| Commit message (Collapse) | Author | Age | Files | Lines |
|---|
| |
|
|
|
|
|
|
|
|
|
| |
Streamline may report that DLSS needs no extra device extensions. The check required a non-empty
list, so that result counted as "extensions unavailable": Streamline was stopped and FSR 3.1 used.
Only require that every listed extension is supported.
Found by CodeRabbit review.
Assisted-by: Claude Opus 5.5 (implementation)
Signed-off-by: Duck912 <350528965+Duck912@users.noreply.github.com>
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
The Streamline start-up and DLSS download code formatted paths with std::filesystem::path::string(),
which on MSVC converts to the active code page and throws when a character has no mapping there,
for example a user name in another script under %APPDATA%. The start-up log runs inside the
renderer's constructor, which only catches vk::Exception, so such a user would crash at game start
with the setting on.
Use Common::FS::PathToUTF8String for every path in these logs and messages.
Found by CodeRabbit review.
Assisted-by: Claude Opus 5.5 (implementation)
Signed-off-by: Duck912 <350528965+Duck912@users.noreply.github.com>
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
The upscaler records its work on the worker thread and reports success once it is queued. When
Streamline then rejects the DLSS options, gives no frame token or fails the evaluation, nothing is
written to the output image, but the caller still draws it into the game's image that frame:
undefined contents after a resize, the previous frame otherwise. FSR 3.1 takes over from the next
frame.
Clear the output on those paths, with barriers around the clear, so that frame is black instead.
Found by CodeRabbit review.
Assisted-by: Claude Opus 5.5 (implementation)
Signed-off-by: Duck912 <350528965+Duck912@users.noreply.github.com>
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Adds a Graphics setting, "Use NVIDIA DLSS in place of a game's DLSS" (nvidia_dlss, off by
default). When it is on, in Windows builds configured with EGLANTINE_STREAMLINE_SDK, the
evaluation Eglantine already replaces with FSR 3.1 goes to NVIDIA DLSS on GeForce RTX cards, with
the same inputs (colour, depth, motion vectors, jitter). When DLSS can't run, FSR 3.1 runs as
before.
Only Streamline's MIT-licensed headers are compiled in; nothing of NVIDIA's is linked or shipped.
At run time Eglantine looks for sl.interposer.dll in the user folder's "streamline" folder (or
EDEN_STREAMLINE_DIR, or "streamline" next to the executable), checks its Streamline signature and
starts Streamline with manual hooking. Streamline's plugins come from that folder or the NVIDIA
app's downloads. DLSS itself is nvngx_dlss.dll in the same folder, else the newest DLSS 310 or
later that the NVIDIA app has downloaded, checked for NVIDIA's Authenticode signature and copied
as streamline/nvidia-app/nvngx_dlss.dll. Like other tools that call DLSS outside a game, it uses
its own fixed project id with the custom engine type.
- Preset K for all modes (EDEN_NX2_DLSS_PRESET overrides it): the default Performance model (M)
and L draw thin dark horizontal lines in some frames at 2x resolution scale; K doesn't.
- HDR input with auto exposure; the mode follows the render/output ratio.
- Buffer device address stays enabled when DLSS needs it, for DLSS only.
- The upscaler waits for the device instead of submitting when it is destroyed: by then the
game's channels are released, and a submit read the state tracker's stale flags (a crash when
stopping a game).
With the setting off, or in builds without the SDK path (the setting is hidden there), nothing
changes.
Testing: Windows 11, RTX 5080, Fire Emblem: Fortune's Weave docked at 1x (720p -> 1440p) and 2x
(1440p -> 2880p), CLI runs and GUI play, stopping and restarting games with DLSS on. Not tested on
other GPUs or games.
Assisted-by: Claude Opus 5.5 (implementation, debugging)
Signed-off-by: Duck912 <350528965+Duck912@users.noreply.github.com>
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
BlitColor(VkImageView, VkImage, ...), which draws FSR 3.1's output into the game's DLSS output,
captured the Framebuffer pointer and read its render pass, handle and size on the worker thread.
The pointer points into the texture cache's framebuffer slots, which can be reallocated before the
worker records the command, so the render pass could begin with a stale or null render pass and
framebuffer. FSR 3.1 records quickly enough that this rarely shows; with a slower first evaluation
(NVIDIA DLSS creating its feature) it crashed in vkCmdBeginRenderPass at the first evaluation in
every run.
Read the three values on the calling thread and capture those instead. Nothing else changes.
Assisted-by: Claude Opus 5.5 (debugging, implementation)
Signed-off-by: Duck912 <350528965+Duck912@users.noreply.github.com>
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
* video_core: demote the texture cache's per-draw debug logs to trace
The invalid sampler and image view index messages fire for unused
binding slots on every draw. Filtered LOG_DEBUG messages still cost a
call and argument packing: in a Fire Emblem: Fortune's Weave city scene
(about 2,100 draws per frame) they were 3.1% of the GPUBack thread's
samples. LOG_TRACE compiles them out of release builds.
Assisted-by: Claude Opus 5.5 (profiling, porting and drafting)
Signed-off-by: Duck912 <350528965+Duck912@users.noreply.github.com>
* texture_cache: read a descriptor with one page lookup
DescriptorTable::Read runs for every texture and sampler of every draw,
and it read each 32-byte TIC/TSC entry through the generic
ReadBlockUnsafe page walk. In a Fire Emblem: Fortune's Weave city scene
(about 2,100 draws per frame) that was 5.0% of the GPUBack thread's
samples. An entry within one page is contiguous in host memory, so it
is now a single page lookup and memcpy. Entries that cross a page and
pages without host memory still take the block read, so unmapped and
unbacked entries still read as zero.
The same change in another Eden-based build cut these reads from 6.4%
to 0.4% of its GPU thread.
Assisted-by: Claude Opus 5.5 (profiling, porting and drafting)
Signed-off-by: Duck912 <350528965+Duck912@users.noreply.github.com>
---------
Signed-off-by: Duck912 <350528965+Duck912@users.noreply.github.com>
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
(#361)
* buffer_cache: don't end the render pass after mapped buffer uploads
Every upload set any_buffer_uploaded, and the next draw or dispatch then
recorded a transfer-to-shader barrier outside the render pass. On Vulkan
the uploads are mapped: each copy either records its own barriers
(CopyBuffer with barrier set) or goes to the upload command buffer,
which ends with a barrier before the frame's commands run. The extra
barrier only ended the render pass.
On tiled GPUs every such split writes the attachments out and reads
them back. Pokopia's first gameplay scene had ~50 of these splits a
frame. Without them, an M3 MacBook Air (MoltenVK, 1x docked) goes from
~35 to ~29.5 ms of GPU time a frame and from 28.5 to 32 fps.
OpenGL uploads immediately and keeps the barrier.
Assisted-by: Claude Opus 5.5 (profiling, implementation and commit message)
* vulkan: log a failed GPU wait before it ends the process
When the device is lost, vkWaitSemaphores in MasterSemaphore::Wait
returns VK_ERROR_DEVICE_LOST and the exception goes uncaught on the
fence thread or the GPU thread, so the process aborts without a line in
the log: ReportLoss is only reached from submit and present. A Pokopia
crash on macOS left the log ending a second before the abort, with no
hint that the GPU was lost.
Log the wait's result before rethrowing. On macOS, MoltenVK prints the
Metal reason (a GPU hang or page fault) to stderr only.
Assisted-by: Claude Opus 5.5 (diagnosis, implementation and commit message)
|
| |
|
|
|
|
| |
Fsr3Upscaler::Pass::Dispatch took a scheduler parameter that shadows
Fsr3Upscaler::scheduler, which fails the Windows MinGW PGO build under
-Werror=shadow.
|
| |\
| |
| | |
FSR 3.1 implementation for DLSS games
|
| | |\
| | |
| | |
| | |
| | |
| | | |
# Conflicts:
# src/video_core/renderer_vulkan/blit_image.cpp
# src/video_core/renderer_vulkan/blit_image.h
|
| | | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
Fire Emblem: Fortune's Weave's DLSS network can't run here, and the
temporal stand-in is only an approximation of it. FSR 3.1 takes the
same inputs DLSS does: the jittered colour, motion vectors, the jitter
and the depth, which the game's first DLSS pass reads through
cbuf0+0x298 (Z32, reversed, the colour's size). Kepler compute keeps it
next to the motion vectors.
Fsr3Upscaler ports the host side of the SDK's ffx_fsr3upscaler.cpp: the
eight passes (prepare inputs, luma and shading-change pyramids, shading
change, reactivity, luma instability, accumulate, RCAS), the internal
images and the constants, recorded through the scheduler with the
shaders from externals/FidelityFX-SDK. It takes HDR colour, render-size
motion vectors and reversed depth, computes its own exposure, and uses
the no-wave SPD variants on devices without compute quad operations.
The SDK's GLSL uses separate textures and samplers, so DescriptorBankInfo
gains sampled image and sampler counts (default 0). FSR's RCAS follows
the RCAS Sharpness slider. The result goes to FSR's own RGBA16F image
and is drawn into the game's output, which keeps its format.
TextureCacheRuntime::TemporalUpscale runs FSR 3.1 first. Without the
depth, at mismatched scales, on a device lacking FSR's formats or
features, or if its pipelines fail to build, the stand-in runs instead,
and the bilinear blit stays last. EDEN_NX2_DLSS_REPLACEMENT=standin
forces the stand-in. A gap of more than 250 ms between evaluations
restarts the history.
The BlitColor overload that takes a raw image view transitions its
source to SHADER_READ_ONLY_OPTIMAL but described it as GENERAL in the
descriptor (a validation error); it now describes the layout it uses.
Assisted-by: Claude Opus 5.5 (debugging, implementation and drafting)
Signed-off-by: Duck912 <350528965+Duck912@users.noreply.github.com>
|
| | | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
Fire Emblem: Fortune's Weave's DLSS input matches Eden's (same 720p
image, same detail), but the stand-in's 1440p output was blurrier than a
plain bilinear 2x of it (Laplacian 0.0066 vs 0.0081; Eden's real DLSS
0.0111), and the image drifted darker over time:
- the sample kernel was about twice too wide for the output resolution;
- the history was kept in the output's B10G11R11 format, whose 5-6 bit
mantissas round away a frame's 0.03-0.1 share;
- the 1-sigma neighbourhood clip removed detail that only some jitter
positions sample (railings, thin edges).
The history is now an RGBA16F pair the shader writes as a storage image,
with no per-frame copy. Still pixels use a third of the kernel width and
a 1.5-sigma clip, so detail from many jitter positions accumulates.
Moving pixels keep the old kernel and 1-sigma clip, and so do pixels
that moved recently or whose history falls far outside the
neighbourhood (content that changes without motion vectors), remembered
in the history's alpha. Static output goes from 0.0066 to 0.0102 with
no brightness drift and no new ghosting.
With resolution scaling the stand-in used to fall back to a bilinear
blit. The images are now scaled up as BlitImage does, the jitter is
scaled to host pixels and the history follows the framebuffer's render
area.
RCAS (temporal_sharpen.frag) sharpens a copy of each result, never the
history, in range-compressed space since the output is HDR. Its strength
is the sharpness slider, relabelled "RCAS Sharpness" (0% turns it off).
EDEN_NX2_TEMPORAL_KERNEL, _CLIP, _KERNEL_MOTION, _CLIP_MOTION and
_WEIGHT tune the filter (all four filter values at 1 give the old one),
and EDEN_NX2_TEMPORAL_SHARPEN=<stops>|off overrides the slider.
Assisted-by: Claude Opus 5.5 (debugging, implementation and drafting)
Signed-off-by: Duck912 <350528965+Duck912@users.noreply.github.com>
|
| | | |
| | |
| | |
| | |
| | |
| | |
| | | |
FFmpeg 6.1's swscale leaves the chroma plane of an NV12 to NV12 copy at
zero, so movies decoded on the host GPU (Mario Kart World's pre-race
movie, Hyrule Warriors' cutscenes) showed green. CPU-decoded frames
(YUV420P) were not affected.
|
| | | |
| | |
| | |
| | |
| | |
| | | |
stream is H.264 (Yoshi demo)
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
|
| | |/
|/|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
- [x] I have read and followed the [Contribution Guidelines](https://git.eden-emu.dev/eden-emu/eden/src/branch/master/CONTRIBUTING.md#code-contributions).
- [x] I have read and followed the [Coding Guidelines](https://git.eden-emu.dev/eden-emu/eden/src/branch/master/docs/policies/Coding.md) to the best of my ability.
-------------------
https://git.eden-emu.dev/eden-emu/eden/pulls/4542
Reviewed-on: https://git.eden-emu.dev/eden-emu/eden/pulls/4590
Reviewed-by: crueter <crueter@eden-emu.dev>
Reviewed-by: John <john@eden-emu.dev>
(cherry picked from commit 0731f40f4d86227f05da3fffbc44831065b53a8a)
Co-Authored-By: Claude Code <noreply@anthropic.com>
(cherry picked from commit 7c3821da0e678978d55df14d65e047539eafba4c)
|
| |\ \
| | |
| | | |
texture_cache: look up X8Z24 with UNORM, UINT, UNORM, UNORM components
|
| | | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
X8Z24 was only known with UINT in the last three components. Mario
Tennis Fever samples its depth buffer as X8Z24 with components
UNORM, UINT, UNORM, UNORM (as S8Z24 already is in the table), which
hit the "texture format=42" fallback to A8B8G8R8. The depth image was
then copied into a colour image of the wrong format every frame (with
"Unimplemented" layer-count asserts) and read as colour, and distant
scenery around the court (the arch and the castle stadium) wasn't
drawn.
Map it to X8_D24_UNORM like the other X8Z24 entry.
Assisted-by: Claude Opus 5.5 (diagnosis, fix and testing)
|
| |\ \ \
| | | |
| | | | |
video_core: bump the shader cache versions for the FSEL lowering fix
|
| | | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
The disk shader cache keeps the lowered native code. "shader: apply
FSEL's floating operand modifiers" (ef3ae4ae23) changed that code
without a version bump, so caches built before it keep programs that
drop FSEL's negate until the user deletes them by hand.
In Splatoon Raiders those stale programs leave the tutorial island and
the dock with no shadows (the depth linearization ends with
FSEL R0, -R0, 1.0, P and comes out negated), no aiming reticle and no
steam (particle quads pick their corners with FSEL +0.5 / -0.5 and
collapse to a point).
|
| |\ \ \ \
| | | | |
| | | | | |
MKW performance improvements
|
| | | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | | |
Port the coherency changes from alarok/Eglantine 77fc365f2 without its diagnostic framework or shutdown changes. Only GPU-dirty CPU reads become permanently preflushable, clean areas can be trapped on the next write, non-core readers avoid indexing core arrays, and collected CPU writes precede readback.
Assisted-by: Codex (source review, selective integration and verification)
(cherry picked from commit b64e211957180aa218ec6d9a3bcb020332279c3f)
|
| | | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | | |
Adapted from sylvieistaken/selevet-eglantine performance builds, including 8b076b167ffc5ca1ba8ffedcd00e255accbe018c and d8c95fef. Preserve upstream command-list coherency and submission ordering.
Assisted by Codex.
|
| | | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | | |
Avoid zeroing unused preallocation during invalidation and pool refresh. Clear newly exposed words whenever limits grow, bound clearing to allocated storage, and test word/allocation boundaries, shrink/regrowth and malformed large limits.
Assisted-by: Codex (implementation, review, commit text and automated validation)
(cherry picked from commit 4bc1ae01f79cef43be016504c850ca8a6998457a)
|
| | | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | | |
Exercise 40,000 mutations and table switches plus repeated deletion across five channels. Make the existing descriptor read work with a memory reader fixture and add an explicit CPU benchmark with opaque invalidation calls.
Assisted-by: Codex (implementation, review, commit text and automated validation)
(cherry picked from commit 3b62a551da8ec0b4edf19a1c4e14fa2519dd20e2)
|
| | | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | | |
Keep the linear path for sixteen or fewer removals. For larger batches, sort the auxiliary pointer list and binary-search it while preserving storage order and the same invalidation decisions. Existing real-cache differential tests cover both paths.
Assisted-by: Codex (implementation, review, commit text and automated validation)
(cherry picked from commit 7e2c8a535230d9810db07706d79130ce0f1b88f1)
|
| | | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | | |
Adapt the batching from ymatrix1/Eglantine PR #237, retaining all CPU invalidations. Acquire the lock once and remove pending shaders once per write batch. Compare real cache contents with sequential invalidation across randomized overlapping writes and pending removals.
Assisted-by: Codex (implementation, review, commit text and automated validation)
(cherry picked from commit b65f818d7789ffadfb58fe4543b18060fd88ce53)
|
| | | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | | |
command buffer
(cherry picked from commit 4aa5cd89fd1862751b3297036f2d21a0dc924b23)
(cherry picked from commit f1bdd5b13a1bdca81243b4aa1d8529f0485d3671)
|
| | | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | | |
Keep error reporting and debug-build history. EDEN_NX2_3D_HISTORY enables the full ring in release builds without recording every method during normal play.
Assisted by Codex.
(cherry picked from commit cac6081916e8e5461e2754308cd8986deaec774b)
|
| | | |_|/
| |/| |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
Use the owned parameter vector for DMA FIFO appends. Add a native JIT regression covering reallocation and repeated invocations against the copied-input path.
Assisted by Codex.
(cherry picked from commit 5b6803aaab18e332493903ee49cab026c041d5dc)
|
| |\ \ \ \
| | | | |
| | | | | |
nx2: fix pokopia regressions, textures that handle shader computes, and more
|
| | | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | | |
837133792d queues the GMMU maps on the GPU thread (GPU::RunInOrder) so the
service thread never waits for the texture cache. With the GPU split
(Low GPU accuracy) the GPU thread defers queued functions to its back, but
its front reads the command lists: it read the lists submitted after a map
through GPU addresses that weren't mapped yet. Pokopia stopped after its
first frame.
Add GPU::MapInOrder for the GMMU maps and the sparse tile remaps. It runs
in order like RunInOrder, and with the split it runs on the front after
draining the back, before the front reads further command lists. The
service thread still never takes the texture cache lock, and the back is
idle while the map changes the caches. Unmaps and the other queued work
still run on the back.
Assisted-by: Claude Opus 5.5 (bisect, fix, testing and commit message)
|
| | | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | | |
e38b738b84 repacks the 69-byte vertex stride to 72 only with dynamic
vertex input, unless EGLANTINE_EXPERIMENTAL_STATIC_VERTEX_REPACK=1.
MoltenVK has no VK_EXT_vertex_input_dynamic_state and Metal needs strides
that are a multiple of 4, so the stride was never fixed on Macs.
Enable the static path automatically when the device's minimum vertex
binding stride alignment does not divide 69. The device reports 4 for
MoltenVK, matching its portability subset limit. The environment variable
still overrides it: 1 forces the repack on, 0 forces it off.
Assisted-by: Claude Sonnet 5.5 (implementation and commit message)
|
| | | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | | |
5de94bdd1d cut the boot preload to the last two sessions and 2f94fb9680
evicts pipelines unused for 9000 frames, both because rebuilding one
costs 2-3 ms through the driver's pipeline cache and each holds ~0.45 MB
of driver memory. On MoltenVK a pipeline built while playing is
translated from SPIR-V to MSL and compiled by Metal, which takes far
longer even with a VkPipelineCache and shows as a stutter.
On MoltenVK, keep the previous eight sessions' pipelines in
vulkan_used.bin (be76402ab5's window) so they are built at boot, and
do not evict. The memory is bounded by what a title builds; the stutter
is not. Other drivers are unchanged and
EDEN_NX2_PRELOAD_ALL_PIPELINES=1 still builds every cached pipeline.
Assisted-by: Claude Sonnet 5.5 (implementation and commit message)
|
| | | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | | |
In the aliased mode (AMD) a SparseTiles image is created unbound, and only
InitSparseTiles binds every block to the shared zeroed one. When it fails
(an unusual block shape, no memory for the zeroed block or the mip tail, a
failed bind), the constructor went on to the next branch and flagged the
image UnwrittenSparse with nothing bound. AMD's drivers can't sample such
an image, and it is sampled before its first write replaces it, which is
the hang a855e8373f avoided for large sparse images.
Create the image again as an ordinary backed image in that case, as before
sparse tiles, and drop the Unbacked flag. The other modes keep falling
back to UnwrittenSparse, since their drivers sample unbound memory.
Follow-up to 130c255ed9 (with a855e8373f and 1c794af564).
Assisted-by: Claude Sonnet 5.5 (implementation and commit message)
|
| | | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | | |
Writing any register outside the instance loop's own methods now draws the
waiting instances first, so the batch sees the base instance it was begun
with. When a title writes another register in the middle of a loop, such
as a per-instance constant buffer, the loop is drawn in several batches,
and each later batch started again at global_base_instance_index with one
extra instance, so it used the instance IDs of the first records.
Count the instances already drawn for the current loop and start the next
batch after them, without the extra first-instance draw. A loop that is
not split is drawn exactly as before.
Follow-up to 2ad0830d2f and eb99da6072.
Assisted-by: Claude Sonnet 5.5 (implementation and commit message)
|
| | | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | | |
DisableGraphicsUniformBuffer tested !current.buffer_id to detect a
binding that is already NULL_BINDING. SlotId converts to true unless it
is the invalid id, and NULL_BUFFER_ID is BufferId{0}, so the check was
false for NULL_BINDING and the early return never triggered. Compare
against NULL_BUFFER_ID instead.
Follow-up to 46b2577bf7.
Assisted-by: Claude Sonnet 5.5 (implementation and commit message)
|
| | | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | | |
SetVertexRepackLayout returned old_mask | new_mask, which stays non-zero
on every draw that uses a 69-byte binding. PrepareDraw then re-dirtied the
vertex buffers and vertex input each time, so the repack compute pass ran
and the render pass ended once per draw, which is costly on tiled GPUs.
Return old_mask ^ new_mask so only a layout change re-dirties the
bindings. The repacked copy only stays valid while the source data is
unchanged, so the buffer cache now flags the vertex buffers overlapping a
CPU write, inline or DMA write, or GPU write, and re-dirties the repacked
ones before binding. Command buffer submission already invalidates all
vertex buffer state, which keeps the staging copy within its tick.
Also pass the source buffer size to the repack pass and clamp the
word-rounded storage descriptor range to it; it could run up to three
bytes past the end of the buffer. A binding that does not fit is skipped
like one over the device limits.
Follow-up to e38b738b84.
Assisted-by: Claude Sonnet 5.5 (implementation and commit message)
|
| | | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | | |
cce6d44af2 executes each command list entry from a copy of its words taken
at submit. Games submit command lists before writing them: the lists wait
behind a semaphore acquire, and the game writes their commands just before
it releases the acquire (Mario Kart World after present). The copies of
those lists were stale. Pokopia stopped after its first frame on macOS: its
GPU never reached the fence the game spun on (EDEN_NX2_PB_SNAPSHOT=0 made it
run again).
Count each channel's suspensions on an acquire and the resumes after them,
and record the count with each copy. A list copied while its channel was
suspended, or whose channel suspended since, reads guest memory instead.
Lists submitted and run without a wait in between keep the copy.
EDEN_NX2_PB_SNAPSHOT=0 still disables it altogether.
Assisted-by: Claude Sonnet 5.5 (first version); Claude Opus 5.5 (bisect, suspension count, testing and commit message)
|
| | | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | | |
0f14696044 made PresentClientBuffer also wait for the current frame's
fence whenever the previous frame's wasn't reached. Mario Kart World's
GPU waits on semaphores its CPU releases only after the present returns,
so in that state the current fence can't be reached until the present
returns, and each such present stalled up to the 4 s deadline.
Wait for the current frame by default, but give up after a few ms once a
GPU channel was seen suspended on a semaphore acquire and wait for the
previous frame instead. EDEN_NX2_FRAMES_AHEAD=1 always waits for the
previous frame (with 0f14696044's current-frame wait when the GPU is
behind), =0 always for the current one. The scheduler gains a lock-free
flag so the present thread can see suspended channels.
Assisted-by: Claude Sonnet 5.5 (implementation and commit message)
|
| | | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | | |
10e7063905 moved native program capture from one ReadBlock per instruction to
ReadBlockUnsafe per 4 KiB page. ReadBlockUnsafe skips the rasterizer flush, so
code that a GPU copy had written and not yet flushed to guest memory was
captured stale. The speed-up comes from reading a page at a time, not from
skipping the flush: read each page (and the odd single instruction outside a
fully mapped page) with ReadBlock, which flushes once per page instead of once
per instruction.
Follow-up to 10e7063905.
Assisted-by: Claude Sonnet 5.5 (implementation and commit message)
|
| | | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | | |
049c9633ed made TextureHeaderIndex take index / 32 for any texture index past the
pool that is a multiple of 32, in VisitImageView and in the shader environment.
That was written for Switch 2 handles that hold a header byte offset, but it ran
for Switch 1 games too, where such an out-of-range index used to give a null view
(and fail the shader environment's range check) and now binds an unrelated
header. Pass whether the memory manager uses native tiling, which is how Switch 2
titles are set up, and keep the previous behaviour otherwise.
Follow-up to 049c9633ed.
Assisted-by: Claude Sonnet 5.5 (implementation and commit message)
|
| | |/ / /
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
561f87abec only gave TextureDescriptor a handle_program. Image load/store/atomics
and texture/image buffers whose handle comes from a storage buffer still got the
placeholder constant buffer address (index 0, offset 0), so the pipelines read
cbuf 0 word 0 as the handle (and asserted if cbuf 0 was unbound). Add
handle_program to TextureBufferDescriptor, ImageBufferDescriptor and
ImageDescriptor, compare it when merging, and let all four pipelines evaluate it
for every descriptor kind instead of only TextureDescriptor.
ImageQueryDimensions on such a handle was always declared Color2D, so array,
cube and 3D views got a 2D declaration. The instruction carries no type, so take
it from another instruction using the same handle program, falling back to
Color2D when there is none.
Follow-up to 561f87abec. The OpenGL pipeline edits are the same change as the
Vulkan ones but were not compiled here.
Assisted-by: Claude Sonnet 5.5 (implementation and commit message)
|
| |\ \ \ \
| | | | |
| | | | | |
buffer_cache: allow 32 storage buffers per stage
|
| | | |/ /
| |/| |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
Native Switch 2 compute shaders (such as Donkey Kong Bananza) bind up to 18 global memory ranges, causing out-of-bounds binding errors:
HW.GPU <Error> buffer_cache.h:BindComputeStorageBuffer: Storage buffer index 16 exceeds maximum storage buffer count
HW.GPU <Error> buffer_cache.h:BindComputeStorageBuffer: Storage buffer index 17 exceeds maximum storage buffer count
followed by GPU driver access violations (0xc0000005).
The enabled and written storage buffer bitmasks (enabled_compute_storage_buffers and written_compute_storage_buffers) are already u32, so widening NUM_STORAGE_BUFFERS from 16 to 32 fits naturally within the existing 32-bit masks.
|
| |\ \ \ \
| |_|/ /
|/| | | |
nx2: decode sRGB in the upscale fallback when the program does (washed-out Splatoon Raiders)
|
| | | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
Splatoon Raiders' upscaler program ends by decoding its result from
sRGB to linear (HSETP2 against 0.04045, HMUL2 by 2.4 between MUFU.LG2
and MUFU.EX2) before its surface store. The bilinear copy that stands in
for it while HMMA isn't emulated wrote the sRGB input as-is, so the
later passes encoded it a second time: outdoor scenes were washed out,
with a pale sky and a white haze.
When the program carries those two immediates, the copy now decodes
like it, through a new copy operation of the blit helper (a fragment
shader with the piecewise sRGB decode). Programs without them, such as
Kirby Air Riders' upscaler, keep the plain copy. The OpenGL backend
ignores the operation.
|
| | | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
The trees fix (#52) drains the split and flushes for every command list
on a GPU-written page: in a Mario Kart World race ~13,500 lists and ~400
GPU round trips a second, during which the back has nothing to do. The
replay ran at ~40 fps instead of ~45.
The front now marks the pages those lists are on. The buffer cache notes
writes to them; after a compute dispatch or DMA copy that writes one it
copies the page into host memory and submits. The front waits for the
back only up to the last compute or copy work it sent (not everything
queued), then reads the copy once the GPU is done with it.
Pages the GPU wrote once, long before, and the CPU now writes lists to
have no copy: their first list drains and flushes the whole page, which
is then read from guest memory until GPU work writes it again. CPU
writes and inline uploads reach guest memory, so they keep such a page
synced but make a copy stale. A page a draw writes always drains.
Mario Kart World time-trial replay, Radeon 780M: ~40 -> ~45 fps, trees
drawn, no list drains left.
|
| |\ \ \ \
| | |/ /
| |/| /
| |_|/
|/| |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
Resolved by area:
- audio feature_support.h: nx2's native revision handling.
- fiber.cpp: ours, which also drops the unused rewind stack.
- host_memory.cpp: ours granule bitmap and HintCold, plus nx2's
fallocate hole punching per whole dirty granule on Linux.
- texture_cache.h: nx2's old-age eviction pass, then our Collect().
- kepler_compute.cpp: shared upscale slots, then Fortune's Weave's,
then the Kirby Air Riders scan.
- nx2_gsv.h: ours, plus nx2's GetBufferMaxResolution and graphics
device ids (1-based).
- nvhost_ctrl_gpu.cpp: engine support flag bits 40-44.
- lower_to_maxwell.h: ours as the base, with nx2's seven lowering
commits applied one at a time. Second texture descriptor buffers
are read bindless in TEX, TLD.SCR (level zero) and SUST; FSEL
keeps its operand modifiers; TXD takes any channel mask.
lower_to_maxwell.h lowers 837 Fortune's Weave and 1183 Mario Kart
World shader dumps with the same rejects as nx2 and ours (12 and 0).
Assisted-by: Claude Sonnet 5.5 (conflict analysis and drafting)
|
| | | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
buffer cache crash (#260)
* host_memory: punch holes in the backing on Linux instead of zero filling freed memory
Freeing guest memory zero filled the memfd backing with memset, which
commits every page it touches. fallocate(PUNCH_HOLE) hands the pages back
to the kernel instead, and they read as zero afterwards. Only the partial
pages at the ends are still cleared by hand.
Mario Kart World on Linux (RX 9070 XT, 31 GB): peak RSS 23.1 -> 15.9 GB.
EDEN_PUNCH_HOLE=0 restores the memset.
Assisted-by: Claude Opus 5.5 (debugging, profiling and code drafts)
* fiber: don't zero fill the stacks, untouched pages stay unbacked
The value initialised stack arrays were zero filled when each fiber was
made, which commits the whole stack. Left uninitialised, the pages a
fiber never touches stay unbacked. Together with the descriptor ring
change, about 1.2 GB less RSS in Mario Kart World.
Assisted-by: Claude Opus 5.5 (debugging, profiling and code drafts)
* vulkan: don't zero fill the descriptor update ring
Same as the fiber stacks: make_unique value initialised the whole ring
for every frame in flight. Entries are always written before they are
read, so allocate it without initialising it.
Assisted-by: Claude Opus 5.5 (debugging, profiling and code drafts)
* vulkan: wait for the GPU after an ASTC transcode decode like the other ASTC decodes
TranscodeAstcToBcn (astc_recompression Bc1/Bc3 with GPU decode) was the
only ASTC decode recorded with wait = false. On RADV (RX 9070 XT, Mesa)
Mario Kart World with BC3 flashed heavily all over the screen, while the
uncompressed GPU decode, which waits, did not. CPU decode with BC3 does
not flash either. Waiting after the transcode removes the flashing; full
pipeline barriers around its copy did not, so the wait hides a race
somewhere else rather than fixing one in the transcode itself.
Assisted-by: Claude Opus 5.5 (debugging, profiling and code drafts)
* vulkan: submit after an ASTC decode instead of waiting for the GPU by default
Every ASTC decode ended with scheduler.Finish(). Over a 19 minute Mario
Kart World Grand Prix that stalled the GPU thread 30.7 s in total, up to
1.2 s out of every 5 s while new textures stream in (track sections,
race starts), which is where the hitches were. Flushing keeps the
ordering that stops the BC3 flashing on RADV without the CPU waiting:
no flashing in play, and frames over 100 ms went from dozens to 3 in
about 7 minutes. EDEN_NX2_ASTC_FLUSH=0 restores the wait.
Assisted-by: Claude Opus 5.5 (debugging, profiling and code drafts)
* texture_cache: over the expected use, evict costly images unused for 1800 frames first
The collector never deletes CostlyLoad images outside aggressive mode, so
once a cup moves on, the previous tracks' textures stay resident and the
cache sits over its expected use deleting the cheap images still in use.
When over the expected use, costly images unused for 1800 frames (half a
minute) now go first.
Mario Kart World Grand Prix on an RX 9070 XT: frames over 20 ms 8.8% ->
4.0%, 58.7 fps average. EDEN_NX2_TC_OLD_AGE=<frames> sets the age, 0
turns it off.
Assisted-by: Claude Opus 5.5 (debugging, profiling and code drafts)
* buffer_cache: no buffer for device addresses past the page table
FindBuffer indexed page_table, which covers 2^36 bytes, with any address.
Mario Kart World crashed after about 4.5 minutes with a uniform buffer at
0x72253b2f2890 (FindBuffer <- UpdateUniformBuffers <- DrawEnd from a
macro), reading far past the table. Treat such an address as no buffer.
Where that address comes from is still unknown; this only stops the
crash.
Assisted-by: Claude Opus 5.5 (debugging, profiling and code drafts)
|
| | | |
| | |
| | |
| | | |
disable (#264)
|
