Run the game on x86-64-v2 CPUs without AVX2

-march=x86-64-v3 kept the port and the installer to CPUs with AVX2 and FMA. --cpu v2 compiles the generated code and the runtime for SSE4.2 instead. Without FMA3 the FP lowering calls fma/fmaf, so runtime/fma.c provides correctly rounded ones (musl, MIT) that honour MXCSR's rounding mode, FTZ and DAZ; the UCRT's software fallback rounds incorrectly by up to half an ulp. simd.h's fast paths keep their assumptions.

The installer's host check asks for x86-64-v2 and selects libtotkrt.a or libtotkrt-v2.a; both release scripts build and ship the pair. The differential tests compile fma.c as well, so a --cflags=-march=x86-64-v2 run tests it.

Tested on an i7-3770K (no AVX2, FMA or BMI) with an RTX 3060 Ti, game 1.4.3: the installer ran end to end, --selftest passed and play holds 40-60 fps at 1080p. fma and fmaf matched 152,288 exact vectors over two seeds, including ties, subnormals and every rounding mode, FTZ and DAZ combination. The v3 build is unchanged: fma.c is an empty translation unit there.
This commit is contained in:
CHEN Feng committed 2026-11-01 07:09:13 +08:00
1 parent 7d62cbfa85
commit 34b4faee7c
19 files changed
+402 -58

No files matched your search

+10 -2
View File
@@ -1,7 +1,9 @@
# Building from source
Windows 10 or later, an x86-64 CPU with AVX2, and a GPU with Vulkan 1.3 and `VK_EXT_external_memory_host`. Allow about
60 GB for the work folder. Extraction wants about 55 GB free while it runs and keeps about 35 GB afterwards.
Windows 10 or later, an x86-64 CPU with SSE4.2, and a GPU with Vulkan 1.3 and `VK_EXT_external_memory_host`. The
default build targets x86-64-v3 (AVX2/FMA); a CPU without them builds with `--cpu v2` instead and runs slower (see
"Build"). Allow about 60 GB for the work folder. Extraction wants about 55 GB free while it runs and keeps about 35 GB
afterwards.
Players who only want to play can use the [installer](installer.md) instead: it builds the game on their machine
without this toolchain.
@@ -89,6 +91,9 @@ From the repository root, in PowerShell. Every step that depends on the game ver
.venv\Scripts\python tools\build_game.py --version <ver>
```
`--cpu v2` targets x86-64-v2 (SSE4.2, no AVX2/FMA) for older CPUs; the default `v3` needs AVX2. The v2 build links
`runtime/fma.c`, a correctly rounded software fma the FP lowering needs, and runs slower.
Lifting takes about 15 s, a clean compile about 200 s on 24 threads; later builds recompile only what changed.
A compiler cache in front of clang shortens it further: with `SCCACHE` set to an sccache executable, `build_game.py`
writes it into the ninja file ahead of `cc`, and every lift rewrites the generated chunks with the same contents, so
@@ -195,6 +200,9 @@ change passes these, all run with `.venv`:
| `tests\excl_mt.py` | the exclusive monitor helpers under contention and ABA, on several threads | no |
| `tests\jumptables_check.py` | `recomp/jumptables.py` against a Unicorn oracle; must report 0 disagreements, on 1.4.3 too | yes |
A `--cpu v2` build passes the same differential tests with `--cflags=-march=x86-64-v2`; they compile `runtime/fma.c`
too, so they check the software fma the v2 build links rather than the C runtime's.
Runtime probes, each a mode of `totk.exe` (the list is at the top of `runtime/main.c`): `--probe-kernel`,
`--probe-vkexec`, `--probe-nvn-tex` and `--probe-shader` cover most runtime changes. `--probe-shader` and the draw
part of `--probe-vkexec` translate shaders of the census below: they look for `build/<ver>/shaders/jobs.json` of the
+8 -6
View File
@@ -41,14 +41,15 @@ keeps `profile/`. Saves carry over both ways: 1.4.3 converts a 1.2.1 save on loa
1. **Check** before writing anything: the files open with the keys, the update is 1.2.1 or 1.4.3 (by hash, as
`tools/extract.py`), the keys are new enough for the update, enough space is free on the target disk (about 36 GiB at the peak of a
first install, 20 GiB after it), the CPU has
AVX2 (`-march=x86-64-v3`), a Vulkan 1.3 GPU exists. Each failure is one clear message.
first install, 20 GiB after it), the CPU is at least x86-64-v2 (SSE4.2; with AVX2/FMA the faster `v3` build is used,
otherwise the `v2` one), a Vulkan 1.3 GPU exists. Each failure is one clear message.
2. **Extract** ExeFS and RomFS into the install folder and verify the hashes.
3. **Lift** the ExeFS to C (`recomp/gen.py`, `recomp/patches-v121.txt` or `patches-v143.txt`) into a temp folder next
to the installer.
4. **Compile** the generated C, `modules.c`, `romfs.c`, `patches.c` and `ov_zstd.c` (they bake the per-install layout, RomFS
hashes and the game version's patch page) and link
them with the prebuilt runtime library into `totk.exe` (`build_game.py --runtime-lib`; about 3 min on 24 cores).
them with the prebuilt runtime library for the host's level (`libtotkrt.a` or `libtotkrt-v2.a`) into `totk.exe`
(`build_game.py --runtime-lib`; about 3 min on 24 cores).
5. **Prepare** the RomFS patches and the shaders (`tools/make_romfs_patches.py`, `tools/prepare_shaders.py`,
`totk.exe --prepare-shaders`) in the install folder, then fold the shader cache's loose files (four per shader, about
117,000) into `shader-cache/shader-cache.pack` (`tools/pack_shader_cache.py`). totk.exe reads the pack first and
@@ -97,7 +98,8 @@ installer/ the installer: steps, CLI, window, en/ru text catalog
python/ Python 3.14.9 embeddable + numpy, lz4, cryptography, cffi, pycparser (+ tkinter for the window)
recomp/, tools/ the lifter and the build, extraction and shader scripts the steps run (copied from the repo)
llvm/ llvm-mingw, x86_64-only subset (Apache-2.0 with LLVM exception; mingw-w64 parts permissive)
runtime/ libtotkrt.a and the headers the generated code and modules.c/romfs.c/patches.c include
runtime/ libtotkrt.a, libtotkrt-v2.a and the headers the generated code and modules.c/romfs.c/patches.c
include
bin/ totk_xlate.exe, hactool.exe, SDL3.dll, libopus-0.dll, ninja.exe
data/ shader_list.tsv (MacNativeGames, MIT)
THIRD_PARTY_NOTICES.md
@@ -124,8 +126,8 @@ python/ python-build-standalone 3.14 (with tkinter) + the installe
tools/ and recomp/ on sys.path
llvm/ LLVM 23 (clang, lld, llvm-ar) and llvm/sysroot: the files of the Steam Runtime 3 "sniper" SDK the
game's compile and link read (about 130 files)
runtime/ libtotkrt.a, headers, the install-side sources, totk.png; deps/ with the Vulkan headers, VMA and a
static zstd
runtime/ libtotkrt.a, libtotkrt-v2.a, headers, the install-side sources, totk.png; deps/ with the Vulkan
headers, VMA and a static zstd
bin/ totk_xlate, hactool, ninja, libSDL3.so.0
```
+2 -1
View File
@@ -6,7 +6,8 @@ from it may be shared.
## What you need
- Windows 10 or later, an x86-64 CPU with AVX2, about 60 GB free.
- Windows 10 or later, an x86-64 CPU with SSE4.2 (AVX2/FMA recommended: they build the default, faster version), about
60 GB free.
- A GPU with Vulkan 1.3 and `VK_EXT_external_memory_host`, with a current driver.
- *Tears of the Kingdom* base game (1.0.0) and the 1.2.1 or 1.4.3 update, dumped from your own Switch as NSP or XCI.
- `prod.keys` dumped from your own Switch. Never send it to anyone, and never paste it into a log or an issue.