aboutsummaryrefslogtreecommitdiffstats
path: root/src
Commit message (Collapse)AuthorAgeFilesLines
* radeonsi: enable SPIR-V and GL 4.6 for NIRMarek Olšák2019-11-271-6/+5
| | | | Reviewed-by: Pierre-Eric Pelloux-Prayer <[email protected]>
* radeonsi/nir: support interface output types to fix SPIR-V xfb piglitsMarek Olšák2019-11-271-1/+1
| | | | Acked-by: Pierre-Eric Pelloux-Prayer <[email protected]>
* radeonsi/nir: fix location_frac handling for TCS outputsMarek Olšák2019-11-271-1/+1
| | | | Acked-by: Pierre-Eric Pelloux-Prayer <[email protected]>
* radeonsi/nir: don't rely on data.patch for tess factorsMarek Olšák2019-11-271-1/+4
| | | | | | GLCTS SPIR-V tests have this issue. Acked-by: Pierre-Eric Pelloux-Prayer <[email protected]>
* radeonsi/nir: validate is_patch because SPIR-V doesn't set it for tess factorsMarek Olšák2019-11-271-10/+21
| | | | Acked-by: Pierre-Eric Pelloux-Prayer <[email protected]>
* radeonsi: simplify get_tcs_tes_buffer_address_from_generic_indicesMarek Olšák2019-11-271-27/+19
| | | | Acked-by: Pierre-Eric Pelloux-Prayer <[email protected]>
* radeonsi: simplify the interface of get_dw_address_from_generic_indicesMarek Olšák2019-11-271-29/+19
| | | | Acked-by: Pierre-Eric Pelloux-Prayer <[email protected]>
* radeonsi/nir: implement subgroup system values for SPIR-VMarek Olšák2019-11-274-0/+11
| | | | Acked-by: Pierre-Eric Pelloux-Prayer <[email protected]>
* ac/nir: don't rely on data.patch for tess factorsMarek Olšák2019-11-271-2/+6
| | | | Acked-by: Pierre-Eric Pelloux-Prayer <[email protected]>
* drirc: Set vs_position_always_invariant for Shadow of Mordor on IntelKenneth Graunke2019-11-271-0/+10
| | | | | | | | | | | | | | | | | | | | | | When drawing the main character in Shadow of Mordor, the game appears to draw Talion with one vertex shader, and the Wraith with another. If the compiler optimizes those in different ways which lead to slight imprecisions, then the resulting positions may not line up, leading to Z-fighting occurring as the game decides which of the two are in front. brw_nir_opt_peephole_ffma looks at usages of multiply adds across the entire shader, and may make different decisions between the two, leading to such imprecisions and Z-fighting. This started happening recently after a NIR change to eliminate unnecessary MOVs (7025dbe7), but that change simply exposed the existing problem. Improves performance on Skylake GT4e by 1.22945% +/- 0.398672% (n=3), likely due to the fixed rendering. Closes: https://gitlab.freedesktop.org/mesa/mesa/issues/1985 Fixes: 7025dbe794b ("nir: Skip emitting no-op movs from the builder.") Reviewed-by: Eric Anholt <[email protected]> Reviewed-by: Ian Romanick <[email protected]>
* driconf, glsl: Add a vs_position_always_invariant optionKenneth Graunke2019-11-278-0/+23
| | | | | | | | | | | | | | | | | | | | Many applications use multi-pass rendering and require their vertex shader position to be computed the same way each time. Optimizations may consider, say, fusing a multiply-add based on global usage of an expression in a shader. But a second shader with the same expression may have different code, causing that optimization to make the other choice the second time around. The correct solution is for applications to mark their VS outputs 'invariant', indicating they need multiple shaders to compute that output in the same manner. However, most applications fail to do so. So, we add a new driconf option - vs_position_always_invariant - which forces the gl_Position output in vertex shaders to be marked invariant. Fixes: 7025dbe794b ("nir: Skip emitting no-op movs from the builder.") Reviewed-by: Eric Anholt <[email protected]> Reviewed-by: Ian Romanick <[email protected]>
* turnip: Disable timestamp queries for now.Eric Anholt2019-11-271-2/+2
| | | | | | | They're not implemented, and not critical to bring up immediately. Avoids failures in the CTS when nothing gets written to the query. Reviewed-by: Bas Nieuwenhuizen <[email protected]>
* freedreno/perfcntrs/fdperf: add missing a2xx case in select_counterJonathan Marek2019-11-271-0/+1
| | | | | Signed-off-by: Jonathan Marek <[email protected]> Reviewed-by: Rob Clark <[email protected]>
* freedreno/perfcntrs/fdperf: add missing a20x compatibleJonathan Marek2019-11-271-0/+1
| | | | | Signed-off-by: Jonathan Marek <[email protected]> Reviewed-by: Rob Clark <[email protected]>
* freedreno/perfcntrs/fdperf: fix u64 print on 32-bit buildsJonathan Marek2019-11-271-1/+2
| | | | | Signed-off-by: Jonathan Marek <[email protected]> Reviewed-by: Rob Clark <[email protected]>
* freedreno/perfcntrs: add a2xx MH countersJonathan Marek2019-11-271-4/+186
| | | | | Signed-off-by: Jonathan Marek <[email protected]> Reviewed-by: Rob Clark <[email protected]>
* freedreno/registers: add missing MH perfcounter enum for a2xxJonathan Marek2019-11-271-0/+185
| | | | | Signed-off-by: Jonathan Marek <[email protected]> Reviewed-by: Rob Clark <[email protected]>
* v3d: fix indirect BO allocation for uniformsIago Toral Quiroga2019-11-271-3/+8
| | | | | | | | | | | | | | | | | | | We were always ensuring a minimum size of 4 bytes for uniforms for the case where we don't have any, to account for hardware pre-fetching of the uniform stream, however, pre-fetching could also lead to to out of bounds reads when have read the last uniform in the stream, so we probably want to have the extra 4 bytes to prevent the kernel from observing invalid memory accesses when the uniform stream sits right at the end of a page. This seems to fix MMU exceptions reported with a Linux 5.4 kernel. Credit goes to Phil Elwell for identifying the problem and narrowing it down to memory accesses in the uniform stream. Reported-by: Phil Elwell <[email protected]> Tested-by: Phil Elwell <[email protected]> Reviewed-by: Eric Anholt <[email protected]>
* radv: enable VK_KHR_shader_subgroup_extended_types on GFX10Samuel Pitoiset2019-11-271-1/+1
| | | | | Signed-off-by: Samuel Pitoiset <[email protected]> Reviewed-by: Bas Nieuwenhuizen <[email protected]>
* ac: add 8-bit and 16-bit supports to ac_build_permlane16()Samuel Pitoiset2019-11-271-8/+16
| | | | | Signed-off-by: Samuel Pitoiset <[email protected]> Reviewed-by: Bas Nieuwenhuizen <[email protected]>
* radv/gfx10: fix implementation of exclusive scansSamuel Pitoiset2019-11-271-24/+52
| | | | | | | | | | | This implementation is loosely based on ROCm. https://github.com/RadeonOpenCompute/ROCm-Device-Libs/blob/master/ockl/src/wfredscan.cl This fixes dEQP-VK.subgroups.arithmetic.*.subgroupexclusive* on GFX10. Fixes: 227c29a80de ("amd/common/gfx10: implement scan & reduce operations") Signed-off-by: Samuel Pitoiset <[email protected]> Reviewed-by: Bas Nieuwenhuizen <[email protected]>
* radv: fix enabling sample shading with SampleID/SamplePositionSamuel Pitoiset2019-11-271-7/+24
| | | | | | | | | | When a fragment shader includes an input variable decorated with SampleId or SamplePosition, sample shading should be enabled because minSampleShadingFactor is expected to be 1.0. Cc: 19.2, 19.3 <[email protected]> Signed-off-by: Samuel Pitoiset <[email protected]> Reviewed-by: Bas Nieuwenhuizen <[email protected]>
* turnip: fix integer render targetsJonathan Marek2019-11-261-1/+3
| | | | | | | | | | | | | Add missing required bits. Fixes at least: dEQP-VK.pipeline.render_to_image.dedicated_allocation.1d.small.r16g16_sint_d24_unorm_s8_uint dEQP-VK.pipeline.render_to_image.dedicated_allocation.2d.mipmap.r16g16_sint_d24_unorm_s8_uint dEQP-VK.renderpass.dedicated_allocation.attachment.4.401 dEQP-VK.renderpass2.suballocation.formats.r16_uint.load.draw dEQP-VK.synchronization.op.single_queue.barrier.write_draw_read_copy_image_to_buffer.image_128x128_r16_uint Signed-off-by: Jonathan Marek <[email protected]> Reviewed-by: Eric Anholt <[email protected]>
* anv: Push constants are relative to dynamic state on IVBJason Ekstrand2019-11-261-0/+17
| | | | | | Fixes: aecde2351 "anv: Pre-compute push ranges for graphics pipelines" Closes: #2136 Reviewed-by: Lionel Landwerlin <[email protected]>
* gallium/auxiliary: Fix uses of gnu struct = {} extensionDylan Baker2019-11-265-8/+8
| | | | | | | | | | Most of these will never actually be compiled by windows, but in the interest of being able to make using struct foo = {}; an error and avoiding breaking windows removing a handful of safe uses seems like a good trade off. Reviewed-by: Kristian H. Kristensen <[email protected]> Acked-by: Eric Engestrom <[email protected]>
* st/mesa: add st_variant base class to simplify code for shader variantsMarek Olšák2019-11-268-307/+149
| | | | Reviewed-by: Timothy Arceri <[email protected]>
* st/mesa: don't use ** in the st_nir_link_shaders signatureMarek Olšák2019-11-261-20/+20
| | | | Reviewed-by: Timothy Arceri <[email protected]>
* st/mesa: simplify looping over linked shaders when linking NIRMarek Olšák2019-11-261-48/+28
| | | | Reviewed-by: Timothy Arceri <[email protected]>
* st/mesa: propagate gl_PatchVerticesIn from TCS to TES before linking for NIRMarek Olšák2019-11-261-2/+2
| | | | Reviewed-by: Timothy Arceri <[email protected]>
* st/mesa: don't call ProgramStringNotify in glsl_to_nirMarek Olšák2019-11-262-13/+16
| | | | Reviewed-by: Timothy Arceri <[email protected]>
* st/mesa: don't use redundant stp->state.ir.nirMarek Olšák2019-11-263-25/+12
| | | | Reviewed-by: Timothy Arceri <[email protected]>
* st/mesa: don't serialize all streamout state if there are no SO outputsMarek Olšák2019-11-261-4/+15
| | | | Reviewed-by: Timothy Arceri <[email protected]>
* iris: Disable VF cache partial address workaround on Gen11+Kenneth Graunke2019-11-262-0/+14
| | | | | | | | | | | | | | | The vertex cache uses the full 48-bit address on Gen11+. See the documentation for 3DSTATE_VERTEX_BUFFERS, which describes the workaround and lists it as pre-Icelake. Interestingly, the docs don't mention index buffers as needing a workaround at all. So either we've been overzealous, or the docs never got updated to record that. Which begs the question of whether the issue there was fixed, if there was one... Cuts 40% of the PIPE_CONTROLs from Civilization VI's benchmark; appears that it improves performance by about 1-2% on Icelake 8x8 (not frequency locked).
* freedreno: switch to layout helperRob Clark2019-11-2627-199/+190
| | | | | | | | | | | | The slices table and most of the other layout fields in the freedreno_resource moves into fdl_layout. v2: Changes by anholt to not have duplicate fields, which was introducing a surprising behavior change in resource layout (using the level_linear helper before the setup of the shadowed fields) Reviewed-by: Eric Anholt <[email protected]> Acked-by: Rob Clark <[email protected]>
* freedreno/a6xx: Log the tiling mode in resource layout debug.Eric Anholt2019-11-261-2/+2
| | | | | | | This was important for figuring out what went wrong with the layout refactor. Acked-by: Rob Clark <[email protected]>
* freedreno: Convert the slice struct to the new resource header.Eric Anholt2019-11-2623-71/+43
| | | | | | | | This gets the worst of the sed required for shared resource layout out of the way. The texture layout comment is dropped now that we're referencing the shared header, which has a more complete description. Acked-by: Rob Clark <[email protected]>
* freedreno: Introduce a resource layout header.Eric Anholt2019-11-261-0/+164
| | | | | | | | This will be used for sharing resource layout code between freedreno and tu. Mostly copied from a commit by Rob, with a new location and the slice struct renamed for consistency. Acked-by: Rob Clark <[email protected]>
* freedreno: Introduce a fd_resource_tile_mode() helper.Eric Anholt2019-11-268-36/+24
| | | | | | | | Multiple places were doing the same thing to get the tile mode of a level, so refactor it out. This will make the shared resource helper transition cleaner. Acked-by: Rob Clark <[email protected]>
* freedreno: Introduce a fd_resource_layer_stride() helper.Eric Anholt2019-11-262-11/+15
| | | | | | | This factors out a bit of duplicated code, but will also make the shared resource layout transition process clearer. Acked-by: Rob Clark <[email protected]>
* freedreno: use rsc->slice accessor everywhereRob Clark2019-11-2614-45/+58
| | | | | | | This will make it easier to extract the slice table out into a layout helper. Acked-by: Rob Clark <[email protected]>
* nir: Make algebraic backtrack and reprocess after a replacement.Eric Anholt2019-11-262-22/+97
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | The algebraic pass was exhibiting O(n^2) behavior in dEQP-GLES2.functional.uniform_api.random.3 and dEQP-GLES31.functional.ubo.random.all_per_block_buffers.13 (along with other code-generated tests, and likely real-world loop-unroll cases). In the process of using fmul(b2f(x), b2f(x)) -> b2f(iand(x, y)) to transform: result = b2f(a == b); result *= b2f(c == d); ... result *= b2f(z == w); -> temp = (a == b) temp = temp && (c == d) ... temp = temp && (z == w) result = b2f(temp); nir_opt_algebraic, proceeding bottom-to-top, would match and convert the top-most fmul(b2f(), b2f()) case each time, leaving the new b2f to be matched by the next fmul down on the next time algebraic got run by the optimization loop. Back in 2016 in 7be8d0773229 ("nir: Do opt_algebraic in reverse order."), Matt changed algebraic to go bottom-to-top so that we would match the biggest patterns first. This helped his cases, but I believe introduced this failure mode. Instead of reverting that, now that we've got the automaton, we can update the automaton's state recursively and just re-process any instructions whose state has changed (indicating that they might match new things). There's a small chance that the state will hash to the same value and miss out on this round of algebraic, but this seems to be good enough to fix dEQP. Effects with NIR_VALIDATE=0 (improvement is better with validation enabled): Intel shader-db runtime -0.954712% +/- 0.333844% (n=44/46, obvious throttling outliers removed) dEQP-GLES2.functional.uniform_api.random.3 runtime -65.3512% +/- 4.22369% (n=21, was 1.4s) dEQP-GLES31.functional.ubo.random.all_per_block_buffers.13 runtime -68.8066% +/- 6.49523% (was 4.8s) v2: Use two worklists, suggested by @cwabbott, to cut out a bunch of tricky code. Runtime of uniform_api.random.3 down -0.790299% +/- 0.244213% compred to v1. v3: Re-add the nir_instr_remove() that I accidentally dropped in v2, fixing infinite loops. Reviewed-by: Connor Abbott <[email protected]>
* nir: Refactor algebraic's block walkEric Anholt2019-11-261-31/+31
| | | | | | | | | My motivation was to clarify the changes in the following commit, but incidentally, it reduces runtime of dEQP-GLES2.functional.uniform_api.random.3 (an algebraic-heavy testcase) by -5.39524% +/- 2.21179% (n=15) Reviewed-by: Connor Abbott <[email protected]>
* nir: Maintain the algebraic automaton's state as we work.Connor Abbott2019-11-262-38/+78
| | | | | | | | In order to have nir_opt_algebraic be able to do further algebraic work on the output of a replacement, we need to maintain the automaton's state. Reviewed-by: Eric Anholt <[email protected]>
* etnaviv: support 3d/array/integer formats in texture descriptorsJonathan Marek2019-11-261-3/+19
| | | | | Signed-off-by: Jonathan Marek <[email protected]> Reviewed-by: Christian Gmeiner <[email protected]>
* etnaviv: blt: fix partial ZS clears with TSJonathan Marek2019-11-261-4/+7
| | | | | | | | If not all bits are cleared, then BLT needs to be given the current clear value and not the new one. Signed-off-by: Jonathan Marek <[email protected]> Reviewed-by: Christian Gmeiner <[email protected]>
* aco: don't value-number instructions from within a loop with ones after the ↵Daniel Schürmann2019-11-261-1/+6
| | | | | | | | | | loop. Fixes: Wolfenstein:Youngblood (w/o shader_ballot) dEQP-VK.descriptor_indexing.combined_image_sampler_in_loop_with_lod Reviewed-by: Rhys Perry <[email protected]>
* aco: set dlc/glc correctly for image loadsRhys Perry2019-11-261-0/+3
| | | | | Signed-off-by: Rhys Perry <[email protected]> Reviewed-By: Timur Kristóf <[email protected]>
* aco: allow constant offsets for global/scratch instructions on GFX10Rhys Perry2019-11-262-2/+5
| | | | | | | | | | I don't think the bug applies for global/scratch instructions and load_barycentric_at_sample selection expects this feature to work. Fixes various dEQP-VK.pipeline.multisample_interpolation.* tests on GFX10. Signed-off-by: Rhys Perry <[email protected]> Reviewed-By: Timur Kristóf <[email protected]>
* radv: Enable VK_KHR_buffer_device_address.Bas Nieuwenhuizen2019-11-262-3/+25
| | | | | | Still no capture/replay or multi device support. Reviewed-by: Samuel Pitoiset <[email protected]>
* radv: fix reporting subgroup size with VK_KHR_pipeline_executable_propertiesSamuel Pitoiset2019-11-261-3/+2
| | | | | Signed-off-by: Samuel Pitoiset <[email protected]> Reviewed-by: Bas Nieuwenhuizen <[email protected]>
generated by cgit v1.2.3 (git 2.39.1) at 2024-12-27 11:51:46 +0000