Ported onto vk_dlaa_plan (#794): the scene end in the inventory also records its submission and op (end_fence,
end_pos) so chunked replays see it at the same point. DLSS-only: everything sits behind dl_active().
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
entry_of decides at record time whether a new texture takes the content of a live GPU-written texture at its storage
(alias_copy) or the storage bytes. In a chunked replay the source's GPU ownership is set when a chunk records its
render, so the choice followed the order the threads got to the chunks: a later chunk could take the old bytes before
an earlier chunk recorded the render, and a mirrored texture keeps them. Now a chunk that makes an entry while a live
colour texture of its size sits at the storage waits until the chunks before it are recorded (vk_exec_chunk_settle,
the texture lock released for the wait), skips sources the scan revives at a later op, and ignores GPU ownership a
later chunk set in this submission. slot_resolve and vk_texture_by_id_locked no longer read tslots through a pointer
taken before the wait.
TOTK_REPLAY_SKEW_US (debug) makes the even chunks wait before they record, so the chunk after each one records
first in every submission. With it at 3000 us and 4 threads the field save runs clean (grass drawn, 0 validation errors); the alias path
did not fire there (0 copied, 0 filled from storage), so this does not reproduce the missing grass reported on
RADV, and the fix there stays unconfirmed.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
DLSS kept the replay on one thread: its decisions follow the passes in replay order, its jitter and camera are noted
draw by draw, and its evaluation reads the per-object motion list the scene's draws build. Now the scan takes DLSS's
decisions on its own state (vk_dlaa_plan, split out of vk_dlaa_pass) and notes the scene camera in replay order
(vk_dlaa_scan_draw), and the copy of the lit scene and the evaluation are records of the chunk that holds the op
(vk_dlaa_exec). A chunk's draws are jittered while they come before the op the scan ended the scene at. The
evaluation waits until the chunks before it are recorded, and the chunks after it wait for it (a hold), since they
sample its output. Chunks note their draws for the per-object motion one at a time (mo_lock in vk_exec_draw.c); the
uniform overrides of a motion draw and shader_for's last-frame variant flag are per thread, as the other chunks' draws
go on meanwhile; the FXAA shader table takes concurrent lookups. fxaa = 0 runs in chunks too.
With dlaa = 1 at 4K (DLSS Quality) on 1.2.1 and an RTX 5070, replay_workers = 4 takes the frame rate from 70 to 78
fps where the CPU is the limit, with no one-frame glitch in 241 consecutive frames.
(cherry picked from commit 04f8bb1b57)
replay_workers = N (0, the default, keeps the sequential replay) records a submission of enough draws in chunks that start between passes, on N threads and in twice as many chunks, and submits them in order. In the town at 4K on an RTX 5070, 4 threads take the replay from 8.2 to 4.5 ms a frame and the frame rate from 71 to 79.
A sequential scan on the submitting thread applies only what is queue state (fixed-function and bind ops, render targets, render enable and conditions, samples-passed counting) and keeps a copy of it where each chunk starts, with how far its lists got: inline uniform updates, RENDER_COND results, contact shadow marches and the ops that drop uniform copies. A chunk takes those from the lists instead of from the chunks before it, starts with every image in GENERAL and ends with a full barrier, so its recording depends only on the copy and the lists.
Each thread records with a context of its own: the replay's memos, descriptor state, occlusion query segments and per-command-buffer state moved from file statics into VkExec and VkExecDraw, and the tables the contexts share (stand-ins, depth twins, views, mirrors, timestamps, completion callbacks, crumbs) take locks or atomics. The scan makes render targets live in replay order and notes the op (vk_texture_revive), since a chunk can sample a storage before the chunk that renders it got there or after the scan went past a later target there: a chunk's texture slots resolve as of its own op, and the chunks after a texture with no entry yet wait for the chunk that makes it. Uniform copies are dropped at the op that writes them (vk_usnap_drop_at).
Contact shadows plan their march in replay order (vk_contact_plan) and the chunk that holds its op records it; the HDR source keeps the copy that comes last in replay order. DLSS and fxaa = 0 act on the frame draw by draw and keep the replay sequential. TOTK_REPLAY_STATE_CHECK=1 compares the state after the last chunk with the scan's.
A draw or dispatch whose shaders write storage buffers drops the uniform copies they cover from its own op on, as the other GPU writes do (scan_writes). The scan still translates no shader: it reads the shaders the recording contexts translated from a table shared with them (vk_exec_draw_writes), and an op whose shader no context translated yet holds the chunks after it until it is recorded. Before, the chunk that recorded such a write dropped the copies for every position at that moment, so a later chunk recorded first bound the copy taken at the submit call: on an RX 7800 XT on RADV the grass near the camera came out dark or unshaded in some runs with 4 threads.
TOTK_TRACE=perf counts the chunks held per submission by reason and logs when each chunk of the last submission began. The settings menu has the thread count under Advanced, Scheduling, with the count this run uses under it.
The recorder thread begins, ends and submits the submission's open part (vk_rec.c rec_end_cb), so a chunk's mirror prologue stays off it: vk_sync_part waits until the recorder is idle, as the part pool allows one thread at a time, then records, ends and submits the prologue itself. Through the recorder the prologue could reach the queue before it was ended, and the GPU lost the device at the title screen.
(cherry picked from commit b94e2e9642)
FCMP and the vector compares no longer flush denormal operands under FPCR.FZ: rt_fpcr_sync keeps MXCSR.DAZ equal to
FPCR.FZ, and x86 compares under DAZ read a denormal as a zero of its sign, as ARM does under FZ. The directed RMode
check of FN(op) and FN(vop) moved into the result check's branch, and min/max skip it, since they do not round.
Hot lifted functions shrink by about 10% of their x86 instructions; the town camera-pan bench (1080p, two runs each)
goes from 10.25 to 9.78 ms per frame. simd.py (60000 encodings), lift.py and func.py report 0 failures.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
With vb_used the streams a draw does not read got the zero buffer, so draws in a row that read different streams
bound all 17 slots again almost every draw. On an RTX 5060 Laptop that cost 16 % of the frame rate (issue #974). Now
only the read streams are compared, one bind covers the changed ones, and the others keep the buffer bound earlier
in the command buffer, which the deferred frees keep alive until the submission is done.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The setup wrote romfs-patches-pc for 16:9 only, so prompts = pc on a 21:9, 32:9, 16:10 or 3:2 screen kept the game
resolution at 1920x1080 (cfg.c) or, with output_res set to the screen's shape, fell back to the Switch prompts.
- make_romfs_patches.py --reuse <folder>: a file of an earlier overlay whose contents match is hard-linked (copied
where links fail) instead of compressed again; the output is byte for byte the same.
- The setup builds romfs-patches-pc-<w>x<h> for each of ASPECT_SIZES from romfs-patches-pc with --reuse: only the
Common archive is compressed, so the fifteen folders take about two minutes and 6 MiB each instead of 22 minutes
and 73 MiB each.
- cfg.c: the display-shaped output_res default applies with prompts = pc when romfs-patches-pc-<w>x<h> is there
(romfs_shaped_pc_ready); Xbox prompts keep 1920x1080 as before.
An i5-1235U or Core 3 100U has 2 P-cores. Binding the five frame-critical threads there, with a spinning recorder
among them, crowds two cores and their hyperthreads while the E-cores stay idle. plat_perf_core_setup now counts the
allowed physical P-cores and selects nothing below PC_MIN_CORES (4), so the scheduler places those threads as before.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
rec_submit queued each part's submission and then drained the ring until the recorder had submitted it: 0.29 ms a
frame in the town on a Ryzen 7900X, 0.34 ms on 4 cores. Now it publishes the submission and returns
(TOTK_REC_ASYNC=1, default). The queue order the present, uploads, readbacks and device waits rely on is kept by
vk_queue_lock: a thread's first lock waits until the recorder submitted what was queued before (rec_queue_wait).
The recorder itself and a thread that already holds the lock skip the wait. rec_submit reports the first failed
submission the recorder made. Town, in-run A/B: replay -4.1% on all cores, -2.9% on 4 cores / 8 threads; queue
users waited about 0.02 ms a frame. fight2 with TOTK_REC_CHECK and the town under sync validation are clean.
Where the recorder does not poll (fewer than 6 usable logical processors) the producer now publishes the ring head
every 8192 words instead of 64, as Savage Wind Recomp does (runtime/nvn/vulkan/record.c rec_init): each publish
there costs an event set and a wake-up. Town on 4 logical processors: frame interval -1.3%. TOTK_REC_PUB=ab
compares both; the perf log counts the recorder's wake-ups and the queue waits.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
On an Intel CPU with P-cores and E-cores a thread on an E-core runs at about 60% of a P-core's speed, and the frame
waits for its slowest critical thread. The game's main thread, the thread that presents, and the vk replay, recorder
and present threads now select the CPU sets of the P-core logical processors the process affinity allows. Nothing is
selected on a CPU with one efficiency class or when the affinity leaves only one kind of core. TOTK_PCORE=0 turns it
off; "power:" log lines say what was selected and which threads took it. LAS Recomp does the same for its game,
render and vsync threads.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The executable exports NvOptimusEnablement and AmdPowerXpressRequestHighPerformance, which NVIDIA Optimus and AMD
switchable graphics read at start to run it on the dedicated GPU; Optimus can otherwise hide its GPU from a program
it put on the integrated one. At start the process opts out of EcoQoS (E-cores and low clocks on hybrid Intel CPUs
when Windows deems it in the background) and of Windows 11 ignoring its timer resolution while the window is hidden.
One "power:" log line says whether Windows accepted it.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Guest waits and waits for a virtual core usually end within microseconds. A thread now polls for 20 us with the
global lock released before it sleeps on its condition variable, so a quick wake costs no OS wake-up
(TOTK_KSPIN_US, 0 turns it off). The recorder polls by the clock for 60 us instead of 4096 pauses, which last a few
us on Zen 4 and about 140 us on Coffee Lake (TOTK_REC_SPIN_US). Both stay off with fewer than 6 usable logical
processors. Savage Wind Recomp polls the same way before it parks.
Machine A, town save, in-run A/B: frame interval -1.8% on average for the kernel poll (5 of 7 block pairs), -0.7%
with the process limited to 4 cores / 8 threads; the recorder poll is neutral there.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Shadow storage is on by default under Wine as well, and the region snapshots were only measured on a native Linux
host. TOTK_SHADOW_SNAP=1 still turns them on under Wine.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Each 2 MiB region is now queried at most once per guest submission, shared with the texture checks, instead of once
per pool at every part of a submission. On a Steam Deck at Lookout Landing the replay thread's write tracking fell
from about 4.4 to 1.3 ms a frame and the frame rate rose about 7.5%. TOTK_SHADOW_SNAP=0 restores the old queries.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
vk_pool_fini freed import, physical and virtual pools right after vkQueueWaitIdle, which covers submitted work only;
with the synchronous replay another game thread may be recording a submission that names the pool. Every pool kind
now retires through the completion queue at the later of the recording's value and the next one.
destroy_tex freed the texture record while other threads still held &e->pub (render targets for a pass, storage
bindings writing wr_fence through it): they read a stale VkImage and cached views by that handle. The record is now
freed with its image's stamp (vk_defer_free), and views of a dropped entry are deferred at once instead of cached.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
With the synchronous replay (the default since 7be68ec) a game thread finalizing a texture destroys released texture
cache entries while another game thread records a submission that slot_resolve may already have handed one of them.
defer() saw the recording's value only on the recording thread and stamped the others with the last submitted value,
so the completion thread freed the image before the submission naming it ran: "write invalid" device faults on NVIDIA
(#788, #802). vk_sync_submit now notes the value at the start of the recording, and every deferral made from then on
waits for it. A fight run counted 10 such deferrals during the world load.
Device faults now log the deferred-free state, and with VK_EXT_device_address_binding_report the object that held the
faulting address and whether it was unbound before the fault (TOTK_ADDR_REPORT=0 turns the map off). The census line
counts deferrals made off the recording thread during a recording.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The copy time bounds how long a compute miss on the drawing thread can wait for the save now that builds and the copy exclude each other; a put-off autosave is logged too.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The replay thread deswizzled every block-linear texture on the CPU, and entering a new area stalled it for 50-236 ms. Now a compute pass reads the texture straight from the pool's imported buffer and writes a linear copy into a device-local scratch pool; the upload batch copies it into the image, or the ASTC decoder takes it. Large texture loads submit the batch early so the staging and scratch pools do not grow on the replay thread, and both pools are made at device creation.
On fight2 the render thread waits about 270 ms on textures per run instead of 900, and the load frame drops from 145 to about 80 ms. Mean fps is unchanged. TOTK_DESWIZZLE=cpu turns it off and check compares every upload with the CPU result (0 of 196k regions differ). Shadow mode (Linux/Wine) and textures outside the device's limits stay on the CPU.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
An exception no handler took (a fault in a GPU driver, a C++ exception a driver let escape) listed only the return addresses in totk.exe, and those match only the reporter's own build. #879 and #285 crash in the AMD driver with nothing in the log saying which driver code failed or what it threw.
The unhandled-exception filter now lists the frames of every module on the faulting thread with the module name and offset, up to 12 outside totk.exe so this executable's frames still fit. For a C++ exception thrown by MSVC-built code it reads the thrown type's decorated name from the ThrowInfo, and what() for a std::exception, with every read checked. The guest pc of a host function call is printed with that host function's name.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
While a game menu has the pointer the left mouse button is A, and it was Y's only default key, so keyboard and mouse players could not sort the inventory or press Y anywhere in the menus. T is now Y's second default key.
A bind line in totk.ini now also removes its keys from the defaults of buttons without a line, as the remapping rows already do: with bind_X = key:t, T presses X only, not X and Y.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The bindings dropped every mouse button while the cursor was not captured, and the menu pointer releases the capture,
so side buttons bound to L and R turned the menu's tabs only in the frames before the pointer appeared. While the menu
pointer is shown and over the window, the middle and side buttons now count; the left and right buttons stay the
menu's A and B.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Each change of pipeline state in prepare_draw re-recorded the whole dynamic state through vk_pipeline_set_dynamic:
about twenty commands, the vertex input packet rebuilt, and the blend enable, equation and write mask of every
attachment, however few fields differed. Every replay pipeline declares the same dynamic states, which a pipeline bind
leaves as they were, so vk_pipeline_set_dynamic_delta compares the values with what the command buffer already holds
and records only the commands that differ; the vertex input is rebuilt only when the vertex shader, attributes or
streams change. The record is per command buffer (B.dyn_valid, cleared with the rest of B on a new one or a split).
TOTK_DYN_DELTA=0 records the full set as before; ab alternates; TOTK_TRACE=perf counts calls, commands and calls that
changed nothing.
1.4.3, RTX 4070 Ti, 2560x1440 DLAA, standing still on Great Sky Island, TOTK_AB=10 (tools/ab_summary.py): 13.75 ms
a frame with the full sets, 12.59 ms with the deltas (-8.5%, 72.7 to 79.4 fps); replay 7.77 to 6.67 ms. An A/A run
with the same method moved the frame by 0.08 ms.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01BspNcezsiV8zT8wcnXrrME
After vk_exec_ub_flush the full barrier barrier_to records makes the copied bytes visible, so later bindings can read
the real buffer. The overlays stayed until the submission's end instead: every later binding inside their span
scanned them, and a block an overlay covered in part got a merged scratch copy that the descriptor build skip then
left alone. With the updates between passes overlaid too (Great Sky Island: 274 of 594 per submission) that was 290
merged copies and 1.1 ms more replay a frame, more than the barriers saved on the GPU (0.3 ms).
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01BspNcezsiV8zT8wcnXrrME
An update between passes was copied at once behind a full barrier of its own, and the copy it recorded made the next
update or pass begin need another: a run of k updates before a pass cost k + 1 full barriers, about 250 updates a
frame on Great Sky Island. They now become overlays like the updates inside a pass; every GPU reader of the real
buffer after them comes behind barrier_to, which records the copies first. TOTK_UB_OVERLAY_OUT: 1, 0, ab; the perf
line counts the overlays made between passes.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01BspNcezsiV8zT8wcnXrrME
vk_exec_ub_flush recorded the copies of a pass's overlaid uniform updates back to back behind one barrier. Copies
with no barrier between them may run in any order, and ov_add only drops an older pending copy of exactly the same
range, so two updates that overlap in part could leave the older bytes in the buffer where they overlap. A copy that
overlaps one recorded since the last barrier now waits for it behind a transfer-to-transfer barrier; the perf line
counts those barriers.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01BspNcezsiV8zT8wcnXrrME
resolve_menu_lang falls back from a region tag the catalog lacks to its base language (ru-RU: ru), but probe_menu_text expected English for such a tag, so --probe-services failed on any Windows whose UI language is ru-RU, de-DE and the like; en-US maps directly and passed.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
963901c added the Speakers row (stereo or 5.1) in English and Russian only, so --probe-services failed: menu text ids 189 to 193 lacked pt-BR and zh-Hans, and the menu showed them in English in those languages.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
tools/check_platform.py failed on vk_dlaa.c: it read the NGX core's folder from the registry, loaded it with LoadLibraryA and GetProcAddress and named the executable's folder with GetModuleFileNameW. platform.h gets plat_ngx_core_path (the registry folder plus _nvngx.dll on Windows, libnvidia-ngx.so.1 on Linux), plat_exe_dir_wide (the folder as NGX's wchar_t string; the UTF-8 to UTF-32 conversion moves from vk_dlaa.c to linux.c) and plat_lib_error, and the core loads through plat_lib_open and plat_lib_sym on both hosts.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The game keeps about ten vertex streams bound while a draw reads a few of them, and bind_vertex_buffers resolved and
compared all 17 slots on every draw. vk_pipeline_set_dynamic now returns the bindings of the state's vertex input,
from the same vertex_input that builds the dynamic input or the pipeline's, and the other slots get the zero buffer.
A stream is taken from the previous bind only when it was resolved there. In the fort, bind_vertex_buffers drops from
9.0% to 7.2% of the replay thread's work (two profiles, recorder on). TOTK_VB_USED=0 binds every stream as before.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>