- Emulating f16 with f32 is not ideal and requires a lot of value clamping
- Using native data type can significantly improve performance and accuracy
- With openGL, check for the compatible extensions NV_gpu_shader5 and
AMD_gpu_shader_half_float
- With Vulkan, enable this functionality in the deviceFeatures if
applicable. (VK_KHR_shader_float16_int8 extension)
- Temporarily disable hw fp16 for vulkan
- Removes CPU-only transforms that broke GPU-side code.
-- Channels in GPU compute are laid out in cell-order, but CPU was uploading in favorable order and compensating with swizzles.
-- This leads to 2 different layouts depending on the location of the data (CPU vs GPU)
- Implement R8G8_R8B8 interleaved format decode
- General improvements
- gl: Include an execution state wrapper to ensure state changes are consistent. Also removes a lot of required 'cleanup' for helper methods
- texture_cache: Make execition context a mandatory field as it is required for all operations. Also removes a lot of situations where duplicate argument is added in for both fixed and vararg fields
- Explicit read/write barrier for framebuffer resources depending on
usage. Allows for operations like optional memory initialization before
reading
- Implicitly invoke a memory barrier if actively reading from an unsynchronized texture
- Simplify memory transfer operations
- Should allow more games to work without strict mode
- Implements a mirror view of D24S8 data that accesses the stencil components.
Finishes the implementation of TEX2D_DEPTH_RGBA as the stencil component was previously missing from the reconstructed data
- Add a few missing destructors
Image classes are inherited a lot and I forgot to make the dtors virtual
- Matching attachments with resource id fails because drivers are reusing
handles!
- Properly sets up stale fbo ref counting and removal
- Properly sets up resource reference test with subsequent removal to
avoid using a broken fbo entry
- Forcefully downloads and reuploads data from the CPU in case of unexpected overlaps
- Properly detect correct size of newly created blit targets
- Remember to clear any existing views when changing the default component map!
- Ignore unlocked blit sections [TODO]
- Do not attempt blit on hw if bytesize is unsupported
- gl: Implement typeless memory transfers
Uses pbo to handle type-agnostic memory transfer
gl/vk: Bump shader cache version
gl/vk: Disable anisotropic override when strict mode enabled as it is proven to alter some games negatively
gl: Clamp buffer view range to not exceed the backing buffer size. Also add assert for the same condition
- gl: Do not call makeCurrent every flip - it is already called in set_current()
- gl: Improve ring buffer behaviour; use sliding window to view buffers larger than maximum viewable hardware range
NV hardware can only view 128M at a time
- gl/vk: Bump transform constant heap size When lots of draw calls are issued, the heap is exhaused very fast (8k per draw)
- gl: Remove CLIENT_STORAGE_BIT from ring buffers. Performance is marginally better without this flag (at least on windows)
- Implement flush-always behaviour to partially fix readback from a currently bound fbo
- Without this, only the first read is correct, as more draws are added the results become 'wrong'
- Fixes WCB and cpublit behviour
- Synchronize blit_dst surfaces to avoid data loss when gpu texture scaling is used
- Its still faster in such cases to disable gpu texture scaling but some types cannot be disabled without force cpu blit (e.g framebuffer transfers)
- Memory management tuning
- rsx: on-demand texture cache rescanning for unprotected sections
- rsx: Only framebuffer resources are upscaled
- Do not resize regular blit engine resources
- Lazy initialize readback buffer when using opengl
-- These measures should help minimize vram usage
- opengl driver optimization for nvidia. On nvidia glTextureBufferRange performance is horrendous
-- Initialize texture buffer to whole buffer at startup and use absolute offsets to read data instead
-- Over 2x performance in some cases (Resogun, TNT racers)
- gl/vk: Do not flip non-existent display buffers. Fixes spec violation at boot in TNT racers demo
- whitespace fixes for sys_rsx
- Handle aliased depth + color target by disabling depth writes. This looks to be the correct way
- Add support for generic passes that cannot be done using general imaging operations. Lays the framework for tons of features and effects
- Implement RGBA->D24D8 casting. Sometimes games will split depth texture into RGBA8 then use the new RGBA8 as a depth texture directly
-- This happens alot in ps3 games and I'm not sure why. Its likely the ps3 did not sample fp values with linear filtering so this is a workaround
-- Only implemented for openGL at the moment
-- Requires a workaround for an AMD driver bug
gl/vk/rsx: Refactoring; unify texture cache code
gl: Fixups
- Removes rsx::gl::texture class and leave gl::texture intact
- Simplify texture create and upload mechanisms
- Re-enable texture uploads with the new texture cache mechanism
rsx: texture cache - check if bit region fits into dst texture before attempting to copy
gl/vk: Cleanup
- Set initial texture layout to DST_OPTIMAL since it has no data in it anyway at the start
- Move structs outside of classes to avoid clutter
rsx: Blit engine improvements
- Always handle blits to and from framebuffers through the GPU
- Handle depth surfaces properly when using GL
- Check for format mismatches when blitting to the surface store [WIP]
Fix rsx offscreen-render-to-display-buffer-blit surface reads
- Also, properly scale display output height if reading from compressed tile
gl: Fix broken dst height computation
- The extra padding is only there to force power-of-2 sizes and isnt used
gl: Ignore compression scaling if output is rendered to in a renderpass
rsx/gl/vk: Cleanup for GPU texture scaling. Initial impl [WIP]
- TODO: Refactor more shared code into RSX/common
- Significant gains from greatly reduced CPU work
- Also reorders command submission in end() to improve throughput
- Refactors most of the vertex buffer handling
- All vertex processing is moved GPU side
