Triforce v0.7.0 "Sage Of Courage"

The 0.6 crashes after warps and area changes fixed (no write-back into memory the capture leaves out), invisible objects fixed (occlusion gate), DLSS frame generation flicker fixed, achievements, the renderer built with a play profile, faster translated game code, helper threads judged on settled scenes, memory, shader preparation, audio effects, launcher, controls and Linux fixes.
This commit is contained in:
ruckus7382sfgsgdf
2026-11-22 11:59:06 -06:00
parent 70c210cee1
commit e6f2f248f2
529 changed files with 55958 additions and 1898 deletions

View File

@@ -2,6 +2,8 @@
Upscaler, and the names it shows for DLSS's (and FSR 3.1's and FSR 4's) modes. No Vulkan, no menu: tests/test_upscaler_rows.cpp checks them. */
#pragma once
#include <cstdint>
namespace mkwvk::upscaler_rows {
/// The Upscaler row (opts.dlss, RX_VK_DLSS): the game's own network, NVIDIA DLSS, NVIDIA DLAA, AMD FSR 3.1 (built from AMD's
@@ -189,7 +191,18 @@ inline int window_filter(unsigned picture_w, unsigned picture_h, unsigned dest_w
/// generation transformer, preset M (NVIDIA's own pick for Performance); 2 the same generation as tuned for Ultra
/// Performance, preset L; 3 NVIDIA's default, preset 0 (the runtime picks per mode, and a runtime update can change it)
enum { MODEL_K = 0, MODEL_M = 1, MODEL_L = 2, MODEL_DEFAULT = 3, kModels = 4 };
inline unsigned model_preset(int model) {
/// (Triforce, 11-22, the performance audit's PRESENT-36, after Source A's and Source B's) The K row on a card with less
/// than 7 GB of video memory asks for preset E, NVIDIA's last CNN model: less GPU time and less of NGX's own memory than
/// the transformer there (Source A's, a 6 GB laptop RTX 3050 at 1080p DLSS Performance: 15.75 to 14.20 ms of GPU a frame,
/// 85 to 58 MB), with more ghosting on grass and thin objects at the smallest render sizes, and NVIDIA marks it deprecated.
/// vram: the card's video memory in bytes; 0 when the rule is off (vk_dlss.cpp dlss_cnn_vram: RX_VK_DLSS_CNN_VRAM is not
/// 1, the default) or the size is not known: K as chosen. M, L and NVIDIA's default stay as chosen, and RX_VK_DLSS_PRESET
/// goes over all of it
constexpr unsigned kPresetE = 5;
constexpr uint64_t kCnnVramBytes = 7ull << 30;
inline bool cnn_for_vram(int model, uint64_t vram) { return model == MODEL_K && vram && vram < kCnnVramBytes; }
inline unsigned model_preset(int model, uint64_t vram = 0) {
if (cnn_for_vram(model, vram)) return kPresetE;
switch (model) {
case MODEL_M: return 13;
case MODEL_L: return 12;
@@ -202,8 +215,8 @@ inline unsigned model_preset(int model) {
inline bool preset_env(const char* env) {
return env && ((*env >= 'A' && *env <= 'Z') || (*env >= 'a' && *env <= 'z') || (env[0] == '0' && !env[1]));
}
inline unsigned dlss_preset(int model, const char* env) {
if (!preset_env(env)) return model_preset(model);
inline unsigned dlss_preset(int model, const char* env, uint64_t vram = 0) {
if (!preset_env(env)) return model_preset(model, vram);
if (*env == '0') return 0;
return (unsigned)(*env >= 'a' ? *env - 'a' + 1 : *env - 'A' + 1);
}