Add AMD FSR 4 upscaling through the libraries of AMD's FSR SDK 2.4, as an Upscaling choice beside FSR 3.1.

runtime/vk_ffx4.cpp loads amd_fidelityfx_loader_vk.dll from beside the executable and asks it which upscaler
versions the device can run. Where a 4.x version is listed (Radeon RX 6000, RX 7000 and RX 9000 cards with AMD's
driver 26.10.1 or newer) the menu's Upscaling row gains AMD FSR 4 Native AA, Quality, Balanced, Performance and
Ultra Performance, and fsr4 = 1 runs it in FSR 3.1's place with the same inputs. Its contexts are made with that
version asked for by its id and the provider is asked afterwards which version it runs, so no other upscaler can
be shown under the name. When a context or a dispatch fails, FSR 3.1 makes the frame, and after three failures
FSR 4 is off for the run. Each upscaler starts its history anew on the frame after a change from one to the other.
The setting is off by default and the Linux build compiles stubs.

On an RX 7900 XT (driver 26.10.1, 3440x1440, HDR) the library runs FSR 4.1.2: a context takes 54 ms to make and
217 MiB of video memory. Standing in Hyrule Field: Quality 47.8 fps where FSR 3.1 gives 51.5, Native AA 41.4.
Screenshots of 601 consecutive frames standing, 801 during a camera sweep, 701 each in Kakariko and the
graveyard, and 1,301 across changes between FSR 3.1 and FSR 4, their modes and off show no flicker that FSR 3.1
does not have, and it runs together with FSR 3.1 frame generation.

The SDK's API headers are in runtime/third_party/ffx_sdk, unmodified, with the licence. tools/make_kit.py and
tools/make_launcher.py take AMD's two signed libraries with --ffx4, and tools/make_game.py copies them beside the
game when the kit has both.

Co-Authored-By: Claude Fable 5.2 <noreply@anthropic.com>
This commit is contained in:
epicmeme546
2026-11-11 03:21:51 -05:00
parent 3fe6965df9
commit 4d5454d380
19 changed files with 1444 additions and 25 deletions

View File

@@ -1,6 +1,7 @@
"""Build a release of the NX2navi launcher (docs/launcher.md, "Releases").
python tools/make_launcher.py --xlate build/xlate [--ffx amd_fidelityfx_vk.dll] [--dlss nvngx_dlss.dll | --no-dlss]
[--ffx4 FOLDER]
[--prompt-art FOLDER [FOLDER ...]] [--out build/launcher-release]
[--feed owner/name]
[--notes notes.md] [--publish]
@@ -17,7 +18,8 @@ The release folder (--out) holds:
python/ python.org's embeddable Python 3.14.9, tkinter and Tcl/Tk 9.0 from its tcltk.msi,
the packages of tools/requirements-launcher.txt, runtime-id.txt
app/current.txt, app/<version>/ launcher/ (without native/), runtime/, recomp/, the tools make_game.py uses,
bin/nx2navi_xlate.exe and .id (and amd_fidelityfx_vk.dll with --ffx),
bin/nx2navi_xlate.exe and .id (and amd_fidelityfx_vk.dll with --ffx, AMD's two
FSR 4 libraries with --ffx4),
bin/nvngx_dlss.dll and its notice (NVIDIA DLSS at the version tools/dlss_pin.py
pins, downloaded from NVIDIA's DLSS SDK repository and checked, or --dlss; not with
--no-dlss),
@@ -177,7 +179,7 @@ def build_python(py):
# ---------------------------------------------------------------- app/<version>/
def build_app(app, ver, xlate, feed_repo, lock, ffx=None, dlss=None, art=None, dlssg=None):
def build_app(app, ver, xlate, feed_repo, lock, ffx=None, dlss=None, art=None, dlssg=None, ffx4=None):
# The runtime's own shaders are source in the repository and SPIR-V headers in a release: a player's PC compiles
# runtime/ (make_game.py) and has no shader compiler. Build the missing and the stale ones before the copy.
build_shaders.build(ROOT / 'runtime')
@@ -198,6 +200,9 @@ def build_app(app, ver, xlate, feed_repo, lock, ffx=None, dlss=None, art=None, d
shutil.copy2(xlate / n, app / 'bin' / n)
if ffx: # FSR 3.1 loads it from beside the game (runtime/vk_ffx.cpp); make_game.py copies it there
shutil.copy2(ffx, app / 'bin' / 'amd_fidelityfx_vk.dll')
if ffx4: # FSR 4 loads them from beside the game (runtime/vk_ffx4.cpp); make_game.py copies them there
for n in ('amd_fidelityfx_loader_vk.dll', 'amd_fidelityfx_upscaler_vk.dll'):
shutil.copy2(ffx4 / n, app / 'bin' / n)
if dlss: # DLSS: NGX loads it from beside the game (runtime/vk_dlaa.c); make_game.py copies it there
dlss_pin.place(dlss, app / 'bin')
if dlssg: # DLSS Frame Generation (runtime/vk_dlssg.c): make_game.py place_dlssg copies it beside the game
@@ -314,6 +319,8 @@ def main():
ap = argparse.ArgumentParser(description=__doc__.splitlines()[0])
ap.add_argument('--xlate', type=Path, required=True, help='folder with nx2navi_xlate.exe and nx2navi_xlate.id')
ap.add_argument('--ffx', type=Path, help="AMD's signed amd_fidelityfx_vk.dll (FidelityFX SDK 1.1.4, MIT) for FSR 3.1")
ap.add_argument('--ffx4', type=Path, help="the folder with AMD's signed amd_fidelityfx_loader_vk.dll and "
'amd_fidelityfx_upscaler_vk.dll (FSR SDK 2.4.0, MIT) for FSR 4')
ap.add_argument('--dlss', type=Path, help=f"NVIDIA's nvngx_dlss.dll (DLSS {dlss_pin.VERSION}) from disk instead of "
"downloading it from NVIDIA's DLSS SDK repository")
ap.add_argument('--no-dlss', action='store_true', help='leave NVIDIA DLSS out of the release')
@@ -356,7 +363,7 @@ def main():
dlssg = (a.dlssg, streamline_pin.resolve(a.streamline, DL)) if a.dlssg else None
say(f'DLSS Frame Generation: {a.dlssg}, Streamline {streamline_pin.VERSION}: {dlssg[1]}' if dlssg
else 'DLSS Frame Generation left out (no --dlssg)')
build_app(folder / 'app' / ver, ver, a.xlate, a.feed, lock, a.ffx, dlss, a.prompt_art, dlssg)
build_app(folder / 'app' / ver, ver, a.xlate, a.feed, lock, a.ffx, dlss, a.prompt_art, dlssg, a.ffx4)
(folder / 'app' / 'current.txt').write_text(ver + '\n', encoding='utf-8')
build_python(folder / 'python')
cc, windres = native_compiler(lock)