network natively on HIP with Vulkan->HIP shared memory, extend K kernels to
gfx11 (RDNA3), and add dependency/packaging/test tooling around them.
Windows: FSR RR backend, native HIP K runner, gfx11 K kernels, tooling
FSR Ray Regeneration (experimental)
- Add fsr_rr_backend.{h,cpp}, fsr_rr_passes.h and the fsr_rr/ prepare and
compose compute shaders. They run AMD's FidelityFX denoiser
(amd_fidelityfx_denoiser_dx12.dll, SDK 2.3, not bundled) on a private
vkd3d-proton device.
- Add the amdxc64 shim (tools/windows/amdxc64_shim) and two vkd3d-proton
patches for libcxx includes. AMD's amdxc64.dll is never loaded.
- d4r_nvngx_windows: RrFeature now selects between the native DLSS RR
backend and the FSR backend (D4R_RR_BACKEND=fsr). After the denoiser it
runs an internal Super Sampling stage (feature 1) to upscale. If SR is
unavailable and output != render size, the frame fails; it is never
returned noisy.
- d4r.ini / player_config: new [RayReconstruction] FsrDenoiser,
FsrStabilityBias (auto or >= 0) and FsrDetail (0-1), with validation, and
matching D4R_FSR_* environment variables.
- Add the FSR RR fixture, output-check and ngx test scripts, and a shader
compile script.
Native HIP K runner (Windows)
- CMake: add d4r_engine_k_hip (d4r-k-hip) plus separate volk/engine copies,
because VK_USE_PLATFORM_WIN32_KHR changes volk's entry points.
- Share Vulkan memory with HIP via Win32 opaque handles on Windows
(VK_KHR_external_memory_win32) and via fds on Linux. Handle types are now
intptr_t, and the platform choice is in D4R_EXT_MEMORY_* macros.
- HipNetwork: add debugRepeatHashes and debugNodeMicros, exposed as
D4R_HIP_REPEAT_CHECK, D4R_HIP_REPEAT_ALL and D4R_HIP_NODE_PROFILE in
replay.cpp.
- replay.cpp: display-resolution-motion frames (FrameFlagDisplayMotion) no
longer require near2rnd == out2rnd. Both values must still be in (0, 1].
K kernels on gfx11
- Allow position-token and enc1 tiling on __GFX11__ as well as gfx12.
- Add an opt-in ds_bpermute cross-half exchange
(D4R_WINDOWS_GFX11_DS_EXCHANGE, enabled for K on Windows gfx11). M keeps
its LDS path.
- Fix a block-barrier race in the position layer. Without POS_VSHARE a fast
wave could overwrite its hb rows while a slower wave still read them for a
later head, giving rare wrong results on gfx1101 (enc0 outputs differing
between identical launches).
Packaging and install
- install.bat, d4r.ini, WINDOWS_README: FSR option, and
install_dependencies.py to install pinned NVIDIA dependencies with staged,
hash-verified replacement and rollback.
- Update the release package, local package refresh and test scripts.
- Add the InteropBatchInputCopies option (default false). Batching is faster
on some synthetic runs, but player temporal exactness is unresolved.
Tooling and tests
- K model: capture_k_weights, build_k_model, compare_k_outputs,
benchmark_k_hip, build_capture_runtime, test_k_model and the
zluda-k-output-capture patch. Captured weights and models are private
local assets and must not enter release archives.
- Synthetic probes: test_m_ngx, test_sr_dynamic,
prepare_rr_dynamic_fixture and the d3d12/rr_ngx probe updates.
- Add tests/windows/test_install_dependencies.py.
Docs
- Update the windows-fsr-rr, k-performance, local-package, rr-interop and
rr-model-cache docs. Add windows-parity-handoff.md and
windows-parity-continuation-prompt.md.
Implement native NGX feature-13 RR through a LUID-matched D3D12/Vulkan
shared-memory and timeline-fence bridge. Add owned parameter snapshots,
GPU resource lifetime handling, inherited/enhanced barriers, optional
guides, sampled input formats and depth-plane support with stencil
preservation. Separate RR initialization from the SR CUDA runtime and
expose explicit backend/model selection through player configuration.
Add synthetic bridge, snapshot and NGX lifecycle harnesses covering
camera/history/reset, parameter mutation, delayed submission, early
release, capability gating and unsupported-input rejection.
Add exact gfx11 D/E model generation, verified caches, versioned local
deployment with configuration backups, and isolated portable model
tools with pinned Python, NumPy and glslang dependencies.
Extend Windows K benchmarking and private native-cache selection.
Fix LLVM 22 emitter/helper compatibility and preserve exact-output
validation for local output-kernel optimization.
Add native FSR RR provider diagnostics, local development package
refresh tooling, installer/checker updates, FidelityFX headers and
Windows parity documentation.
Keep NVIDIA binaries, derived models and private optimized kernels out
of distributable packages. Native RR evidence remains synthetic;
game validation, FSR RR dispatch, installer integration and Linux
performance parity remain incomplete.
Add native K/M/RR inference engines, HIP kernels, replay and validation
harnesses, and Windows DLSS setup and runtime support. Include GPU
profiling tools, build and packaging scripts, platform integration
patches, and implementation documentation.
- Set Disable.Watermark on the game path (the Linux shim already does);
[DLSS] ShowWatermark = true keeps NVIDIA's stamp.
- Always write d4r\d4r_nvngx.log in player mode. It was skipped when the
launcher gave the game a valid stderr pipe, e.g. Steam.
- Accept NVIDIA's 310.7.0 + _nvngx 32.0.15.9636 as "less tested" pins.
Native kernels are still gated by PTX identity, so a different DLL
falls back to translated kernels.
- Add d4r-check.ps1 (ok/MISSING/note lines with a fix for each; -Report
writes a bug-report ZIP) with 15 fixture tests.
- package-windows-release.ps1: -Architectures subset, ships the check
script, GPU registry and one GPU-query tool.
- Add a shorter WINDOWS_README.txt; trim README's Windows section and
merge in the RX 9060 XT / distributor-DLL notes.
- Diagnostic probe accepts preset 12 (L).
Tested on RX 7700 XT without a game: loader and shim start-up, K and M
synthetic NGX frames with both DLL pairs (finite output, native kernels
used). Not tested in a game.
Players extract one ZIP into the game folder, copy their two NVIDIA
DLLs, and start the game from Steam, Epic or a shortcut. No launcher
script is needed; settings live in d4r\d4r.ini.
- Shim player mode: when the launcher has not set D4R_HIP_ROOT and
d4r.ini sits beside the shim, derive the backend environment from it.
Codegen flags stay at the Windows-validated values. NVIDIA DLLs must
match an exact SHA256 in dll-pins.txt, with "you have X, you need Y"
errors. Rejects a vkd3d/Wine d3d12.dll and code objects built for
another GPU. [DLSS] Model accepts K or M and is applied by rewriting
the render preset hints.
- d4r.ini exposes only options that exist on Windows: Model, AsyncInterop,
ValidateOutput, Log, CacheDir. Invalid values fall back with a note in
d4r_nvngx.log.
- d4r_nvngx_loader: one _nvngx.dll for all GPUs. It matches the game's
adapter by LUID and loads _nvngx_<arch>.dll from the same folder.
- OptiScaler patch 0006: relative NvngxPath/NvngxDlssPath/
NvngxFeaturePath resolve against OptiScaler's folder.
- package-windows-release.ps1 builds the release from the per-target
game packages and refuses mixed targets or differing shared runtimes.
- Tests: test-player-config.ps1 (ini and DLL-pin decisions, loader
selection on the local GPU) and test-windows-release-package.ps1.
Adapt official LLVM 7a0829e41228 to the pinned LLVM 22 parser and ELF structures. Apply and audit the complete source delta before building, export its provenance, and package the source patch. Add 33 offline public PTX compilation checks across all eleven GPU targets without hardware execution. Preserve strict arithmetic and document RX9070XT K/M regressions.
Share explicit target metadata across build/runtime/package guards, preserve native gfx11 WMMA with a Windows LDS lane exchange, select HIP devices by D3D12 LUID, and reject wrong-target/vkd3d installations before changing game files. Review all issue #10 logs and document compile-only coverage plus RX9070XT regressions.
Co-authored-by: realdody <[email protected]>
Add explicit gfx1200/gfx1201 builds, runtime selection and per-target
code objects. Guard staging, packaging and installation against mixed
architectures; retain gfx1201 compatibility and strict K/M arithmetic.
Adapt realdody's target configuration and MSVC binary16/BOOL portability
fixes from ae530b0. Do not import its forced gfx12 layout on RDNA3.
Add exhaustive binary16 and C++/PowerShell target contract regressions.
Validation: 20 CTest gates on RX 9070 XT; four software-only gfx1200 gates;
MSVC portability compile checks; exact 4K K async and M control outputs;
isolated package install/restore and mixed-target rejection tests.
No physical gfx1200 validation is claimed.
Co-authored-by: realdody <[email protected]>
- Add default-off PreferAccuracy for all native kernel families,
including RDNA4 FP8 variants.
- Restore intermediate FP8 quantization, per-MMA FP16 rounding,
denormal handling and NGX synchronization in accuracy mode.
- Select separately marked accuracy kernels and use a separate JIT
cache; reject mixed sets and never silently substitute fast kernels.
- Add a pinned GLIBC 2.41 build environment for ZLUDA/LLVM, the DLLs
and both kernel sets, with ABI checks and toolchain provenance.
- Support parallel target builds and pre-extracted PTX; correct
ROCm header paths and Clang-CL input handling.
- Bump to 0.1.3 and update release notes, documentation and tests.
Validation: 11 tests passed; all 912 kernels built; E/K/M GPU harness
runs completed with accuracy off/on. Packaged libraries loaded under
GLIBC 2.41 and require no GLIBC version newer than 2.38.
RTX image-quality parity remains unverified.
Fixes#6
Refs #6
- Fix Cyberpunk 2077 preset-K black artifacts by preserving FP16/FP64
denormal requirements while limiting denormal suppression to FP32.
- Fix stale direct-output redirects after array destruction, preventing
reused handles from inheriting an earlier feature's output buffer.
- Log rejected resource formats and dimensions, including failures after
startup, with rate limiting for repeated errors.
- Add opt-in blocking output synchronization with a dedicated bridge
export and fallback to context synchronization.
- Report output-wait CPU and wall time, and defer GPU timing queries
until asynchronous evaluation has completed. The CPU-overhead
mitigation for #7 still needs reporter confirmation.
- Reject release libraries requiring GLIBC newer than the configured
baseline, defaulting to 2.41, and document compatible builds.
Validation: shim and bridge builds pass; targeted redirect-lifetime and
format-logging checks pass.
Fixes#5Fixes#8
Refs #7
ayo