summaryrefslogtreecommitdiffstats
path: root/src/video_core/engines/maxwell_3d.cpp (follow)
Commit message (Collapse)AuthorAgeFilesLines
* Merge pull request #9802 from Kelebek1/wait_data_cachebunnei2023-02-161-0/+4
|\ | | | | Re-add the invalidate_texture_data_cache register
| * Reimplement the invalidate_texture_data_cache registerKelebek12023-02-141-0/+4
| |
* | remove static from pointer sized or smaller types for aesthetics, change constexpr static to static constexpr for consistencyarades792023-02-141-1/+1
| | | | | | | | Signed-off-by: arades79 <scravers@protonmail.com>
* | add static lifetime to constexpr values to force compile time evaluation where possiblearades792023-02-141-1/+1
|/ | | | Signed-off-by: arades79 <scravers@protonmail.com>
* Merge pull request #9556 from vonchenplus/draw_textureliamwhite2023-01-191-0/+1
|\ | | | | video_core: Implement maxwell3d draw texture method
| * video_core: Implement maxwell3d draw texture methodFeng Chen2023-01-051-0/+1
| |
* | Demote maxwell3d Firmware4 call log to debugKelebek12023-01-181-1/+1
| |
* | video_core: Cache GPU internal writes.Fernando Sahmkow2023-01-051-6/+1
|/
* Video_core: Address feedbackFernando Sahmkow2023-01-041-7/+6
|
* Rasterizer: Setup skeleton for Host Conditional renderingFernando Sahmkow2023-01-011-4/+10
|
* Vulkan: Implement Dynamic State 3Fernando Sahmkow2023-01-011-1/+0
|
* Vulkan Implement Dynamic State 2 LogicOp and PatchVerticesFernando Sahmkow2023-01-011-0/+1
|
* DMAPusher: Improve collection of non executing methodsFernando Sahmkow2023-01-011-0/+94
|
* Revert Buffer cache changes and setup additional macros.Fernando Sahmkow2023-01-011-17/+9
|
* MacroHLE: Reduce massive calculations on sizing estimation.Fernando Sahmkow2023-01-011-0/+15
|
* MacroHLE: Add HLE replacement for base vertex and base instance.Fernando Sahmkow2023-01-011-2/+13
|
* MacroHLE: Add Index Buffer size estimation.Fernando Sahmkow2023-01-011-0/+7
|
* MacroHLE: Refactor MacroHLE system.Fernando Sahmkow2023-01-011-0/+37
|
* MacroHLE: Implement DrawIndexedIndirect & DrawArraysIndirect.Fernando Sahmkow2023-01-011-0/+4
|
* Merge pull request #9401 from vonchenplus/draw_managerFernando S2022-12-081-161/+10
|\ | | | | video_core: Implement maxwell3d draw manager and split draw logic
| * video_core: Implement maxwell3d draw manager and split draw logicFeng Chen2022-12-081-161/+10
| |
* | Merge pull request #9360 from Kelebek1/R-E-S-P-E-C-Tliamwhite2022-12-061-29/+39
|\ \ | |/ |/| Respect render mode override
| * Respect render mode overrideKelebek12022-11-301-29/+39
| |
* | Merge pull request #9353 from vonchenplus/draw_indexedliamwhite2022-12-031-27/+21
|\ \ | |/ |/| video_core: Fine tuning the index drawing judgment logic
| * video_core: Fine tuning the index drawing judgment logicFeng Chen2022-12-011-27/+21
| |
* | engines: Remove unnecessary castsLioncash2022-11-291-6/+6
|/ | | | In a few cases we have some casts that can be trivially removed.
* Merge pull request #9288 from vonchenplus/deferred_drawliamwhite2022-11-261-59/+57
|\ | | | | video_core: Fine tune maxwell drawing trigger mechanism
| * video_core: Optimize maxwell drawing trigger mechanismFengChen2022-11-221-59/+57
| |
* | GPU: Fix buffer cache issue, engine upload not inlining memory in multiline and pessismistic invalidation.Fernando Sahmkow2022-11-241-7/+1
|/
* maxwell3d: full HLE for multi-layer clearsLiam2022-11-171-3/+3
|
* Fix regs regression with OpenGL two-sided stencil, and re-add data invalidation regKelebek12022-11-111-0/+5
|
* video_core: Fix drawing trigger mechanism regressionFengChen2022-10-311-32/+25
|
* video_core: Fix drawing trigger mechanism regressionFengChen2022-10-271-61/+68
|
* Merge pull request #9112 from vonchenplus/deferred_drawliamwhite2022-10-251-151/+134
|\ | | | | video_core: Reimplementing the maxwell drawing trigger mechanism
| * video_core: Implement maxwell inline_index methodFengChen2022-10-221-66/+94
| |
| * video_coare: Reimplementing the maxwell drawing trigger mechanismFengChen2022-10-211-151/+106
| |
* | Merge pull request #9095 from FernandoS27/meat-good-vegetable-badFernando S2022-10-221-10/+7
|\ \ | |/ |/| Maxwell3D/Puller: Fix regressions and syncing issues.
| * Maxwell3D/Puller: Fix regressions and syncing issues.Fernando Sahmkow2022-10-191-10/+7
| |
* | Fix stencil func registers, make clip control equivalent to how it was before, but surely wrong.Kelebek12022-10-101-4/+4
|/
* Update 3D regsKelebek12022-10-071-164/+172
|
* NVDRV: Further improvements.Fernando Sahmkow2022-10-061-13/+5
|
* DMA & InlineToMemory Engines Rework.bunnei2022-10-061-2/+3
|
* Maxwell3D: Add small_index_2Fernando Sahmkow2022-10-061-0/+2
|
* VideoCore: Refactor fencing system.Fernando Sahmkow2022-10-061-3/+21
|
* Maxwell3D: Fix 3D semaphore counter type 0 handlingBilly Laws2022-06-021-2/+2
| | | | Counter type 0 actually releases the semaphore payload rather than a constant zero as was previously thought. This is required by Skyrim.
* video_core/macro: clear code on upload address assignmentLiam2022-05-101-0/+2
|
* general: Convert source file copyright comments over to SPDXMorph2022-04-231-3/+2
| | | | | This formats all copyright comments according to SPDX formatting guidelines. Additionally, this resolves the remaining GPLv2 only licensed files by relicensing them to GPLv2.0-or-later.
* maxwell3d: add small_index_2 registerLiam2022-04-141-0/+5
|
* Merge pull request #8023 from ameerj/kirby-pop-inFernando S2022-03-161-56/+11
|\ | | | | maxwell_3d: Implement a safer CB data upload
| * maxwell_3d: Implement a safer CB data uploadameerj2022-03-151-56/+11
| | | | | | | | This makes constant buffer uploads safer and more accurate by updating the GPU memory as soon as the CB Data method is invoked. The previous implementation was deferring the updates until a different maxwell 3d method was detected, then writing all CB data at once.
* | Maxwell3D: restore original topology when topology overrides are disabledbyte[]2022-03-141-0/+2
| |
* | Maxwell3D: Use override constants from nouveauLiam2022-03-141-1/+22
| | | | | | | | This fixes some incorrect rendering in Sunshine
* | Maxwell3D: Restrict topology override effect to after the register is setLiam2022-03-121-1/+4
| |
* | Maxwell3D: mark index buffers as dirty after updating countsLiam2022-03-111-0/+2
| |
* | Maxwell3D: read small-index draw and primitive topology override registersLiam2022-03-111-0/+14
|/ | | | This allows Galaxy and Sunshine to render for the first time.
* Rasterizer: Implement Inline2Memory Acceleration.Fernando Sahmkow2022-01-291-0/+1
|
* video_core/macro: Remove unused parameter from Execute()Lioncash2022-01-251-1/+1
| | | | Simplifies the function interface.
* shader: Unify shader stage typesReinUsesLisp2021-07-231-1/+0
|
* shader: Remove old shader managementReinUsesLisp2021-07-231-38/+0
|
* buffer_cache: Simplify uniform disabling logicameerj2021-06-011-2/+6
|
* video_core: Reimplement the buffer cacheReinUsesLisp2021-02-131-9/+8
| | | | | | | | | | | | | | | | | | | | | | | | | | | | Reimplement the buffer cache using cached bindings and page level granularity for modification tracking. This also drops the usage of shared pointers and virtual functions from the cache. - Bindings are cached, allowing to skip work when the game changes few bits between draws. - OpenGL Assembly shaders no longer copy when a region has been modified from the GPU to emulate constant buffers, instead GL_EXT_memory_object is used to alias sub-buffers within the same allocation. - OpenGL Assembly shaders stream constant buffer data using glProgramBufferParametersIuivNV, from NV_parameter_buffer_object. In theory this should save one hash table resolve inside the driver compared to glBufferSubData. - A new OpenGL stream buffer is implemented based on fences for drivers that are not Nvidia's proprietary, due to their low performance on partial glBufferSubData calls synchronized with 3D rendering (that some games use a lot). - Most optimizations are shared between APIs now, allowing Vulkan to cache more bindings than before, skipping unnecesarry work. This commit adds the necessary infrastructure to use Vulkan object from OpenGL. Overall, it improves performance and fixes some bugs present on the old cache. There are still some edge cases hit by some games that harm performance on some vendors, this are planned to be fixed in later commits.
* gpu: Report renderer errors with exceptionsReinUsesLisp2021-02-131-2/+2
| | | | | | Instead of using a two step initialization to report errors, initialize the GPU renderer and rasterizer on the constructor and report errors through std::runtime_error.
* maxwell_3d: Silence array bounds warningsReinUsesLisp2021-01-241-34/+34
|
* video_core: Rewrite the texture cacheReinUsesLisp2020-12-301-33/+12
| | | | | | | | | | | | | | The current texture cache has several points that hurt maintainability and performance. It's easy to break unrelated parts of the cache when doing minor changes. The cache can easily forget valuable information about the cached textures by CPU writes or simply by its normal usage.The current texture cache has several points that hurt maintainability and performance. It's easy to break unrelated parts of the cache when doing minor changes. The cache can easily forget valuable information about the cached textures by CPU writes or simply by its normal usage. This commit aims to address those issues.
* video_core: Remove unnecessary enum class casting in logging messagesLioncash2020-12-071-6/+5
| | | | | | | fmt now automatically prints the numeric value of an enum class member by default, so we don't need to use casts any more. Reduces the line noise a bit.
* maxwell_3d: Use insert instead of loop push_backReinUsesLisp2020-11-111-3/+1
| | | | | | This reduces the overhead of bounds checking on each element. It won't reduce the cost of allocation because usually this vector's capacity is usually large enough to hold whatever we push to it.
* maxwell_3d: Move code to separate functionsReinUsesLisp2020-11-111-151/+116
| | | | | Deduplicate some code and put it in separate functions so it's easier to understand and profile.
* General: Make use of std::nullopt where applicableLioncash2020-09-221-1/+1
| | | | | | | | Allows some implementations to avoid completely zeroing out the internal buffer of the optional, and instead only set the validity byte within the structure. This also makes it consistent how we return empty optionals.
* video_core: Initialize renderer with a GPUReinUsesLisp2020-08-221-15/+20
| | | | | Add an extra step in GPU initialization to be able to initialize render backends with a valid GPU instance.
* Macro HLE supportDavid Marcec2020-06-241-1/+1
|
* Merge pull request #4049 from ReinUsesLisp/separate-samplersbunnei2020-06-131-1/+4
|\ | | | | shader/texture: Join separate image and sampler pairs offline
| * shader/texture: Join separate image and sampler pairs offlineReinUsesLisp2020-06-051-1/+4
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Games using D3D idioms can join images and samplers when a shader executes, instead of baking them into a combined sampler image. This is also possible on Vulkan. One approach to this solution would be to use separate samplers on Vulkan and leave this unimplemented on OpenGL, but we can't do this because there's no consistent way of determining which constant buffer holds a sampler and which one an image. We could in theory find the first bit and if it's in the TIC area, it's an image; but this falls apart when an image or sampler handle use an index of zero. The used approach is to track for a LOP.OR operation (this is done at an IR level, not at an ISA level), track again the constant buffers used as source and store this pair. Then, outside of shader execution, join the sample and image pair with a bitwise or operation. This approach won't work on games that truly use separate samplers in a meaningful way. For example, pooling textures in a 2D array and determining at runtime what sampler to use. This invalidates OpenGL's disk shader cache :) - Used mostly by D3D ports to Switch
* | Merge pull request #4009 from ogniK5377/macro-jit-prodbunnei2020-06-041-10/+7
|\ \ | |/ |/| video_core: Implement Macro JIT
| * Default init labels and use initializer list for macro engineDavid Marcec2020-06-041-1/+1
| |
| * Mark parameters as constDavid Marcec2020-06-031-1/+1
| |
| * Pass by reference instead of copying parametersDavid Marcec2020-06-021-4/+6
| |
| * Implement macro JITDavid Marcec2020-05-301-12/+7
| |
* | Merge pull request #3998 from ReinUsesLisp/init-3dbunnei2020-06-011-0/+4
|\ \ | |/ |/| maxwell_3d: Initialize more registers to their expected value
| * maxwell_3d: Initialize line widthsReinUsesLisp2020-05-271-0/+2
| | | | | | | | Initialize line widths to avoid setting a line width of zero.
| * maxwell_3d: Initialize polygon modesReinUsesLisp2020-05-271-0/+2
| | | | | | | | | | NVN expects this to be initialized as Fill, otherwise games that never bind a rasterizer state will log an invalid polygon mode.
* | maxwell_3d: Reduce severity of logs that can be spammedReinUsesLisp2020-05-281-6/+7
|/ | | | | These logs were killing performance on some games when they were spammed. Reduce them to Debug severity.
* Merge pull request #3885 from ReinUsesLisp/viewport-swizzlesbunnei2020-05-081-0/+6
|\ | | | | video_core: Implement viewport swizzles with NV_viewport_swizzle
| * maxwell_3d: Add viewport swizzlesReinUsesLisp2020-05-041-0/+6
| |
* | Merge pull request #3815 from FernandoS27/command-list-2bunnei2020-05-051-16/+12
|\ \ | |/ |/| GPU: More optimizations to GPU Command List Processing and DMA Copy Optimizations
| * VideoCore/Engines: Refactor Engines CallMethod.Fernando Sahmkow2020-04-281-16/+12
| |
* | {maxwell_3d,buffer_cache}: Implement memory barriers using 3D registersReinUsesLisp2020-04-281-0/+4
|/ | | | | | | | | | | | | Drop MemoryBarrier from the buffer cache and use Maxwell3D's register WaitForIdle. To implement this on OpenGL we just call glMemoryBarrier with the necessary bits. Vulkan lacks this synchronization primitive, so we set an event and immediately wait for it. This is not a pretty solution, but it's what Vulkan can do without submitting the current command buffer to the queue (which ends up being more expensive on the CPU).
* Maxwell3D: Process Macros on MultiMethod.Fernando Sahmkow2020-04-231-25/+47
|
* DMAPusher: Propagate multimethod writes into the engines.Fernando Sahmkow2020-04-231-0/+52
|
* FenceManager: Manage syncpoints and rename fences to semaphores.Fernando Sahmkow2020-04-221-2/+2
|
* Rasterizer: Document SignalFence & ReleaseFences and setup skeletons on Vulkan.Fernando Sahmkow2020-04-221-1/+0
|
* GPU: Fix rebase errors.Fernando Sahmkow2020-04-221-4/+3
|
* OpenGL: Implement Fencing backend.Fernando Sahmkow2020-04-221-11/+5
|
* GPU: Delay Fences.Fernando Sahmkow2020-04-221-1/+9
|
* GPU: Refactor synchronization on Async GPUFernando Sahmkow2020-04-221-2/+6
|
* maxwell_3d: Initialize format attributes constant as oneReinUsesLisp2020-04-171-0/+4
| | | | nouveau expects this to be true but it doesn't set it.
* apply replay logic to all writes. remove replay from MacroInterpreter::Send (@fincs)namkazy2020-03-221-6/+9
|
* maxwell_3d: init shadow_statenamkazy2020-03-221-0/+2
|
* maxwell_3d: this seem more correct.namkazy2020-03-221-2/+2
|
* maxwell_3d: update comments for shadow ram usagenamkazy2020-03-221-1/+1
|
* maxwell_3d: track shadow ram ctrl and hw reg valueNguyen Dac Nam2020-03-221-0/+10
|
* const_buffer_engine_interface: Store component typesReinUsesLisp2020-03-091-1/+1
| | | | | This is required for Vulkan. Sampling integer textures with float handles is illegal.
* maxwell_3d: Flatten cull and front face registersReinUsesLisp2020-02-281-3/+3
|
* video_core: Reintroduce dirty flags infrastructureReinUsesLisp2020-02-281-1/+13
|
* gl_rasterizer: Remove dirty flagsReinUsesLisp2020-02-281-180/+1
|
* Merge pull request #3414 from ReinUsesLisp/maxwell-3d-drawbunnei2020-02-191-2/+2
|\ | | | | maxwell_3d: Unify draw methods
| * maxwell_3d: Unify draw methodsReinUsesLisp2020-02-141-2/+2
| | | | | | | | | | Pass instanced state of a draw invocation as an argument instead of having two separate virtual methods.
* | gl_query_cache: Optimize query cacheReinUsesLisp2020-02-141-3/+8
| | | | | | | | Use a custom cache instead of relying on a ranged cache.
* | gl_query_cache: Implement host queries using a deferred cacheReinUsesLisp2020-02-141-18/+23
| | | | | | | | | | | | | | | | | | | | | | | | | | | | Instead of waiting immediately for executed commands, defer the query until the guest CPU reads it. This way we get closer to what the guest program is doing. To archive this we have to build a dependency queue, because host APIs (like OpenGL and Vulkan) use ranged queries instead of counters like NVN. Waiting for queries implicitly uses fences and this requires a command being queued, otherwise the driver will lock waiting until a timeout. To fix this when there are no commands queued, we explicitly call glFlush.
* | maxwell_3d: Slow implementation of passed samples (query 21)ReinUsesLisp2020-02-141-13/+26
|/ | | | Implements GL_SAMPLES_PASSED by waiting immediately for queries.
* GPU: Address Feedback.Fernando Sahmkow2020-02-131-5/+2
|
* GPU: Implement GPU Clock correctly.Fernando Sahmkow2020-02-101-1/+2
|
* Maxwell3D: Correct query reporting.Fernando Sahmkow2020-02-101-44/+49
|
* Shader_IR: Allow constant access of guest driver.Fernando Sahmkow2020-01-241-0/+4
|
* GPU: Implement guest driver profile and deduce texture handler sizes.Fernando Sahmkow2020-01-241-0/+4
|
* vk_graphics_pipeline: Set front facing properlyReinUsesLisp2020-01-181-0/+1
| | | | | | Front face was being forced to a certain value when cull face is disabled. Set a default value on initialization and drop the forcefully set front facing value with culling disabled.
* yuzu: Remove Maxwell debuggerReinUsesLisp2020-01-031-31/+0
| | | | | This was carried from Citra and wasn't really used on yuzu. It also adds some runtime overhead. This commit removes it from yuzu's codebase.
* gl_rasterizer: Implement RASTERIZE_ENABLEReinUsesLisp2019-12-181-3/+3
| | | | | | | | RASTERIZE_ENABLE is the opposite of GL_RASTERIZER_DISCARD. Implement it naturally using this. NVN games expect rasterize to be enabled by default, reflect that in our initial GPU state.
* video_core: Unify ProgramType and ShaderStage into ShaderTypeReinUsesLisp2019-11-231-16/+17
|
* texture_cache: Use a table instead of switch for texture formatsReinUsesLisp2019-11-151-8/+0
| | | | | | Use a large flat array to look up texture formats. This allows us to properly implement formats with different component types. It should also be faster.
* video_core: Silence implicit conversion warningsReinUsesLisp2019-11-081-1/+2
|
* maxwell_3d/kepler_compute: Remove unused arguments in GetTextureReinUsesLisp2019-10-281-17/+4
|
* video_core/textures: Remove unused index entry in FullTextureInfoReinUsesLisp2019-10-281-1/+0
|
* maxwell_3d: Remove unused method GetStageTexturesReinUsesLisp2019-10-281-39/+0
|
* maxwell_3d: Silence implicit conversion warningsReinUsesLisp2019-10-271-24/+23
| | | | While we are at it, unify types for dirty reg pointers.
* Shader_IR: allow lookup of texture samplers within the shader_ir for instructions that don't provide itFernando Sahmkow2019-10-251-0/+18
|
* VideoCore: Unify const buffer accessing along engines and provide ConstBufferLocker class to shaders.Fernando Sahmkow2019-10-251-1/+2
|
* maxwell_3d: Reduce FlushMMEInlineDraw logging to TraceReinUsesLisp2019-10-201-1/+1
|
* maxwell_3d: Silence truncation warningsLioncash2019-10-151-1/+2
| | | | | A trivial warning caused by not using size_t as the argument types instead of u32.
* maxwell_3d: Add dirty flags for depth bounds valuesReinUsesLisp2019-10-051-0/+5
| | | | | This is useful in Vulkan where we want to update depth bounds without caring if it's enabled or disabled through vkCmdSetDepthBounds.
* Merge pull request #2870 from FernandoS27/multi-drawDavid2019-09-221-1/+101
|\ | | | | Implement a MME Draw commands Inliner and correct host instance drawing
| * Maxwell3D: Corrections and refactors to MME instance refactorFernando Sahmkow2019-09-221-31/+38
| |
| * Rasterizer: Refactor and simplify DrawBatch Interface.Fernando Sahmkow2019-09-191-2/+2
| |
| * VideoCore: Corrections to the MME Inliner and removal of hacky instance management.Fernando Sahmkow2019-09-191-8/+27
| |
| * Video Core: initial Implementation of InstanceDraw PackagingFernando Sahmkow2019-09-191-0/+74
| |
* | Mark DrawArrays as LOG_TRACEDavid Marcec2019-09-211-1/+1
|/ | | | There's no reason to clog logs with DrawArray.
* Merge pull request #2851 from ReinUsesLisp/srgbFernando Sahmkow2019-09-151-0/+3
|\ | | | | renderer_opengl: Fix sRGB blits
| * renderer_opengl: Fix sRGB blitsReinUsesLisp2019-09-111-0/+3
| | | | | | | | | | | | | | | | | | Removes the sRGB hack of tracking if a frame used an sRGB rendertarget to apply at least once to blit the final texture as sRGB. Instead of doing this apply sRGB if the presented image has sRGB. Also enable sRGB by default on Maxwell3D registers as some games seem to assume this.
* | Merge pull request #2824 from ReinUsesLisp/mmeFernando Sahmkow2019-09-151-0/+12
|\ \ | |/ |/| Revert "Revert #2466" and stub FirmwareCall 4
| * maxwell_3d: Update firmware 4 call stub commentaryRodrigo Locatti2019-09-151-1/+2
| |
| * Revert "Revert #2466" and stub FirmwareCall 4ReinUsesLisp2019-09-041-0/+11
| |
* | maxwell_3d: Avoid moving macro_paramsReinUsesLisp2019-09-041-3/+4
|/
* maxwell_3d: Fix macro binding cursorReinUsesLisp2019-09-011-9/+3
|
* video_core: Silent miscellaneous warnings (#2820)Rodrigo Locatti2019-08-301-5/+5
| | | | | | | | | | | | | | | | | | | | | | | | * texture_cache/surface_params: Remove unused local variable * rasterizer_interface: Add missing documentation commentary * maxwell_dma: Remove unused rasterizer reference * video_core/gpu: Sort member declaration order to silent -Wreorder warning * fermi_2d: Remove unused MemoryManager reference * video_core: Silent unused variable warnings * buffer_cache: Silent -Wreorder warnings * kepler_memory: Remove unused MemoryManager reference * gl_texture_cache: Add missing override * buffer_cache: Add missing include * shader/decode: Remove unused variables
* Merge pull request #2592 from FernandoS27/sync1bunnei2019-07-261-2/+3
|\ | | | | Implement GPU Synchronization Mechanisms & Correct NVFlinger
| * video_core: Implement GPU side SyncpointsFernando Sahmkow2019-07-051-2/+3
| |
* | Merge pull request #2704 from FernandoS27/conditionalbunnei2019-07-241-0/+44
|\ \ | | | | | | maxwell3d: Implement Conditional Rendering
| * | maxwell3d: Implement Conditional RenderingFernando Sahmkow2019-07-171-0/+44
| |/ | | | | | | | | | | Conditional Rendering takes care of conditionaly clearing or drawing depending on a set of queries. This PR implements the query checks to stablish if things can be rendered or not.
* | Maxwell3D: Reorganize and address feedbackFernando Sahmkow2019-07-201-0/+4
| |
* | Maxwell3D: Address FeedbackFernando Sahmkow2019-07-171-10/+7
| |
* | GL_Rasterizer: Corrections to Clearing.Fernando Sahmkow2019-07-171-1/+1
| |
* | Maxwell3D: Correct marking dirtiness on CB uploadFernando Sahmkow2019-07-171-0/+1
| |
* | GL_Rasterizer: Rework RenderTarget/DepthBuffer clearingFernando Sahmkow2019-07-171-1/+0
| |
* | Maxwell3D: Implement State Dirty Flags.Fernando Sahmkow2019-07-171-0/+82
| |
* | Maxwell3D: Rework CBData UploadFernando Sahmkow2019-07-171-8/+34
| |
* | Maxwell3D: Rework the dirty system to be more consistant and scaleableFernando Sahmkow2019-07-171-47/+89
|/
* maxwell_3d: Partially implement texture buffers as 1D texturesReinUsesLisp2019-06-211-8/+4
|
* video_core/engines: Move ConstBufferInfo out of Maxwell3DReinUsesLisp2019-06-081-3/+1
|
* Merge pull request #2472 from FernandoS27/ticHexagon122019-05-191-1/+1
|\ | | | | maxwell_3d: reduce severity of different component formats assert.
| * maxwell_3d: reduce sevirity of different component formats assert.Fernando Sahmkow2019-05-141-1/+1
| | | | | | | | | | | | | | This was reduced due to happening on most games and at such constant rate that it affected performance heavily for the end user. In general, we are well aware of the assert and an implementation is already planned.
* | video_core/engines/maxwell3d: Get rid of three magic values in CallMethod()Lioncash2019-05-141-3/+3
| | | | | | | | We can use the named constant instead of using 32 directly.
* | video_core/engines/maxwell_3d: Simplify for loops into ranged for loops within InitializeRegisterDefaults()Lioncash2019-05-141-15/+15
|/ | | | | | Lessens the amount of code that needs to be read, and gets rid of the need to introduce an indexing variable. Instead, we just operate on the objects directly.
* Corrections and stylingFernando Sahmkow2019-04-231-1/+1
|
* Implement Maxwell3D Data UploadFernando Sahmkow2019-04-231-2/+14
|
* Merge pull request #2411 from FernandoS27/unsafe-gpubunnei2019-04-221-2/+2
|\ | | | | GPU Manager: Implement ReadBlockUnsafe and WriteBlockUnsafe
| * Use ReadBlockUnsafe on TIC and TSC readingFernando Sahmkow2019-04-161-2/+2
| | | | | | | | | | Use ReadBlockUnsafe on TIC and TSC reading as memory is never flushed from host GPU there.
* | Move ConstBufferAccessor to Maxwell3d, correct mistakes and clang format.Fernando Sahmkow2019-04-081-2/+10
| |
* | Implement Bindless Handling on SetupTextureFernando Sahmkow2019-04-081-13/+18
|/
* maxwell_3d: Reduce severity of ProcessSyncPointReinUsesLisp2019-04-061-2/+2
|
* Merge pull request #2317 from FernandoS27/syncbunnei2019-04-061-0/+12
|\ | | | | Implement SyncPoint Register in the GPU.
| * Implement SyncPoint Register in the GPU.Fernando Sahmkow2019-04-061-0/+12
| |
* | video_core/engines: Make memory manager members privateLioncash2019-04-061-2/+2
| | | | | | | | | | These aren't used externally by anything, so they can be made private data members.
* | video_core/engines: Remove unnecessary inclusions where applicableLioncash2019-04-061-2/+1
|/ | | | | | Replaces header inclusions with forward declarations where applicable and also removes unused headers within the cpp file. This reduces a few more dependencies on core/memory.h
* gpu: Rewrite virtual memory manager using PageTable.bunnei2019-03-211-4/+4
|
* video_core: Refactor to use MemoryManager interface for all memory access.bunnei2019-03-161-29/+12
| | | | | | | | | | | # Conflicts: # src/video_core/engines/kepler_memory.cpp # src/video_core/engines/maxwell_3d.cpp # src/video_core/morton.cpp # src/video_core/morton.h # src/video_core/renderer_opengl/gl_global_cache.cpp # src/video_core/renderer_opengl/gl_global_cache.h # src/video_core/renderer_opengl/gl_rasterizer_cache.cpp
* gpu: Use host address for caching instead of guest address.bunnei2019-03-151-1/+4
|
* maxwell_3d: Use std::bitset to manage dirty flagsReinUsesLisp2019-02-261-34/+32
|
* video_core: Remove usages of System::GetInstance() within the enginesLioncash2019-02-161-5/+7
| | | | | Avoids the use of the global accessor in favor of explicitly making the system a dependency within the interface.
* core_timing: Convert core timing into a classLioncash2019-02-161-1/+1
| | | | | | | | | | | Gets rid of the largest set of mutable global state within the core. This also paves a way for eliminating usages of GetInstance() on the System class as a follow-up. Note that no behavioral changes have been made, and this simply extracts the functionality into a class. This also has the benefit of making dependencies on the core timing functionality explicit within the relevant interfaces.
* core_timing: Rename CoreTiming namespace to Core::TimingLioncash2019-02-121-1/+1
| | | | | | Places all of the timing-related functionality under the existing Core namespace to keep things consistent, rather than having the timing utilities sitting in its own completely separate namespace.
* video_core: Assert on invalid GPU to CPU address queriesReinUsesLisp2019-02-031-19/+27
|
* maxwell_3d: Allow sampler handles with TSC id zeroReinUsesLisp2019-02-031-10/+6
|
* maxwell_3d: Allow texture handles with TIC id zeroReinUsesLisp2019-02-031-16/+7
| | | | | Also remove "enabled" field from Tegra::Texture::FullTextureInfo because it would become unused.
* maxwell_3d: Set rt_separate_frag_data to 1 by defaultReinUsesLisp2019-01-221-0/+5
| | | | | | | Commercial games assume that this value is 1 but they never set it. On the other hand nouveau manually sets this register. On ConfigureFramebuffers we were asserting for what we are actually implementing (according to envytools).
* gl_rasterizer_cache: Use dirty flags for the depth bufferReinUsesLisp2019-01-071-0/+10
|
* gl_rasterizer_cache: Use dirty flags for color buffersReinUsesLisp2019-01-071-0/+9
|
* gl_shader_cache: Use dirty flags for shadersReinUsesLisp2019-01-071-0/+8
|
* gpu: Rewrite GPU command list processing with DmaPusher class.bunnei2018-11-271-25/+28
| | | | - More accurate impl., fixes Undertale (among other games).
* Merge pull request #1723 from degasus/dirty_flagsbunnei2018-11-271-0/+16
|\ | | | | gl_rasterizer: Skip VB upload if the state is clean.
| * gl_rasterizer: Skip VB upload if the state is clean.Markus Wick2018-11-171-0/+16
| |
* | maxwell_3d: Initialize rasterizer color mask registers as enabled.bunnei2018-11-211-0/+9
| | | | | | | | - Fixes rendering regression with Sonic Mania.
* | set default value for point size registerRodolfo Bogado2018-11-171-0/+3
| |
* | fix viewport and scissor behaviorRodolfo Bogado2018-11-171-2/+2
|/
* gl_rasterizer: Minor cleanupFrederic L2018-11-131-4/+2
| | | Minor code cleanup from unaddressed feedback in #1654
* Try to fix problems with stencil test in some games, relax translation to opengl enums to avoid crashing and only generate logs of the errors.Rodolfo Bogado2018-11-111-0/+13
|
* Merge pull request #1654 from degasus/dirty_flagsbunnei2018-11-111-0/+8
|\ | | | | gl_rasterizer: Skip VAO binding if the state is clean.
| * gl_rasterizer: Skip VAO binding if the state is clean.Markus Wick2018-11-061-0/+8
| |
* | Implement multi-target viewports and blendingRodolfo Bogado2018-11-051-0/+16
|/
* maxwell_3d: Restructure macro upload to use a single macro code memory.bunnei2018-11-011-8/+18
| | | | | - Fixes an issue where macros could be skipped. - Fixes rendering of distant objects in Super Mario Odyssey.
* global: Use std::optional instead of boost::optional (#1578)Frederic L2018-10-301-5/+5
| | | | | | | | | | | | | | | | * get rid of boost::optional * Remove optional references * Use std::reference_wrapper for optional references * Fix clang format * Fix clang format part 2 * Adressed feedback * Fix clang format and MacOS build
* maxwell_3d: Add code for initializing register defaults.bunnei2018-10-261-1/+19
|
* maxwell_3d: Remove unused variable within ProcessQueryGet()Lioncash2018-10-241-1/+0
|
* engines/maxwell_*: Use nested namespace specifiers where applicableLioncash2018-10-201-4/+2
| | | | | | These three source files are the only ones within the engines directory that don't use nested namespaces. We may as well change these over to keep things consistent.
* Port #4182 from Citra: "Prefix all size_t with std::"fearlessTobi2018-09-151-6/+7
|
* rasterizer: Drop unused handler.Markus Wick2018-09-101-2/+0
| | | | | | | | This virtual function is called in a very hot spot, and it does nothing. If this kind of feature is required, please be more specific and add callbacks in the switch statement within Maxwell3D::WriteReg. There is no point in having another switch statement within the rasterizer.
* maxwell_3d: Remove assert that no longer applies.bunnei2018-09-081-4/+0
|
* maxwell_3d: Use CoreTiming for query timestampZach Hilman2018-09-011-2/+3
|
* core/core: Replace includes with forward declarations where applicableLioncash2018-08-311-2/+1
| | | | | | | | | | | The follow-up to e2457418dae19b889b2ad85255bb95d4cd0e4bff, which replaces most of the includes in the core header with forward declarations. This makes it so that if any of the headers the core header was previously including change, then no one will need to rebuild the bulk of the core, due to core.h being quite a prevalent inclusion. This should make turnaround for changes much faster for developers.
* maxwell3d: Move FinishedPrimitiveBatch event after AcceleratedDrawBatch()Lioncash2018-08-251-4/+4
| | | | | The start and finish events should likely not be right after one another like this, otherwise the batch will appear to complete immediately
* Rasterizer: Implemented instanced rendering.Subv2018-08-151-0/+12
| | | | | | We keep track of the current instance and update an uniform in the shaders to let them know which instance they are. Instanced vertex arrays are not yet implemented.
* maxwell_3d: Ignore macros that have not been uploaded yet.bunnei2018-08-091-4/+9
| | | | - Used by Super Mario Odyssey (in game).
* maxwell_3d: Use correct const buffer size and check bounds.bunnei2018-08-081-0/+2
| | | | - Fixes mem corruption with Super Mario Odyssey and Pokkén Tournament DX.
* maxwell_3d: Remove outdated assert.bunnei2018-08-061-2/+0
|
* video_core: Eliminate the g_renderer global variableLioncash2018-08-041-5/+5
| | | | | | | | | | | | | | We move the initialization of the renderer to the core class, while keeping the creation of it and any other specifics in video_core. This way we can ensure that the renderer is initialized and doesn't give unfettered access to the renderer. This also makes dependencies on types more explicit. For example, the GPU class doesn't need to depend on the existence of a renderer, it only needs to care about whether or not it has a rasterizer, but since it was accessing the global variable, it was also making the renderer a part of its dependency chain. By adjusting the interface, we can get rid of this dependency.
* GPU: Remove the assert that required the CODE_ADDRESS to be 0.Subv2018-07-241-8/+0
| | | | Games usually just leave it at 0 but nouveau sets it to something else. This already works fine, the assert is useless.
* maxwell_3d: Remove unused variable within GetStageTextures()Lioncash2018-07-201-2/+0
|
* gl_rasterizer: Fix check for if a shader stage is enabled.bunnei2018-07-131-21/+0
|
* Merge pull request #609 from Subv/clear_buffersbunnei2018-07-041-0/+12
|\ | | | | GPU: Implemented the CLEAR_BUFFERS register.
| * GPU: Support clears that don't clear the color buffer.Subv2018-07-031-2/+3
| |
| * GPU: Bind and clear the render target when the CLEAR_BUFFERS register is written to.Subv2018-07-031-0/+11
| |
* | Update clang formatJames Rowe2018-07-031-2/+2
| |
* | Rename logging macro back to LOG_*James Rowe2018-07-031-1/+1
|/
* Build: Fixed some MSVC warnings in various parts of the code.Subv2018-06-201-2/+3
|
* GPU: Implement sampling multiple textures in the generated glsl shaders.Subv2018-06-061-0/+34
| | | | | | All tested games that use a single texture show no regression. Only Texture2D textures are supported right now, each shader gets its own "tex_fs/vs/gs" sampler array to maintain independent textures between shader stages, the textures themselves are reused if possible.
* GPU: Partial implementation of long GPU queries.Subv2018-06-041-9/+24
| | | | | | | | Long queries write a 128-bit result value to memory, which consists of a 64 bit query value and a 64 bit timestamp. In this implementation, only select=Zero of the Crop unit is implemented, this writes the query sequence as a 64 bit value, and a 0u64 value for the timestamp, since we emulate an infinitely fast GPU. This specific type was hwtested, but more rigorous tests should be performed in the future for the other types.
* maxwell_3d: Reset vertex counts after drawing.bunnei2018-04-291-0/+10
|
* general: Convert assertion macros over to be fmt-compatibleLioncash2018-04-271-2/+2
|
* GPU: Reduce the number of registers of Maxwell3D to 0xE00.Subv2018-04-251-3/+3
| | | | The rest are just macro shim registers.
* GPU: Move the Maxwell3D macro uploading code to the inside of the Maxwell3D processor.Subv2018-04-251-4/+10
| | | | It doesn't belong in the PFIFO handler.
* video-core: Move logging macros over to new fmt-capable onesLioncash2018-04-251-2/+2
|
* memory_manager: Make GpuToCpuAddress return an optional.bunnei2018-04-241-10/+11
|
* memory_manager: Use GPUVAdddr, not PAddr, for GPU addresses.bunnei2018-04-241-6/+5
|
* GPU: Added asserts to our code for handling the QUERY_GET GPU command.Subv2018-04-241-1/+26
| | | | | This is based on research from nouveau. Many things are currently unknown and will require hwtests in the future. This commit also stubs QueryMode::Write2 to do the same as Write. Nouveau code treats them interchangeably, it is currently unknown what the difference is.
* GPU: Pitch textures are now supported, don't assert when encountering them.Subv2018-04-181-2/+3
|
* Merge pull request #346 from bunnei/misc-gpu-improvementsbunnei2018-04-181-1/+2
|\ | | | | Misc gpu improvements
| * maxwell3d: Allow Texture2DNoMipmap as Texture2D.bunnei2018-04-181-1/+2
| |
* | renderer_opengl: Implement BlendEquation and BlendFunc.bunnei2018-04-181-3/+1
|/
* gl_rasterizer: Implement indexed vertex mode.bunnei2018-04-171-1/+3
|
* GPU: Added a function to determine whether a shader stage is enabled or not.Subv2018-04-151-0/+21
|
* GPU: Assert when finding a texture with a format type other than UNORM.Subv2018-04-071-0/+2
|
* GPU: Use the MacroInterpreter class to execute the GPU macros instead of HLEing them.Subv2018-04-011-100/+8
|
* GPU: Implemented a gpu macro interpreter.Subv2018-04-011-0/+5
| | | | | | The Ryujinx macro interpreter and envydis were used as reference. Macros are programs that are uploaded by the games during boot and can later be called by writing to their method id in a GPU command buffer.
* Maxwell3D: Call AccelerateDrawBatch on DrawArrays.bunnei2018-03-271-1/+8
|
* GPU: Load the sampler info (TSC) when retrieving active textures.Subv2018-03-261-20/+60
|
* GPU: Make the debug_context variable a member of the frontend instead of a global.Subv2018-03-251-11/+13
|
* GPU: Added a function to retrieve the active textures for a shader stage.Subv2018-03-241-44/+49
| | | | TODO: A shader may not use all of these textures at the same time, shader analysis should be performed to determine which textures are actually sampled.
* GPU: Implement the Incoming/FinishedPrimitiveBatch debug breakpoints.Subv2018-03-241-0/+7
|
* GPU: Implement the MaxwellCommandLoaded/Processed debug breakpoints.Subv2018-03-241-0/+10
|
* GPU: Added a method to unswizzle a texture without decoding it.Subv2018-03-241-1/+1
| | | | Allow unswizzling of DXT1 textures.
* GPU: Preliminary work for texture decoding.Subv2018-03-241-0/+45
|
* Clang FixesN00byKing2018-03-191-1/+2
|
* Clean Warnings (?)N00byKing2018-03-191-1/+1
|
* GPU: Implement macro 0xE1A BindTextureInfoBuffer in HLE.Subv2018-03-191-0/+18
| | | | This macro simply sets the current CB_ADDRESS to the texture buffer address for the input shader stage.
* GPU: Implement the BindStorageBuffer macro method in HLE.Subv2018-03-181-0/+21
| | | | | | This macro binds the SSBO Info Buffer as the current ConstBuffer. This buffer is usually bound to c0 during shader execution. Games seem to use this macro instead of directly writing the address for some reason.
* GPU: Handle writes to the CB_DATA method.Subv2018-03-181-0/+36
| | | | | | Writing to this method will cause the written value to be stored in the currently-set ConstBuffer plus CB_POS. This method is usually used to upload uniforms or other shader-visible data.
* GPU: Store uploaded GPU macros and keep track of the number of method parameters.Subv2018-03-181-7/+16
|
* GPU: Macros are specific to the Maxwell3D engine, so handle them internally.Subv2018-03-181-5/+36
|
* GPU: Renamed ShaderType to ShaderStage as that is less confusing.Subv2018-03-181-12/+12
|
* GPU: Store shader constbuffer bindings in the GPU state.Subv2018-03-181-2/+36
|
* GPU: Make the SetShader macro call do the same as the real macro's code.Subv2018-03-181-1/+19
| | | | | | It'll now set the CB_SIZE, CB_ADDRESS and CB_BIND registers when it's called. Presumably this SetShader function is binding the constant shader uniforms to buffer 1 (c1[]).
* GPU: Corrected the parameter documentation for the SetShader macro call.Subv2018-03-171-5/+6
| | | | | | Register 0xE24 is actually a macro that sets some shader parameters in the register structure. Macros are uploaded to the GPU at startup and have their own ISA, we'll probably write an interpreter for this in the future.
* GPU: Handle the SetShader method call (0xE24) and store the shader config.Subv2018-03-171-2/+22
|
* Merge pull request #241 from Subv/gpu_method_callbunnei2018-03-171-0/+17
|\ | | | | GPU: Process command mode 5 (IncreaseOnce) differently from other commands
| * GPU: Process command mode 5 (IncreaseOnce) differently from other commands.Subv2018-03-171-0/+17
| | | | | | | | | | | | Accumulate all arguments before calling the desired method. Note: Maybe we should do the same for the NonIncreasing mode?
* | GPU: Assert that we get a 0 CODE_ADDRESS register in the 3D engine.Subv2018-03-171-0/+8
|/ | | | Shader address calculation depends on this value to some extent, we do not currently know what it being 0 entails.
* GPU: Intercept writes to the VERTEX_END_GL register.Subv2018-03-051-0/+9
| | | | | | This is the register that gets written after a game calls DrawArrays(). We should collect all GPU state and draw using our graphics API here.
* GPU: Partially implemented the QUERY_* registers in the Maxwell3D engine.Subv2018-02-121-1/+39
| | | | Only QueryMode::Write is supported at the moment.
* Make a GPU class in VideoCore to contain the GPU state.Subv2018-02-121-3/+1
| | | | Also moved the GPU MemoryManager class to video_core since it makes more sense for it to be there.
* GPU: Added a command processor to decode the GPU pushbuffers and forward the commands to their respective engines.Subv2018-02-121-0/+15
generated by cgit v1.2.3 (git 2.25.1) at 2024-07-04 09:04:59 +0200