ayo
34e6b0fe71
Add an experimental FSR Ray Regeneration backend for D3D12 RR requests: - Convert depth, motion, normals and albedos into AMD denoiser inputs. - Split and recompose diffuse/specular signals, then run Super Sampling. - Add stability/detail controls, demodulation guards and halo limiting. - Handle AMD provider discovery under NVIDIA GPU spoofing. - Add runtime probes, synthetic reference scenes and per-dispatch profiling. Add exact gfx11 DLSS-RR inference using f16 cooperative matrices: - Reproduce FP8 conversions, accumulation and camera reprojection arithmetic. - Widen model weights and avoid requiring VK_EXT_shader_float8. - Extend model building, replay tools and descriptor metadata. Expand motion preprocessing and experimental AW2 native-motion support: - Add tiled LDS dilation, padded allocation handling and thin-only sky fill. - Order dilation with input waits and NGX on the selected stream. - Add jitter-aligned native composition, hybrid routing and wire history repair. - Add prepared masks, sparse repair and independent motion/repair controls. Add optional fused R11G11B10 input/output conversion, sampled stage timings and profile summarization. Extend configuration, packaging, documentation, FidelityFX headers and CPU/GPU regression coverage.
82 lines
3.7 KiB
Python
82 lines
3.7 KiB
Python
"""Execute the shim's motion routing with native readiness and NGX failures."""
|
|
from pathlib import Path
|
|
import subprocess
|
|
import tempfile
|
|
import unittest
|
|
|
|
ROOT = Path(__file__).resolve().parents[1]
|
|
|
|
|
|
class NativeMotionRoutingTests(unittest.TestCase):
|
|
def test_independent_repair_and_ngx_motion(self):
|
|
source = (ROOT / 'tools/d4r_nvngx_shim.cpp').read_text()
|
|
begin = source.index(' CudaObject nativeMotion = 0;')
|
|
end = source.index(' // Ray Reconstruction: the auxiliary', begin)
|
|
block = source[begin:end]
|
|
begin = source.index(' const bool wireHistoryRequested=nativeMotion')
|
|
end = source.index(';', begin) + 1
|
|
wire = source[begin:end]
|
|
program = r'''
|
|
#include <atomic>
|
|
#include <cassert>
|
|
#include <cstdint>
|
|
#include <cstring>
|
|
using CudaObject = uint64_t;
|
|
static unsigned substitute=1, repair=1;
|
|
static bool available=true, matched=true;
|
|
static CudaObject selected=0;
|
|
static unsigned env_uint(const char* name, unsigned) {
|
|
return std::strcmp(name,"D4R_NATIVE_MV_SUBSTITUTE")==0 ? substitute : repair;
|
|
}
|
|
static void logf(const char*, ...) {}
|
|
struct Feature {
|
|
unsigned flags=0x43, cudaFlags=0x43, outWidth=2560, outHeight=1440;
|
|
void* cudaHandle=(void*)1;
|
|
};
|
|
struct Slot { bool nativeLive=true; };
|
|
struct Params { unsigned renderWidth=853,renderHeight=480; float mvScaleX=853,mvScaleY=480; };
|
|
static CudaObject compose_native_live(Feature&, Slot&, const Params&, unsigned, void*) { return available ? 42 : 0; }
|
|
static bool set_native_motion_flags(Feature& f, Params&, bool native, unsigned) {
|
|
if(matched) f.cudaFlags=native ? f.flags & ~2u : f.flags;
|
|
return matched;
|
|
}
|
|
static CudaObject prepare_dilated_motion(Feature&, Slot&, const Params&) { return 7; }
|
|
struct { struct { bool stabilizeThinWires=true; } cu; } g;
|
|
static void run(bool nativeLive, bool hasField, unsigned useNative, unsigned useRepair,
|
|
bool match, unsigned initialFlags, bool expectedOk,
|
|
CudaObject expectedMotion, bool expectedRepair) {
|
|
Feature state; auto* feature=&state; Slot slot; Params params;
|
|
slot.nativeLive=nativeLive; available=hasField; substitute=useNative;repair=useRepair;matched=match;
|
|
state.cudaFlags=initialFlags;
|
|
bool ok=true; unsigned frame=1; void* stages=nullptr;
|
|
''' + block + wire + r'''
|
|
selected=liveMotion ? liveMotion : dilatedMotion;
|
|
assert(ok==expectedOk);
|
|
assert(selected==expectedMotion);
|
|
assert(wireHistoryRequested==expectedRepair);
|
|
if(liveMotion) {
|
|
assert(params.mvScaleX==2560 && params.mvScaleY==1440);
|
|
assert(state.cudaFlags==0x41);
|
|
} else {
|
|
assert(params.mvScaleX==853 && params.mvScaleY==480);
|
|
if(ok) assert(state.cudaFlags==0x43);
|
|
}
|
|
}
|
|
int main() {
|
|
run(true,true,1,1,true,0x43,true,42,true); // old combined path
|
|
run(true,true,1,0,true,0x43,true,42,false); // native only
|
|
run(true,true,0,1,true,0x43,true,7,true); // game MVs + repair
|
|
run(true,true,0,0,true,0x43,true,7,false); // game MVs, no repair
|
|
run(false,false,1,1,true,0x43,true,7,false); // clean baseline
|
|
run(true,false,0,1,true,0x43,true,7,false); // unavailable native inputs
|
|
run(true,true,0,1,true,0x41,true,7,true); // restore low-res NGX flags
|
|
run(true,true,0,1,false,0x41,false,0,true); // failed restore must not evaluate
|
|
run(true,true,1,1,false,0x43,true,7,false); // preserve old raw-output fallback
|
|
}
|
|
'''
|
|
with tempfile.TemporaryDirectory() as temp:
|
|
runner = Path(temp) / 'routing'
|
|
subprocess.run(['c++', '-std=c++17', '-Wall', '-Wextra', '-Werror', '-x', 'c++', '-',
|
|
'-o', str(runner)], input=program, text=True, check=True, capture_output=True)
|
|
subprocess.run([str(runner)], check=True, timeout=5)
|