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.
96 lines
5.3 KiB
Python
96 lines
5.3 KiB
Python
"""Build an isolated diagnostic ZLUDA runtime using the already-built LLVM tree.
|
|
|
|
Never replaces the delivered runtime or rebuilds LLVM. Output is local under
|
|
build/. The copied runtime is for audited synthetic captures, not game use.
|
|
"""
|
|
import argparse
|
|
import hashlib
|
|
import json
|
|
import os
|
|
from pathlib import Path
|
|
import shutil
|
|
import subprocess
|
|
|
|
|
|
def sha(path):
|
|
with path.open('rb') as stream:
|
|
return hashlib.file_digest(stream, 'sha256').hexdigest()
|
|
|
|
|
|
def main():
|
|
repo = Path(__file__).resolve().parents[2]
|
|
parser = argparse.ArgumentParser(description=__doc__)
|
|
parser.add_argument('--runtime', required=True, type=lambda p: Path(p).resolve())
|
|
parser.add_argument('--output', required=True, type=lambda p: Path(p).resolve())
|
|
parser.add_argument('--jobs', type=int, default=4)
|
|
args = parser.parse_args()
|
|
if args.output.exists() or not args.output.is_relative_to((repo / 'build').resolve()):
|
|
parser.error('use a fresh directory under ignored build/')
|
|
if args.output.is_relative_to(args.runtime) or not 1 <= args.jobs <= 8:
|
|
parser.error('output must be outside input runtime; jobs must be 1..8')
|
|
source = repo / 'external/ZLUDA'
|
|
if '"hiluma_engine_output_depthreg_mvhi_ldr_max_v2_rel" => Some(328)' not in (source / 'zluda/src/impl/d4r_replay.rs').read_text():
|
|
parser.error('apply tools/windows/zluda-k-output-capture.patch to the prepared Windows ZLUDA source first')
|
|
llvm = repo / 'build/zluda-llvm-windows-gnu'
|
|
rust = repo / '.tools/rust-1.98.1-gnu'
|
|
compiler = repo / '.tools/llvm-mingw-20260922-ucrt-x86_64'
|
|
target = repo / 'build/zluda-rust-windows-gnu'
|
|
for path in (llvm / 'bin/llvm-config.exe', rust / 'bin/cargo.exe', source / 'ptx/lib/zluda_ptx_impl.bc'):
|
|
if not path.is_file():
|
|
parser.error('prebuilt dependency missing: ' + str(path))
|
|
args.output.mkdir(parents=True)
|
|
artifact = target / 'x86_64-pc-windows-gnu/release/nvcuda.dll'
|
|
if artifact.exists():
|
|
shutil.copy2(artifact, args.output / 'previous-cargo-nvcuda.dll')
|
|
environment = os.environ.copy()
|
|
environment.update(CARGO_HOME=str(repo / '.tools/cargo-home'), CARGO_TARGET_DIR=str(target),
|
|
CARGO_BUILD_JOBS=str(args.jobs), RUSTC=str(rust / 'bin/rustc.exe'),
|
|
CC=str(compiler / 'bin/x86_64-w64-mingw32-clang.exe'),
|
|
CXX=str(compiler / 'bin/x86_64-w64-mingw32-clang++.exe'),
|
|
AR=str(compiler / 'bin/llvm-ar.exe'), CXXSTDLIB='c++',
|
|
LIBRARY_PATH=str(rust / 'lib/rustlib/x86_64-pc-windows-gnu/lib/self-contained'),
|
|
CARGO_TARGET_X86_64_PC_WINDOWS_GNU_LINKER=str(compiler / 'bin/x86_64-w64-mingw32-clang.exe'),
|
|
CARGO_TARGET_X86_64_PC_WINDOWS_GNU_RUSTFLAGS='-C link-arg=-static',
|
|
ZLUDA_LLVM_BUILD_DIR=str(llvm))
|
|
environment['PATH'] = os.pathsep.join(str(p) for p in (rust / 'bin', compiler / 'bin',
|
|
repo / '.tools/python/cmake/data/bin', repo / '.tools/python/bin', Path('C:/Program Files/Git/usr/bin'))) + os.pathsep + environment['PATH']
|
|
report = dict(passed=False, diagnosticOnly=True, llvmRebuilt=False, deliveredRuntimeChanged=False,
|
|
originalRuntimeSha256=sha(args.runtime / 'zluda/nvcuda.dll'),
|
|
captureSourceSha256=sha(source / 'zluda/src/impl/d4r_replay.rs'),
|
|
llvmConfigSha256=sha(llvm / 'bin/llvm-config.exe'))
|
|
try:
|
|
command = [str(rust / 'bin/cargo.exe'), 'build', '--offline', '--locked', '--release',
|
|
'--target', 'x86_64-pc-windows-gnu', '-p', 'zluda']
|
|
with (args.output / 'cargo.log').open('wb') as log:
|
|
process = subprocess.Popen(command, cwd=source, env=environment, stdout=log, stderr=subprocess.STDOUT)
|
|
try:
|
|
code = process.wait(timeout=1800)
|
|
except subprocess.TimeoutExpired:
|
|
subprocess.run(['taskkill', '/PID', str(process.pid), '/T', '/F'], check=False, timeout=30,
|
|
stdout=subprocess.DEVNULL, stderr=subprocess.DEVNULL)
|
|
process.wait(timeout=30)
|
|
raise RuntimeError('Capture runtime build timed out; process tree terminated')
|
|
if code:
|
|
raise subprocess.CalledProcessError(code, command)
|
|
runtime = args.output / 'runtime'
|
|
shutil.copytree(args.runtime, runtime)
|
|
shutil.copy2(artifact, runtime / 'zluda/nvcuda.dll')
|
|
for name in ('libc++.dll', 'libunwind.dll'):
|
|
shutil.copy2(compiler / 'bin' / name, runtime / 'zluda' / name)
|
|
if sha(args.runtime / 'zluda/nvcuda.dll') != report['originalRuntimeSha256']:
|
|
raise RuntimeError('Original runtime changed during build')
|
|
report.update(passed=True, runtime=str(runtime), runtimeSha256=sha(runtime / 'zluda/nvcuda.dll'),
|
|
sourceBase=subprocess.check_output(['git', 'rev-parse', 'HEAD'], cwd=source, text=True).strip(),
|
|
sourceWorkingTree='local modified; capture source hash recorded, no clean source assertion',
|
|
gpuValidation=False, published=False)
|
|
print(json.dumps(report, indent=2))
|
|
except Exception as error:
|
|
report['error'] = str(error)
|
|
raise
|
|
finally:
|
|
(args.output / 'build.json').write_text(json.dumps(report, indent=2) + '\n')
|
|
|
|
|
|
if __name__ == '__main__':
|
|
main()
|