increase SHA1_DIGEST_LENGTH to 32 (BLAKE3_KEY_LEN)")
Signed-off-by: Christian Gmeiner <cgmeiner@igalia.com>
Reviewed-by: Erik Faye-Lund <erik.faye-lund@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/39192>
That whole comment about Ivy Bridge is not relevant as ANV don't support IVB.
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/39175>
There is no side affects for this shadowing but better fix it.
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/39175>
This reverts commit 45b6aa1eb7.
This is not thread safe and will lead to buffer leaks, eg:
[threadA] _mesa_reference_buffer_object_ ctx=0x60bc07fa33f0 buf=0x60bc09b90020 CtxRefCount-=993187
[threadB] _mesa_glthread_release_upload_buffer ctx=0x60bc07fa33f0 buf=0x60bc09b90020 ref=7768 CtxRefCount=993212 -> 954
[threadB] _mesa_glthread_upload ctx=0x60bc07fa33f0 buf=0x60bc09eb7d00 CtxRefCount=1000000
[threadA] _mesa_reference_buffer_object_ ctx=0x60bc07fa33f0 buf=0x60bc09b90020 CtxRefCount-=993186
../src/mesa/main/bufferobj.h:201: _mesa_reference_buffer_object_: Assertion `oldObj->Ctx == ctx' failed.
The assert is one added by the previous commit.
Closes: https://gitlab.freedesktop.org/mesa/mesa/-/issues/14483
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/39036>
Comparing the ctx values and then updating the refcounts is not
thread-safe so add an assert to make sure the ctx wasn't updated
by another thread (via detach_ctx_from_buffer).
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/39036>
Unifies nir per instruction float control.
In the future this can be split into contract/reassoc/transform
like SPIR-V.
Reviewed-by: Alyssa Rosenzweig <alyssa.rosenzweig@intel.com>
Reviewed-by: Marek Olšák <marek.olsak@amd.com> (except SPIR-V)
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/39103>
It is not actually available to all the platforms mesa can be compiled
to, so let's keep an opt-in list of supported platforms instead, and
compile it out on all other platforms.
Fixes: 48a0478126 ("zink: add renderdoc handling")
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/39176>
