AnyPS5/docs/dev/ENVIRONMENT.md

Environment variables

The code reads environment names through getenv, IsEnvironmentSet, kibSetting, MaskFromEnvironment, EnvInt, GetEnvironmentVariableA, GetEnvironmentVariableW with a wide literal, configured, PipelineSwitch, environmentCount and addWideString (the wide names the Windows entry stub reads). This table lists every APS5_* and ANYPS5_* name it reads, its kind, the value used when it is unset and what it does, one table per library. Regenerate the names with python3 tools/env_reference.py --table; the Conventions check fails when a name the code reads has no row here, or a row's name is no longer read.

core/libs

Name	Kind	Source	Default	Description
ANYPS5_LOADEXEC_PARENT	string	core/libs/tests/GuestLoadExecHost.cpp		variable the Windows entry stub reads; its wide name is built in core/relinker/elfpatcher/src/windows/WindowsEntryStubBuilder.cpp
ANYPS5_LOADEXEC_TEST_ROOT	string	core/libs/tests/GuestLoadExecHost.cpp		directory the load-exec host test reads its guest executable from
ANYPS5_TEST_QUOTED_ARGV0	presence	core/libs/tests/GuestLoadExecHost.cpp		makes the load-exec test's first argument one with quotes, a backslash and non-ASCII text
APS5_EBOOT	string	core/libs/tests/GuestSystemExit.cpp		host path of the executable launched for /app0/eboot.bin, defaulting to the running executable
APS5_SKIP_TIMING_TESTS	presence	core/libs/tests/PreciseWait.cpp		treats the timing tests' runner as preempted, so they skip without probing the clock
APS5_USBD_DEVICE	string	core/libs/tests/GuestUsbd.cpp		selects the USB device a test drives, as vendor:product in hex

core/relinker

Name	Kind	Source	Default	Description
ANYPS5_GUEST_MODULE_ARGS	string	core/relinker/elfpatcher/tests/GuestModuleArgsHost.cpp		path of the file the module-argument fixture appends its events to

core/shader

Name	Kind	Source	Default	Description
ANYPS5_NO_SHADER_CACHE	presence	core/shader/recompiler/ShaderDiskCache.cpp		disables the on-disk shader cache when set to a non-empty value other than 0
ANYPS5_SHADER_CACHE_DIR	string	core/shader/recompiler/ShaderDiskCache.cpp		directory for the on-disk shader cache, overriding the one beside the executable
APS5_BDA_BYTE_READS	presence	core/shader/recompiler/SpirvBackend/src/SpirvBdaRead.cpp		forces one address lookup and byte load per byte for BDA reads instead of a wide dword load
APS5_BDA_DWORD_STORES	presence	core/shader/recompiler/Translation/src/MemoryInstructions/FlatInstructions.cpp		reverts BDA writes to one address store per dword instead of a wide multi-dword store
APS5_DUMP_IR	string	core/shader/recompiler/Recompiler.cpp		hex code address (or all) of the programs whose IR is printed after each front-end pass
APS5_GATHER_CONST	integer	core/shader/recompiler/SpirvBackend/src/SpirvImageEmitter.cpp		integer every non-depth gather returns in all four texels instead of the sampled value
APS5_IMAGE_TABLE_KEY_SCAN	integer	core/shader/recompiler/Optimization/src/ResourceMaterializer.cpp	256	number of image-table entries the key scan examines
APS5_IMAGE_TABLE_STRICT	presence	core/shader/recompiler/Optimization/src/ResourceMaterializer.cpp		sets image-table materialization to its strict form; 0 leaves it off
APS5_NO_ATOMIC_BARRIER	presence	core/shader/recompiler/SpirvBackend/src/SpirvMemory/SpirvFormatConvert.cpp		drops the device-scope acquire/release memory barrier emitted after atomics
APS5_NO_ATOMIC_ZERO_SKIP	presence	core/shader/recompiler/SpirvBackend/src/SpirvMemory/SpirvMemoryInstructions.cpp		restores plain buffer atomics instead of skipping identity-operand atomics
APS5_NO_CODE_HASH_KEY	presence	core/shader/recompiler/CacheKey.hpp		appends every code word to the cache key instead of hashing the code
APS5_NO_FAILURE_MEMO	presence	core/shader/recompiler/Recompiler.cpp		rebuilds a program whose compilation failed instead of reusing the memoized failure
APS5_NO_IMAGE_TABLES	presence	core/shader/recompiler/Optimization/src/ResourceTracker.cpp		disables image tables, leaving table-loaded descriptors unplanned
APS5_NO_PERF_DISK_PLANS	presence	core/shader/recompiler/Recompiler.cpp		disables the disk cache of planning results
APS5_NO_PERF_DOM_INTERSECTION	presence	core/shader/recompiler/ControlFlow/src/Structurizer.cpp		computes dominator intersections into a fresh vector instead of intersecting in place
APS5_NO_PERF_ITERATION_SELECTION	presence	core/shader/recompiler/ControlFlow/src/Structurizer.cpp		recomputes the post-dominator sets when finding a loop's iteration-selection merge instead of reusing them
APS5_NO_RESULT_MEMO	presence	core/shader/recompiler/Recompiler.cpp		materializes a fresh result for every recompile instead of reusing the result memo
APS5_NO_SRT_TAPE	presence	core/shader/recompiler/Optimization/src/SrtWalker/SrtTape.cpp		runs the runtime-source walk of a plan in the IR evaluator instead of the compiled tape
APS5_NO_TABLE_RESOLUTION_REUSE	presence	core/shader/recompiler/Optimization/src/ResourceMaterializer.cpp		stops a result memo miss from reusing an unchanged image table's resolution, so the table is resolved again
APS5_NO_TABLE_SNAPSHOT_REUSE	presence	core/shader/recompiler/Optimization/src/ResourceMaterializer.cpp		turns off the image table snapshot reuse, so every capture reads and compares each table column
APS5_NO_WINDOW_NARROWING	presence	core/shader/recompiler/SpirvBackend/src/SpirvMemory/SpirvModuleSetup.cpp		indexes every slot of the shader's vector window instead of only the slots the program uses
APS5_PROBE	string	core/shader/recompiler/Translation/src/DispatchInstructions.cpp		pc:vgpr[:shift] probe that copies that VGPR into v255 after the instruction so image stores write it as a per-lane trace
APS5_RAYTRACING	string	core/shader/recompiler/Translation/src/DispatchInstructions.cpp		selects the image_bvh_intersect_ray behaviour: strict throws and miss writes 0xffffffff
APS5_SAVEEXEC_WRITE_FIRST	presence	core/shader/recompiler/Translation/src/ControlFlowInstructions.cpp		writes the saved exec to the destination before reading the source in s_*_saveexec
APS5_SINGLE_LANE	string	core/shader/recompiler/Recompiler.cpp		hex code addresses (comma separated, or all) of wave64 programs kept at one guest lane per invocation
APS5_SRT_DEBUG	presence	core/shader/recompiler/Optimization/src/RequestMemoryView.cpp		prints extra SRT descriptor-walk debug output, such as guest reads outside the request memory
APS5_TRACE_BDA	presence	core/shader/recompiler/Optimization/src/ResourceTracker.cpp		prints the first few accesses that make a program take the address-based path
APS5_TRACE_BUFFER_WRITTEN	presence	core/shader/recompiler/Optimization/src/DescriptorBindingBuilder.cpp		prints how many guest buffer elements a program may store to and how many it only loads
APS5_TRACE_IMAGE_TABLE	presence	core/shader/recompiler/Optimization/src/ResourceMaterializer.cpp		traces image-table resolution
APS5_VERIFY_SRT_COMPACT	presence	core/shader/recompiler/Optimization/src/SrtWalker/SrtDescriptorEvaluation.cpp		walks every SRT descriptor a second time through the IR evaluator and throws when the two walks differ
APS5_VERIFY_SRT_TAPE	presence	core/shader/recompiler/Optimization/src/SrtWalker/SrtTape.cpp		walks a plan's runtime sources a second time through the evaluator, adopts its answers where the two differ and prints the first twenty differences
APS5_WAVE_LDS_SUBGROUP	presence	core/shader/recompiler/SpirvBackend/src/SpirvEmitter.cpp		restores subgroup-scope LDS barriers for single-wave workgroups instead of workgroup scope

libSceAgc

Name	Kind	Source	Default	Description
APS5_TRACE_LABEL	integer	core/libs/prx/libSceAgc/Cb/src/Sync.cpp	0	hex guest address of a RELEASE_MEM label whose packet is logged when recorded

libSceAgcDriver

Name	Kind	Source	Default	Description
ANYPS5_GPU	string	core/libs/prx/libSceAgcDriver/Execution/src/VulkanDevice.cpp		restricts physical device selection to device names containing this case-insensitive substring
ANYPS5_OVERLAY	string	core/libs/prx/libSceAgcDriver/Execution/src/PerformanceOverlay.cpp	0	1 shows the performance overlay; 0 or unset hides it, and any other value is refused
ANYPS5_OVERLAY_POS	string	core/libs/prx/libSceAgcDriver/Execution/src/PerformanceOverlay.cpp	tl	the overlay's corner: tl, tr, bl or br
ANYPS5_PRESENT_MODE	string	core/libs/prx/libSceAgcDriver/Execution/src/VulkanDevice.cpp	fifo	selects the swapchain present mode: fifo, fifo_latest_ready, mailbox or immediate
ANYPS5_REQUIRE_DESCRIPTOR_BINDING_FLAGS	presence	core/libs/prx/libSceAgcDriver/tests/ComputePipelineReplay.cpp		fails the replay test instead of accepting a device that refuses descriptor binding flags
ANYPS5_REQUIRE_ROBUST2	presence	core/libs/prx/libSceAgcDriver/tests/execution/LdsMixedWidthAtomics.cpp		fails the LDS lock-budget test when the device does not enable robustBufferAccess2, instead of only printing it
ANYPS5_REQUIRE_VULKAN	presence	core/libs/prx/libSceAgcDriver/tests/BdaDevice.cpp		fails a test instead of skipping when no usable Vulkan device is present
ANYPS5_SCREENSHOTS	string	core/libs/prx/libSceAgcDriver/Execution/src/Screenshot.cpp		directory the screenshot key saves its PNG files to, defaulting to anyps5-screenshots beside the executable
APS5_ADDRESS_COPY_MAX_MIB	integer	core/libs/prx/libSceAgcDriver/Graphics/src/GuestBufferMemory.cpp	64	upper bound in MiB on the guest memory one address-based build copies; exceeding it aborts
APS5_ALL_BUFFERS_WRITTEN	presence	core/libs/prx/libSceAgcDriver/Graphics/src/ShaderResources.cpp		notes every shader buffer element as written, as before bufferWritten existed
APS5_ASSERT_GPU_LOCK	presence	core/libs/prx/libSceAgcDriver/Execution/src/GuestMemory.cpp		aborts when a guarded path is reached without holding GuestMemory::GpuMutex
APS5_ATOMIC_STAGE_MAX_KIB	integer	core/libs/prx/libSceAgcDriver/Graphics/src/GuestBufferMemory.cpp	1024	size cap in KiB on an atomic shader element covered by device-local staging
APS5_BARRIER_VALIDATE	presence	core/libs/prx/libSceAgcDriver/Graphics/src/Recorder.cpp		counts the barriers a hazard-tracking validator would emit and skip, by class and hazard kind
APS5_BATCH_CAP	integer	core/libs/prx/libSceAgcDriver/Execution/src/Driver/Synchronization/DeferredLabels.cpp	64	recorded work since the last submit at which a label batch, flush or reap is forced (0 disables the cap)
APS5_BDA_FLUSH	presence	core/libs/prx/libSceAgcDriver/Graphics/src/GuestBufferMemory.cpp		also flushes pending storage results over address-based shader regions, which are otherwise exempt
APS5_BDA_TABLE_CACHE_ENTRIES	integer	core/libs/prx/libSceAgcDriver/Graphics/src/BdaResources.cpp	8	number of recently used BDA page tables kept in the cache (1 is the previous behaviour)
APS5_BLOCK_WRITEBACK_EACH	presence	core/libs/prx/libSceAgcDriver/Graphics/src/Texture.cpp		stores the touched units only on every write-back instead of widening repeated partial stores
APS5_BLOCK_WRITEBACK_FRAMES	integer	core/libs/prx/libSceAgcDriver/Graphics/src/Texture.cpp	8	window of presents within which repeated partial stores trigger the whole-image widening
APS5_BLOCK_WRITEBACK_PIECES	integer	core/libs/prx/libSceAgcDriver/Graphics/src/Texture.cpp	4	number of partial stores within the window that triggers the whole-image widening
APS5_BLOCK_WRITEBACK_WIDEN	presence	core/libs/prx/libSceAgcDriver/Graphics/src/Texture.cpp		enables the widening that stores a whole image when it keeps being asked for in pieces
APS5_BUFFER_POOL_SHARED	presence	core/libs/prx/libSceAgcDriver/Graphics/src/BufferPool.cpp		uses one buffer tier for every allocation size, as before the tier split
APS5_BUFFER_POOL_SLOTS	integer	core/libs/prx/libSceAgcDriver/Graphics/src/BufferPool.cpp	4096	bounds the buffers each pool tier retains (0 keeps the default of 4096)
APS5_CAPTURE_GPU_COPIES	presence	core/libs/prx/libSceAgcDriver/Graphics/src/GuestBufferMemory.cpp		records GPU buffer copies in the capture trace (requires APS5_CAPTURE_TRACE)
APS5_CAPTURE_INPUTS	presence	core/libs/prx/libSceAgcDriver/Graphics/src/Draw.cpp		records each draw's inputs in the capture trace (requires APS5_CAPTURE_TRACE)
APS5_CAPTURE_TRACE	presence	core/libs/prx/libSceAgcDriver/Execution/include/CaptureTrace.hpp		writes capture events to capture-trace.log
APS5_CHECK_INDIRECT_ARGS	presence	core/libs/prx/libSceAgcDriver/Graphics/src/Draw.cpp		reads GPU-side indirect draw records back on the CPU and prints the first ones
APS5_COMPLETION_STORE_SYNC	presence	core/libs/prx/libSceAgcDriver/Graphics/src/Recorder.cpp		makes a store made by a completion action wait for later batches over its range
APS5_COPY_DESTINATION_STRICT	presence	core/libs/prx/libSceAgcDriver/Execution/src/VulkanDevice.cpp		refuses a CPU copy whose destination has any pending writer, not only an unsettled one
APS5_COPY_HLE_CPU_MAX	integer	core/libs/prx/libSceAgcDriver/Execution/src/Driver/Memory/BufferCopy.cpp	262144	largest copy in bytes the CPU side of the copy HLE handles
APS5_COPY_HLE_GPU_MAX	integer	core/libs/prx/libSceAgcDriver/Execution/src/Driver/Memory/BufferCopy.cpp		largest copy in bytes the GPU side of the copy HLE handles; unset means no limit
APS5_COPY_READ_TRACKING	presence	core/libs/prx/libSceAgcDriver/Graphics/src/Recorder.cpp		notes in-place reads of a range; set to 0 to disable, making the CPU copy path require an idle recorder
APS5_COPY_SYNC	presence	core/libs/prx/libSceAgcDriver/Execution/src/VulkanDevice.cpp		treats any pending writer over a copy's range as needing a wait
APS5_COPY_VERIFY	presence	core/libs/prx/libSceAgcDriver/Execution/src/VulkanDevice.cpp		verifies a CPU copy's source and destination bytes against the pre-copy values after a sync
APS5_COPY_WAIT_SOURCE	presence	core/libs/prx/libSceAgcDriver/Execution/src/VulkanDevice.cpp		waits for the source's producer batch when it is the only one in flight, on non-zero queues
APS5_COUNT_ALL_COMPLETION_LABELS	presence	core/libs/prx/libSceAgcDriver/Graphics/src/Recorder.cpp		counts a completion label also stored on the GPU as pending from its registration
APS5_CPU_COPIES	presence	core/libs/prx/libSceAgcDriver/Graphics/src/GuestBufferMemory.cpp		copies misaligned descriptor sub-ranges on the CPU instead of with a GPU buffer copy
APS5_CPU_DCC_KEYS	presence	core/libs/prx/libSceAgcDriver/Graphics/src/DccMetadata.cpp		always stores DCC keys on the CPU
APS5_CPU_FLIP_IN_ORDER	presence	core/libs/prx/libSceAgcDriver/Execution/src/Driver/Queues/Submission.cpp		orders every CPU flip behind queued work instead of presenting unqueued flips at once
APS5_CPU_STORES	presence	core/libs/prx/libSceAgcDriver/Execution/src/Driver/Packets/PacketMemory.cpp		writes packet stores on the CPU instead of recording them on the GPU
APS5_DCC_KEYS_EACH	presence	core/libs/prx/libSceAgcDriver/Graphics/src/Recorder.cpp		records every DCC key store at once instead of batching them
APS5_DEPTH_COPY_MIN_KIB	integer	core/libs/prx/libSceAgcDriver/Graphics/src/ShaderResources.cpp	1024	minimum surface size in KiB for a depth target to take the depth-copy path
APS5_DISPATCH_CACHE_BYTES	integer	core/libs/prx/libSceAgcDriver/Execution/src/Driver/Dispatch/DispatchCache.cpp	268435456	size of the dispatch cache in bytes (0 keeps the default of 256 MiB)
APS5_DISPATCH_CACHE_ENTRIES	integer	core/libs/prx/libSceAgcDriver/Execution/src/Driver/Dispatch/DispatchCache.cpp	8192	number of dispatch cache entries; 0 picks a default from the variant count
APS5_DISPATCH_STAMP_VALIDATE	presence	core/libs/prx/libSceAgcDriver/Execution/src/Driver/Dispatch/DispatchCache.cpp		validates dispatch cache hits by write stamps, forcing one variant and a 4096-entry cache
APS5_DISPATCH_VARIANTS	integer	core/libs/prx/libSceAgcDriver/Execution/src/Driver/Dispatch/DispatchCache.cpp	4	number of dispatch variants kept per cache entry, clamped to the maximum
APS5_DISPMOVE_MAX_KIB	integer	core/libs/prx/libSceAgcDriver/Execution/src/VulkanDevice.cpp	1024	upper bound in KiB on the read-only buffers a dispatch-cache hit moves into its template, above which the dispatch is prepared again
APS5_DRAIN_ALL	presence	core/libs/prx/libSceAgcDriver/Execution/src/Driver/Packets/PacketMemory.cpp		drains every packet, disabling the batched recording path
APS5_DRAIN_COMPLETION_LABELS	presence	core/libs/prx/libSceAgcDriver/Execution/src/Driver/Synchronization/DeferredLabels.cpp		records completion labels immediately instead of deferring them with the batch
APS5_DRAW_CACHE_ENTRIES	integer	core/libs/prx/libSceAgcDriver/Execution/src/Driver/Draw/DrawCache.cpp	4096	number of entries in the draw cache; 0 keeps the default 4096
APS5_DRAW_DRAIN	presence	core/libs/prx/libSceAgcDriver/Execution/src/Driver/Packets/PacketMemory.cpp		treats draw packets as drains, waiting for them instead of recording ahead
APS5_DRAW_INPUTS_IN_PLACE	presence	core/libs/prx/libSceAgcDriver/Execution/src/VulkanDevice.cpp		binds draw vertex and index ranges in place from host imports also with NVIDIA's Windows driver, which copies them by default
APS5_DRAW_KEY_ALL_USER_WORDS	presence	core/libs/prx/libSceAgcDriver/Execution/src/Driver/Draw/DrawRegisterKey.cpp		includes all 32 user words in a draw's register key instead of the used count
APS5_DRAW_SNAPSHOT_MAX_MIB	integer	core/libs/prx/libSceAgcDriver/Graphics/src/ShaderResources.cpp	64	upper bound in MiB on the snapshot bytes a draw's bindings may take
APS5_DRAW_TRANSITIONS	presence	core/libs/prx/libSceAgcDriver/Graphics/src/Draw.cpp		keeps the per-draw upload and download barriers and layout transitions instead of the lean recorded path
APS5_DRAW_UPLOAD_MAX	integer	core/libs/prx/libSceAgcDriver/Graphics/src/ShaderResources.cpp	4096	largest draw snapshot in bytes copied into the per-recorder upload chunk; 0 turns the chunk off
APS5_DUMP_COMPUTE_MIN_WORDS	integer	core/libs/prx/libSceAgcDriver/Execution/src/VulkanDevice.cpp		only captures compute pipelines whose SPIR-V has at least this many words
APS5_DUMP_COMPUTE_PIPELINES	string	core/libs/prx/libSceAgcDriver/Execution/src/VulkanDevice.cpp		directory in which captured compute pipelines are written
APS5_DUMP_DRAW_SHADERS	integer	core/libs/prx/libSceAgcDriver/Execution/src/Driver/Draw/Draw.cpp	0	hex color-target or slot-0 address whose draw's shaders are dumped
APS5_DUMP_DRAW_SLOT1	integer	core/libs/prx/libSceAgcDriver/Execution/src/Driver/Draw/Draw.cpp	0	hex slot-1 address whose draw's shaders and register state are dumped
APS5_DUMP_FRAMES	integer	core/libs/prx/libSceAgcDriver/Execution/src/VulkanDevice.cpp	0	saves the first n presented frames as frame_.bmp
APS5_DUMP_FRAMES_EVERY	integer	core/libs/prx/libSceAgcDriver/Execution/src/VulkanDevice.cpp	1	dumps every n-th presented frame
APS5_DUMP_QUEUE	integer	core/libs/prx/libSceAgcDriver/Execution/src/Driver/Packets/PacketExecution.cpp		dumps the first submissions of the queue with this hex id
APS5_DUMP_REJECTED	presence	core/libs/prx/libSceAgcDriver/Execution/src/Driver/Packets/CommandValidation.cpp		writes a rejected submission to rejected_submission_.bin
APS5_DUMP_SCALE	integer	core/libs/prx/libSceAgcDriver/Execution/src/VulkanDevice.cpp	4	keeps every n-th pixel in each direction of a dumped frame
APS5_DUMP_SHADERS	presence	core/libs/prx/libSceAgcDriver/Execution/src/Driver/Dispatch/Dispatch.cpp		writes each prepared shader's request to shader_.req
APS5_DUMP_STORAGE	string	core/libs/prx/libSceAgcDriver/Graphics/src/Texture.cpp		selects storage textures to dump by hex address or WxHxL[:format] extent as storage__.raw
APS5_DUMP_SUBMITS	string	core/libs/prx/libSceAgcDriver/Execution/src/SubmissionCapture.cpp		directory in which every submission is captured for replay
APS5_DUMP_TARGETS	integer	core/libs/prx/libSceAgcDriver/Graphics/src/Draw.cpp	0	dumps the first n renders of each non-resident color target as target__.raw in the working directory; a non-zero value also runs every draw synchronously, and in practice it needs APS5_NO_RESIDENT_TARGETS=1
APS5_DUMP_TEXTURE	string	core/libs/prx/libSceAgcDriver/Graphics/src/Texture.cpp		hex addresses or WxH extents whose sampled textures are dumped on first upload as texture__.raw
APS5_EAGER_WRITEBACK	presence	core/libs/prx/libSceAgcDriver/Graphics/src/Texture.cpp		writes back pending storage layers eagerly
APS5_EXACT_DRAW_WRITES	presence	core/libs/prx/libSceAgcDriver/Execution/src/Driver/Memory/WrittenBuffers.cpp		sizes a color or depth target's draw write ranges from its own bytes and slices instead of the slack and unknown-byte approximation; 0 leaves the approximation on
APS5_FILL_CHAIN	presence	core/libs/prx/libSceAgcDriver/Execution/src/VulkanDevice.cpp		seeds and doubles the fill pattern in place in the import
APS5_FILL_CLEAR_EXACT_ONLY	presence	core/libs/prx/libSceAgcDriver/Execution/src/Driver/Memory/BufferFill.cpp		turns a fill into a clear only when it covers the target exactly and nothing else
APS5_FILL_SYNC	presence	core/libs/prx/libSceAgcDriver/Execution/src/VulkanDevice.cpp		waits for every recorded store over the fill's range
APS5_FLIP_INFLIGHT	integer	core/libs/prx/libSceAgcDriver/Execution/src/VulkanDevice.cpp	1	number of presentations that may trail on the GPU (0, 1 or 2)
APS5_FLIP_READ_CHECK	presence	core/libs/prx/libSceAgcDriver/Graphics/src/Recorder.cpp		records the in-place reads of completed batches and checks them before the present blit's fence
APS5_FORCE_SDR	presence	core/libs/prx/libSceAgcDriver/Execution/src/HostHdr.cpp	unset	suppresses advertised HDR availability while retaining detected host capabilities; uses the existing SDR presentation path
APS5_FORCE_STRUCTURIZER_FALLBACK	presence	core/libs/prx/libSceAgcDriver/tests/ControlFlow.cpp		makes the control-flow test expect the reach-flag fallback instead of no cloning
APS5_FULL_BARRIERS	presence	core/libs/prx/libSceAgcDriver/Graphics/src/Recorder.cpp		records every barrier instead of merging redundant ones
APS5_GPU_CHECKPOINTS	presence	core/libs/prx/libSceAgcDriver/Execution/src/VulkanDevice.cpp		attaches device diagnostic checkpoints through VK_NV_device_diagnostic_checkpoints
APS5_GPU_COPY_MAX_KIB	integer	core/libs/prx/libSceAgcDriver/Graphics/src/GuestBufferMemory.cpp	1024	size cap in KiB on a descriptor sub-range copied through the GPU path, larger ones staying on the CPU
APS5_GPU_NO_WRITEBACK	presence	core/libs/prx/libSceAgcDriver/Graphics/src/ShaderResources.cpp		skips writing GPU results back into guest memory
APS5_GPU_WAIT_TIMEOUT_MS	integer	core/libs/prx/libSceAgcDriver/Execution/src/Driver/Synchronization/WaitMemory.cpp	1000	milliseconds a GPU wait may run before a warning is printed
APS5_GROUP_CAPTURE_ALL_QUEUES	presence	core/libs/prx/libSceAgcDriver/Execution/src/Driver/Dispatch/DispatchAhead.cpp		resolves group-ahead capture on queue 0 as well
APS5_HANDOFF_SPIN_US	integer	core/libs/prx/libSceAgcDriver/Execution/include/HandoffSpin.hpp		microseconds a thread spins for a handed-off draw's work or for the GPU mutex before blocking, clamped to 0..10000; unset spins 150 us when the process has more than 8 usable processors and not at all otherwise
APS5_HEAP_MIRROR_MIB	integer	core/libs/prx/libSceAgcDriver/Graphics/src/GuestBufferMemory.cpp	24576	budget in MiB for the heap mirrors held for registered allocations
APS5_HOOK_FLUSH_CPU_BLOCKS	presence	core/libs/prx/libSceAgcDriver/Graphics/src/Recorder.cpp		set to 0 flushes pending images whose every 64 KiB block the CPU wrote, instead of skipping them
APS5_HOOK_FULL_SYNC	presence	core/libs/prx/libSceAgcDriver/Graphics/src/Recorder.cpp		makes the flush hook's recorded-store wait a full Sync under the GpuMutex
APS5_HOOK_LOCKED_WAIT	presence	core/libs/prx/libSceAgcDriver/Graphics/src/Recorder.cpp		keeps the flush hook's pending-write wait under the GpuMutex
APS5_HOST_IMPORT_MIB	integer	core/libs/prx/libSceAgcDriver/Graphics/src/GuestBufferMemory.cpp	0	budget in MiB for pinned host imports; 0 selects the larger of 6144 MiB and half the physical memory, at most the largest host heap minus the larger of 2 GiB and half that heap (4055 MiB on a 16 GB Windows PC with an RTX 4060)
APS5_IGNORE_DEPTH_TEST	presence	core/libs/prx/libSceAgcDriver/Graphics/src/State.cpp		renders depth- and stencil-tested draws without a depth target as if the tests always passed
APS5_IMAGE_POOL_MIB	integer	core/libs/prx/libSceAgcDriver/Graphics/src/BufferPool.cpp		size in MiB of the released image memory the buffer pool keeps for reuse by memory type and size while the device-local heap is under its reported budget, a quarter of the device tier's budget by default; 0 frees image memory at once
APS5_INDIRECT_VERTEX_MIB	integer	core/libs/prx/libSceAgcDriver/Execution/src/Driver/Draw/IndirectDraw.cpp	64	cap in MiB on a vertex attribute extent, above which a GPU-side indirect draw resolves on the CPU
APS5_INSERT_COMPARE	presence	core/libs/prx/libSceAgcDriver/Execution/src/Driver/Dispatch/DispatchCache.cpp		compares a fresh capture's regions against the recorder before inserting it into the cache
APS5_KEPT_DESCRIPTOR_SETS	integer	core/libs/prx/libSceAgcDriver/Graphics/src/ShaderResources.cpp	8	number of freed dedicated image table descriptor sets kept for a later build with the same layout and sizes; 0 keeps none
APS5_KEYFLIP_DROP	presence	core/libs/prx/libSceAgcDriver/Graphics/src/Texture.cpp		drops a pending unit's results on a clear-code to uncompressed key flip, as before
APS5_KEYS_FILL_CLEAR	presence	core/libs/prx/libSceAgcDriver/Graphics/src/Texture.cpp		clears an image from DCC keys whose fill pattern it matches (only when set to 1)
APS5_LABEL_FLUSH_US	integer	core/libs/prx/libSceAgcDriver/Execution/src/Driver/Synchronization/DeferredLabels.cpp	1000	age in microseconds at which a pending label batch is recorded or submitted
APS5_LABEL_GATE_ALL	presence	core/libs/prx/libSceAgcDriver/Execution/src/VulkanDevice.cpp		gates a label store on any pending completion, not only one overlapping the range
APS5_LABEL_LOCK_EACH	presence	core/libs/prx/libSceAgcDriver/Execution/src/Driver/Synchronization/DeferredLabels.cpp		records each label under the GpuMutex instead of deferring it
APS5_LABEL_OWN_LOCK	presence	core/libs/prx/libSceAgcDriver/Execution/src/Driver/Packets/PacketFlush.cpp		makes a label flush take the GpuMutex itself instead of relying on a self-locking packet
APS5_LABEL_REAP_ALL_QUEUES	presence	core/libs/prx/libSceAgcDriver/Execution/src/VulkanDevice.cpp		reaps completion labels on every queue instead of only queue 0
APS5_LABEL_RUNS_INLINE	presence	core/libs/prx/libSceAgcDriver/Graphics/src/Recorder.cpp		records store runs at once and closes each run at the next command
APS5_LABEL_STORE_ALWAYS	presence	core/libs/prx/libSceAgcDriver/Graphics/src/Recorder.cpp		writes back every label store even when a later store already covered its bytes
APS5_LABEL_SUBMIT_NOW	presence	core/libs/prx/libSceAgcDriver/Execution/src/VulkanDevice.cpp		submits every label immediately instead of waiting for the flush deadline
APS5_LABEL_TRUST_LATE	presence	core/libs/prx/libSceAgcDriver/Graphics/src/Recorder.cpp		trusts cache entries with a stamp at or below the wait's stamp under the late rule (off unless set to a non-zero value)
APS5_LABEL_TRY_EACH_PACKET	presence	core/libs/prx/libSceAgcDriver/Execution/src/Driver/Synchronization/DeferredLabels.cpp		attempts to record deferred labels at every self-locking packet instead of deferring to its lock
APS5_LAND_WAITS_ALL_QUEUES	presence	core/libs/prx/libSceAgcDriver/Execution/src/Driver/Synchronization/WaitMemory.cpp		notes pending waits on queue 0 as well
APS5_LAZY_SHADER_PREPARE	presence	core/libs/prx/libSceAgcDriver/Execution/src/Driver/Shaders/ShaderRegistry.cpp		prepares a registered shader on its first use instead of while registering it
APS5_LDS_SLACK	integer	core/libs/prx/libSceAgcDriver/Graphics/src/ShaderInputState.cpp	0	dwords added to every dispatch's LDS allocation
APS5_LOCKED_BUILD	presence	core/libs/prx/libSceAgcDriver/Execution/src/VulkanDevice.cpp		builds the whole dispatch preparation under the GpuMutex, skipping the split path
APS5_LOCKED_DRAW_PREPARE	presence	core/libs/prx/libSceAgcDriver/Execution/src/Driver/Draw/Draw.cpp		takes the GpuMutex before preparing a draw
APS5_LOG_STATS	presence	core/libs/prx/libSceAgcDriver/Execution/include/AsyncLog.hpp		prints the AGC log's line, byte and coalescing statistics every window; 0 turns them off
APS5_LOG_SYNC	presence	core/libs/prx/libSceAgcDriver/Execution/include/AsyncLog.hpp		writes AGC log lines on the calling thread instead of through the asynchronous sink thread
APS5_LOOP_GUARD	presence	core/libs/prx/libSceAgcDriver/Graphics/src/ShaderResources.cpp		enables the fault buffer that guards shader loops running past their bound
APS5_MIMALLOC	presence	core/relinker/elfpatcher/src/windows/WindowsEntryStubBuilder.cpp		loads mimalloc.dll in the Windows entry stub, so the process allocates through mimalloc instead of the C runtime's heap
APS5_MIRROR_REFRESH_THREADS	integer	core/libs/prx/libSceAgcDriver/Graphics/src/GuestBufferMemory.cpp	3	number of helper threads that refresh heap mirrors, clamped to 0 through 16
APS5_NO_ADDRESS_READ_PASSES	presence	core/libs/prx/libSceAgcDriver/Graphics/src/Draw.cpp		stops a draw that only reads through its buffer descriptors from counting as read-only, so it breaks the pass instead of continuing the producer's
APS5_NO_ADDRESS_SPACE_CACHE	presence	core/libs/prx/libSceAgcDriver/Graphics/src/GuestBufferMemory.cpp		disables the address-space cache so every build rebuilds its space
APS5_NO_ADJACENT_GENERATION	presence	core/libs/prx/libSceAgcDriver/Graphics/src/Texture.cpp		leaves adjacent pending images at their generation during a write-back, as before
APS5_NO_ALIAS_BORROW	presence	core/libs/prx/libSceAgcDriver/Graphics/src/Texture.cpp		disables borrowing a pending aliased storage image
APS5_NO_ATOMIC_STAGING	presence	core/libs/prx/libSceAgcDriver/Graphics/src/GuestBufferMemory.cpp		binds atomic shader elements in place instead of staging them through a device-local buffer
APS5_NO_BARRIER_BATCH	presence	core/libs/prx/libSceAgcDriver/Graphics/src/Recorder.cpp		records the staging, write-back and label-run leading barriers that a covering barrier already orders, instead of leaving them out
APS5_NO_BARRIER_ELISION	presence	core/libs/prx/libSceAgcDriver/Graphics/src/Recorder.cpp		records every leading barrier instead of merging it into the following submission
APS5_NO_BDA_TABLE_CACHE	presence	core/libs/prx/libSceAgcDriver/Graphics/src/BdaResources.cpp		builds every BDA descriptor table instead of reusing the recent-table cache
APS5_NO_BLOCK_TRACKING	presence	core/libs/prx/libSceAgcDriver/Graphics/src/Texture.cpp		tracks array layers, or the whole surface, instead of 64 KiB blocks
APS5_NO_BUFFER_CLASSES	presence	core/libs/prx/libSceAgcDriver/Graphics/src/BufferPool.cpp		matches exact sizes in the buffer pool instead of rounding up to size classes
APS5_NO_CACHED_PRECOLLECT	presence	core/libs/prx/libSceAgcDriver/Execution/src/VulkanDevice.cpp		leaves the write-watch walks of a cached object's Revalidate under the GPU mutex
APS5_NO_CLEAR_HAND_OVER	presence	core/libs/prx/libSceAgcDriver/Graphics/src/Texture.cpp		writes other images' pending results back at a clear-key storage refresh instead of handing them over to the refreshed image
APS5_NO_CLEARED_VIEW	presence	core/libs/prx/libSceAgcDriver/Graphics/src/ShaderResources.cpp		keeps fast-cleared surfaces as CPU snapshots instead of viewing the cleared storage image
APS5_NO_COLLECT_MEMO	presence	core/libs/prx/libSceAgcDriver/Execution/src/GuestMemory.cpp		disables the per-thread memo of write-watch collects
APS5_NO_CONTEXT_CACHE	presence	core/libs/prx/libSceAgcDriver/Execution/src/VulkanDevice.cpp		rebuilds the Graphics::Context on every call instead of returning the cached one
APS5_NO_COPY_ALIAS	presence	core/libs/prx/libSceAgcDriver/Execution/src/VulkanDevice.cpp		keeps every device copy a transfer instead of aliasing the source storage image
APS5_NO_COPY_HLE	presence	core/libs/prx/libSceAgcDriver/Execution/src/Driver/Memory/BufferCopy.cpp		disables the copy-kernel HLE, so matching copies run as compute
APS5_NO_COPY_KNOWN_VALUES	presence	core/libs/prx/libSceAgcDriver/Execution/src/Driver/Memory/BufferCopy.cpp		disables serving a copy's destination from source values the driver already knows
APS5_NO_CPU_COPY	presence	core/libs/prx/libSceAgcDriver/Execution/src/Driver/Memory/BufferCopy.cpp		disables the CPU copy path of the copy HLE
APS5_NO_DATA_HITS	presence	core/libs/prx/libSceAgcDriver/Execution/src/Driver/Dispatch/DispatchCache.cpp		disables dispatch-cache hits that differ only in ShaderData or FlattenedSrt words
APS5_NO_DCC_PRESYNC	presence	core/libs/prx/libSceAgcDriver/Graphics/src/ShaderResources.cpp		leaves the DCC key ranges out of the surfaces a build pre-syncs, so only the surface itself waits for its producer
APS5_NO_DCC_WORD_SCAN	presence	core/libs/prx/libSceAgcDriver/Graphics/src/DccMetadata.cpp		restores the byte-by-byte DCC key scan instead of the word-at-a-time compare
APS5_NO_DEFERRED_EVICTION	presence	core/libs/prx/libSceAgcDriver/Execution/src/VulkanDevice.cpp		destroys displaced resource-cache entries under the mutex instead of deferring them to the open batch
APS5_NO_DEFERRED_SETTLES	presence	core/libs/prx/libSceAgcDriver/Graphics/src/GuestBufferMemory.cpp		settles watched private host imports as they open instead of through the deferred block-settling path (Windows only)
APS5_NO_DISPATCH_CACHE	presence	core/libs/prx/libSceAgcDriver/Execution/src/Driver/Dispatch/Dispatch.cpp		disables the dispatch cache so every dispatch takes the uncached path
APS5_NO_DISPATCH_KEY_HYGIENE	presence	core/libs/prx/libSceAgcDriver/Execution/src/Driver/Dispatch/Dispatch.cpp		mixes every shader register into the dispatch key instead of only the registers that affect the dispatch
APS5_NO_DISPATCH_LRU	presence	core/libs/prx/libSceAgcDriver/Execution/src/Driver/Dispatch/DispatchInsert.cpp		clears the whole dispatch cache on overflow instead of evicting the least recently used entry
APS5_NO_DISPATCH_MOVE	presence	core/libs/prx/libSceAgcDriver/Execution/src/VulkanDevice.cpp		leaves a dispatch-cache hit from moving its read-only buffers into the template, so it prepares the dispatch again; also off when APS5_NO_TEMPLATE_DATA_REFRESH is set
APS5_NO_DISPATCH_RECIPE	presence	core/libs/prx/libSceAgcDriver/Execution/src/VulkanDevice.cpp		builds no dispatch recipes, so dispatch-cache hits repeat the recorded work
APS5_NO_DRAW_DATA_HITS	presence	core/libs/prx/libSceAgcDriver/Execution/src/Driver/Draw/DrawCache.cpp		disables draw-cache hits that differ only in ShaderData words
APS5_NO_DRAW_INPUTS_IN_PLACE	presence	core/libs/prx/libSceAgcDriver/Execution/src/VulkanDevice.cpp		copies draw vertex and index ranges instead of binding them in place from host imports, on any device; wins over APS5_DRAW_INPUTS_IN_PLACE
APS5_NO_DRAW_KEY	presence	core/libs/prx/libSceAgcDriver/Execution/src/Driver/Draw/DrawCache.cpp		registers no draw key, so every draw decodes its programs instead of hitting the draw cache
APS5_NO_DRAW_KEY_TRIM	presence	core/libs/prx/libSceAgcDriver/Graphics/src/Draw.cpp		keys the render target and index ranges into the draw resource key instead of repeating the alias checks on a hit
APS5_NO_DRAW_PRECHECK	presence	core/libs/prx/libSceAgcDriver/Execution/src/Driver/Draw/DrawPrecheck.cpp		disables the draw precheck that rejects empty or invalid draws early
APS5_NO_DRAW_RECIPE	presence	core/libs/prx/libSceAgcDriver/Graphics/src/Draw.cpp		builds no draw recipes, so draw-cache hits repeat the recorded stages
APS5_NO_DRAW_RELOCATION	presence	core/libs/prx/libSceAgcDriver/Execution/src/Driver/Draw/DrawRelocate.cpp		disables draw relocation, which rewrites a draw's moved user words from a learned variant instead of preparing the draw again
APS5_NO_DRAW_RESOURCE_CACHE	presence	core/libs/prx/libSceAgcDriver/Graphics/src/Draw.cpp		builds every draw's shader resources instead of reusing a cached identical build
APS5_NO_DRAW_SRT_ENTRIES	presence	core/libs/prx/libSceAgcDriver/Execution/src/Driver/Draw/DrawCache.cpp		disables the draw cache's SRT entries, so every draw decodes its programs
APS5_NO_DRAW_STAGE_DATA_HITS	presence	core/libs/prx/libSceAgcDriver/Execution/src/Driver/Draw/DrawCache.cpp		disables draw-cache hits that differ only in a stage's ShaderData words
APS5_NO_EPOCH_COLLECT_MAP	presence	core/libs/prx/libSceAgcDriver/Execution/src/GuestMemory.cpp		drops the per-epoch map of completed collects, leaving only the fixed-size thread memo to answer a repeated collect
APS5_NO_EPOCH_REVALIDATE	presence	core/libs/prx/libSceAgcDriver/Graphics/src/ShaderResources.cpp		runs every Revalidate's per-element and per-region checks instead of the epoch gate
APS5_NO_FAST_REVALIDATE	presence	core/libs/prx/libSceAgcDriver/Graphics/src/ShaderResources.cpp		always repeats the resource lookups instead of using the fast proof
APS5_NO_FILL_CLEAR	presence	core/libs/prx/libSceAgcDriver/Execution/src/Driver/Memory/BufferFill.cpp		disables the clear-through-DCC-keys path in a fill
APS5_NO_FILL_HLE	presence	core/libs/prx/libSceAgcDriver/Execution/src/Driver/Memory/BufferFill.cpp		disables the fill-kernel HLE, so matching fills run as compute
APS5_NO_FLUSH_PRETEST	presence	core/libs/prx/libSceAgcDriver/Graphics/src/Texture.cpp		lists every pending image on FlushPending instead of testing whether a pending unit lies inside the range
APS5_NO_FOREIGN_WRITERS	presence	core/libs/prx/libSceAgcDriver/Execution/src/Driver/Memory/WriteEvidence.cpp		records no draw-written ranges as foreign writers for capture validation
APS5_NO_GATE_RETRY	presence	core/libs/prx/libSceAgcDriver/Execution/src/Driver/Memory/WriteEvidence.cpp		retries no validation whose entry moved, leaving the dispatch or draw unvalidated
APS5_NO_GPL	presence	core/libs/prx/libSceAgcDriver/Execution/src/VulkanDevice.cpp		draws without graphics pipeline libraries; the device still enables their extensions, so the driver builds its pipelines as for a default device
APS5_NO_GPU_DUMP	presence	core/libs/prx/libSceAgcDriver/Execution/src/VulkanDevice.cpp		decodes dump frames from the tiled guest buffer on the CPU
APS5_NO_GPU_INDIRECT	presence	core/libs/prx/libSceAgcDriver/Execution/src/Driver/Dispatch/IndirectDispatch.cpp		reads indirect dispatch arguments on the CPU instead of letting the GPU count them
APS5_NO_GPU_INDIRECT_DRAW	presence	core/libs/prx/libSceAgcDriver/Graphics/src/Draw.cpp		reads indirect draw arguments through the flush hook on the CPU instead of from the GPU
APS5_NO_GPU_SAMPLE_DUMPS	presence	core/libs/prx/libSceAgcDriver/Execution/src/VulkanDevice.cpp		dumps occlusion-counter samples on the CPU instead of through the GPU copy path
APS5_NO_GPU_THREAD_DIMENSIONS	presence	core/libs/prx/libSceAgcDriver/Execution/src/Driver/Dispatch/IndirectDispatch.cpp		reads the thread counts of an indirect dispatch with USE_THREAD_DIMENSIONS on the CPU instead of turning them into group counts and a thread limit on the GPU
APS5_NO_HOLD_PROFILE	presence	core/libs/prx/libSceAgcDriver/Execution/src/GuestMemory.cpp		turns off GpuMutex hold-time profiling, keeping only the wait timing
APS5_NO_HOOKSYNC_PROFILE	presence	core/libs/prx/libSceAgcDriver/Graphics/src/Recorder.cpp		leaves only the plain hook-sync counts under APS5_PROFILE_DRAW
APS5_NO_HOOK_COMPLETION_GUARD	presence	core/libs/prx/libSceAgcDriver/Graphics/src/Recorder.cpp		lets the flush hook record the accessing worker's queued labels even from inside a completion action
APS5_NO_HOOK_SNAPSHOT	presence	core/libs/prx/libSceAgcDriver/Graphics/src/Recorder.cpp		takes the GpuMutex on every access instead of using the pending-write snapshot
APS5_NO_HOST_IMPORT	presence	core/libs/prx/libSceAgcDriver/Execution/src/VulkanDevice.cpp		disables importing guest memory, so address-based shaders copy instead of using it in place
APS5_NO_HW_F32_FLUSH	presence	core/libs/prx/libSceAgcDriver/Execution/src/VulkanDevice.cpp		skips the startup f32 denormal flush probe and keeps the emulated flush in every shader
APS5_NO_HW_F32_PRESERVE	presence	core/libs/prx/libSceAgcDriver/Execution/src/VulkanDevice.cpp		stops requesting DenormPreserve for 32-bit floats, so the device keeps its default f32 denormal handling
APS5_NO_IMAGE_PAGE_CACHE	presence	core/libs/prx/libSceAgcDriver/Execution/src/GuestMemory.cpp		leaves the main guest image's page-protection table uncached
APS5_NO_INDIRECT_DRAW	presence	core/libs/prx/libSceAgcDriver/Execution/src/Pm4.cpp		rejects indirect draw packets at submission validation
APS5_NO_JOIN_WAW_CHECK	presence	core/libs/prx/libSceAgcDriver/Graphics/src/Recorder.cpp		lets a contiguous store join the pending one without the write-after-write check
APS5_NO_KEEP_NEW_STORAGE	presence	core/libs/prx/libSceAgcDriver/Graphics/src/ShaderResources.cpp		does not keep a newly created storage image alive with the open batch; such an image evicted before the batch completes can hand its memory to another image through the buffer pool while the batch still writes it
APS5_NO_KEYS_FILL_MEMO	presence	core/libs/prx/libSceAgcDriver/Graphics/src/DccMetadata.cpp		records no memo of GPU-side DCC key fills
APS5_NO_KEYS_REFILL	presence	core/libs/prx/libSceAgcDriver/Graphics/src/Texture.cpp		keeps a covered image's uploaded keys as they were instead of marking them uncompressed after a matching key fill
APS5_NO_KEYS_RESCAN	presence	core/libs/prx/libSceAgcDriver/Graphics/src/ShaderResources.cpp		reuses the DCC keys scanned before the lookup's refresh instead of discarding them for a rescan
APS5_NO_KEY_FAST_PATH	presence	core/libs/prx/libSceAgcDriver/Graphics/src/DccMetadata.cpp		scans a surface's DCC keys on every call instead of using and storing a proof
APS5_NO_LABEL_BATCHING	presence	core/libs/prx/libSceAgcDriver/Execution/src/Driver/Synchronization/DeferredLabels.cpp		turns off every label-batching behaviour, the label runs, queued labels and batched submit
APS5_NO_LABEL_BATCH_SUBMIT	presence	core/libs/prx/libSceAgcDriver/Execution/src/Driver/Synchronization/DeferredLabels.cpp		disables batched label submission
APS5_NO_LABEL_PACKET_FLUSH	presence	core/libs/prx/libSceAgcDriver/Execution/src/Driver/Packets/PacketFlush.cpp		disables the label flush between packets
APS5_NO_LABEL_RUNS	presence	core/libs/prx/libSceAgcDriver/Graphics/src/Recorder.cpp		records every store in a run of its own between a barrier pair
APS5_NO_LABEL_SHORTCUT	presence	core/libs/prx/libSceAgcDriver/Execution/src/Driver/Synchronization/WaitMemory.cpp		disables the pending-label table shortcut in a memory wait
APS5_NO_LAYER_CLEAR	presence	core/libs/prx/libSceAgcDriver/Graphics/src/Texture.cpp		tracks every surface as one layer instead of per array layer
APS5_NO_LAYER_REFUSAL_MEMO	presence	core/libs/prx/libSceAgcDriver/Graphics/src/Texture.cpp		re-checks a layer whose surface check failed at its current generation on every fill
APS5_NO_LAYOUT_CACHE	presence	core/libs/prx/libSceAgcDriver/Graphics/src/ShaderResources.cpp		builds descriptor set layouts fresh instead of through the descriptor cache
APS5_NO_LEASE_MIRROR	presence	core/libs/prx/libSceAgcDriver/Graphics/src/GuestBufferMemory.cpp		restores copies instead of image mirrors for address-based builds
APS5_NO_LIBRARY_PREBUILD	presence	core/libs/prx/libSceAgcDriver/Execution/src/Driver/Shaders/ShaderRegistry.cpp		disables prebuilding registration-form shader libraries (only 1 disables it)
APS5_NO_LOCK_PROFILE	presence	core/libs/prx/libSceAgcDriver/Execution/src/GuestMemory.cpp		turns off GpuMutex wait profiling under APS5_PROFILE_DRAW
APS5_NO_MOVED_BUFFER_TEMPLATES	presence	core/libs/prx/libSceAgcDriver/Graphics/src/Draw.cpp		keys the full descriptor words of read-only guest buffers instead of allowing a hit when they move
APS5_NO_OBSERVATION_GUARD	presence	core/libs/prx/libSceAgcDriver/Execution/src/Driver/Memory/WriteEvidence.cpp		observes every validated range, not only the ranges whose flush hook waited
APS5_NO_OPPORTUNISTIC_REAP	presence	core/libs/prx/libSceAgcDriver/Execution/src/VulkanDevice.cpp		skips retiring already-finished batches before a submit
APS5_NO_OWN_IMAGE_REFRESH	presence	core/libs/prx/libSceAgcDriver/Graphics/src/ShaderResources.cpp		lets a Pending failure of the fast proof take the full walk instead of the own-object refresh
APS5_NO_PARALLEL_MIRROR_REFRESH	presence	core/libs/prx/libSceAgcDriver/Graphics/src/GuestBufferMemory.cpp		disables the mirror-refresh thread pool, so mirror refreshes run on the calling thread
APS5_NO_PATTERN_FILL_HLE	presence	core/libs/prx/libSceAgcDriver/Execution/src/Driver/Memory/BufferFill.cpp		disables the dword-pattern fill HLE
APS5_NO_PAUSE_SPIN	presence	core/libs/prx/libSceAgcDriver/Execution/src/Driver/Synchronization/WaitMemory.cpp		disables the CPU pause-spin in a blocked memory wait
APS5_NO_PENDING_HTILE_FILLS	presence	core/libs/prx/libSceAgcDriver/Graphics/src/DepthSurface.cpp		reads the HTILE words of a depth target bind even when recorded fills are pending over them, instead of working the words out from the fills
APS5_NO_PERF_FRONTEND_PAIR	presence	core/libs/prx/libSceAgcDriver/tests/ShaderMemory.cpp		disables frontend-pair reuse in the shader-memory test
APS5_NO_PERF_LAYOUT_PAIR	presence	core/libs/prx/libSceAgcDriver/tests/ShaderMemory.cpp		disables layout-pair reuse in the shader-memory test
APS5_NO_PIPELINED_INDIRECT	presence	core/libs/prx/libSceAgcDriver/Execution/src/Driver/Draw/Draw.cpp		drains the draw pipeline at every CPU-path indirect draw instead of queuing it behind the pipeline
APS5_NO_PIPELINE_CACHE	presence	core/libs/prx/libSceAgcDriver/Graphics/src/Pipeline.cpp		builds a pipeline per draw instead of reusing cached ones
APS5_NO_PIPELINE_JOURNAL	presence	core/libs/prx/libSceAgcDriver/Graphics/src/PipelineJournal.cpp		keeps the pipeline journal off; 0 leaves it on, and it is also off while the shader disk cache is disabled
APS5_NO_PIPELINE_LTO	presence	core/libs/prx/libSceAgcDriver/Graphics/src/PipelineLibrary.cpp		disables background link-time optimization of graphics pipeline libraries
APS5_NO_POOL_CACHE	presence	core/libs/prx/libSceAgcDriver/Graphics/src/ShaderResources.cpp		allocates descriptor pools fresh instead of through the descriptor cache
APS5_NO_PRECOLLECT	presence	core/libs/prx/libSceAgcDriver/Graphics/src/ShaderResources.cpp		disables the stage-A precollect pass
APS5_NO_PREPARED_FIND	presence	core/libs/prx/libSceAgcDriver/Execution/src/VulkanDevice.cpp		looks the resource-cache key up again instead of using the object stage A found
APS5_NO_PRESENT_EPOCH	presence	core/libs/prx/libSceAgcDriver/Execution/src/VulkanDevice.cpp		keeps the presenter at collect epoch 0 instead of adopting the newest epoch a worker started
APS5_NO_PRESYNC	presence	core/libs/prx/libSceAgcDriver/Execution/src/VulkanDevice.cpp		disables the stage-A pre-sync of pending writes
APS5_NO_PROC_TABLE	presence	core/libs/prx/libSceAgcDriver/Execution/src/VulkanDevice.cpp		resolves the device's Vulkan entry points per call instead of filling the function table once
APS5_NO_QUEUED_DRAW_READ_ORDER	presence	core/libs/prx/libSceAgcDriver/Execution/src/Driver/Draw/DrawPipeline.cpp		stops queued draws and DMA copies from holding back write-backs over the guest memory they read at commit
APS5_NO_QUEUED_LABELS	presence	core/libs/prx/libSceAgcDriver/Execution/src/Driver/Synchronization/DeferredLabels.cpp		disables the queued-label table
APS5_NO_RECORDED_CLEAR	presence	core/libs/prx/libSceAgcDriver/Graphics/src/Texture.cpp		waits for a storage image's fast clear in a batch of its own instead of recording it
APS5_NO_RECORDER_SYNC_DRAWS	presence	core/libs/prx/libSceAgcDriver/Graphics/src/Draw.cpp		gives the draws the recorder waits for their own batch instead of recording them with the wait
APS5_NO_RECORD_COPIED_DRAWS	presence	core/libs/prx/libSceAgcDriver/Graphics/src/Draw.cpp		waits for copied-write draws at once instead of listing their write-back for later
APS5_NO_RECT_INDIRECT	presence	core/libs/prx/libSceAgcDriver/Execution/src/Driver/Draw/IndirectDraw.cpp		sends an indirect rectangle-list draw down the CPU path instead of the GPU-side path
APS5_NO_RESIDENT_LINEAR	presence	core/libs/prx/libSceAgcDriver/Graphics/src/Draw.cpp		drops the residency of a linear color target whose address already has a host import, so it is copied in and out again
APS5_NO_RESIDENT_PRESENT	presence	core/libs/prx/libSceAgcDriver/Execution/src/VulkanDevice.cpp		always presents through guest memory instead of the resident image
APS5_NO_RESIDENT_TARGETS	presence	core/libs/prx/libSceAgcDriver/Graphics/src/Draw.cpp		copies every render target in and out again instead of keeping it resident
APS5_NO_RESOURCE_CACHE	presence	core/libs/prx/libSceAgcDriver/Execution/src/VulkanDevice.cpp		disables the shared resource cache, so every dispatch builds its resources fresh
APS5_NO_SAMPLED_FROM_STORAGE	presence	core/libs/prx/libSceAgcDriver/Graphics/src/ShaderResources.cpp		keeps every sampled texture a CPU snapshot instead of viewing the surface's storage image
APS5_NO_SAMPLER_CACHE	presence	core/libs/prx/libSceAgcDriver/Graphics/src/ShaderResources.cpp		creates a sampler per binding instead of caching them
APS5_NO_SEPARATE_QUEUED_LABELS	presence	core/libs/prx/libSceAgcDriver/Graphics/src/Recorder.cpp		lets a queued label enter the recorded table and replace a recorded entry of its dword
APS5_NO_SINGLE_PASS_COLLECT	presence	core/libs/prx/libSceAgcDriver/Execution/src/GuestMemory.cpp		collects written pages in one cleared pass instead of a second confirming walk
APS5_NO_SITE_SAMPLING	presence	core/libs/prx/libSceAgcDriver/Execution/src/Driver/Memory/WriteEvidence.cpp		disables per-site sampling of write evidence
APS5_NO_SNAPSHOT_CHECK	presence	core/libs/prx/libSceAgcDriver/Graphics/src/GuestBufferMemory.cpp		skips the live compare that proves a captured snapshot still equals the guest bytes
APS5_NO_SORTED_SNAPSHOT_LOOKUP	presence	core/libs/prx/libSceAgcDriver/Graphics/src/GuestBufferMemory.cpp		scans the snapshot's regions in AddSnapshot instead of searching them
APS5_NO_SPACE_STAMP_ONCE	presence	core/libs/prx/libSceAgcDriver/Graphics/src/GuestBufferMemory.cpp		stamps every direct import of a cached address space on each hit instead of once per space
APS5_NO_SPARSE_HEAP_MIRROR	presence	core/libs/prx/libSceAgcDriver/Graphics/src/GuestBufferMemory.cpp		disables sparse heap mirrors, mirroring only whole heap ranges
APS5_NO_STAGE_A_IMAGES	presence	core/libs/prx/libSceAgcDriver/Graphics/src/ShaderResources.cpp		leaves every image element to the stage-B full lookup instead of a stage-A record
APS5_NO_STORAGE_DEDUPE	presence	core/libs/prx/libSceAgcDriver/Graphics/src/ShaderResources.cpp		looks each identical consecutive storage descriptor up, and refreshes its image, separately
APS5_NO_SYNC_SITES	presence	core/libs/prx/libSceAgcDriver/Graphics/src/Recorder.cpp		leaves only the per-source and per-thread sync counts, with no per-site attribution
APS5_NO_SYNC_THROUGH	presence	core/libs/prx/libSceAgcDriver/Graphics/src/Recorder.cpp		makes every pending-write sync a full Sync instead of the targeted unlocked wait
APS5_NO_TABLE_REPLAY	presence	core/libs/prx/libSceAgcDriver/Graphics/src/ShaderResources.cpp		makes a reused bound image table rerun each element's checks instead of replaying its recorded bind outcomes
APS5_NO_TABLE_REUSE	presence	core/libs/prx/libSceAgcDriver/Graphics/src/ShaderResources.cpp		disables reusing a materialized image table
APS5_NO_TARGET_WRITEBACK	presence	core/libs/prx/libSceAgcDriver/Graphics/src/Draw.cpp		skips writing render targets back to guest memory after the draw
APS5_NO_TEMPLATE_DATA_REFRESH	presence	core/libs/prx/libSceAgcDriver/Execution/src/VulkanDevice.cpp		keys the ShaderData and FlattenedSrt words into compute keys instead of refreshing a template's data buffers
APS5_NO_TEXTURE_CACHE	presence	core/libs/prx/libSceAgcDriver/Execution/src/VulkanDevice.cpp		builds every texture and storage-image object fresh instead of caching it
APS5_NO_TEXTURE_HASH	presence	core/libs/prx/libSceAgcDriver/Graphics/src/ShaderResources.cpp		finds texture-cache entries by scanning the list instead of the hash index
APS5_NO_TIMELINE	presence	core/libs/prx/libSceAgcDriver/Execution/src/VulkanDevice.cpp		leaves timeline semaphores off, so drains wait under the mutex
APS5_NO_UNIT_SHADOW	presence	core/libs/prx/libSceAgcDriver/Graphics/src/UnitShadow.cpp		makes every unit shadow primitive inert
APS5_NO_UNLOCKED_DEVICE	presence	core/libs/prx/libSceAgcDriver/Execution/src/Driver/Dispatch/Dispatch.cpp		loads the device and runs a dispatch under the GPU mutex instead of using an unlocked device
APS5_NO_UNLOCKED_DRAIN	presence	core/libs/prx/libSceAgcDriver/Execution/src/Driver/Packets/PacketMemory.cpp		keeps the GPU mutex held while draining a packet instead of the unlocked drain
APS5_NO_UNLOCKED_VALIDATE	presence	core/libs/prx/libSceAgcDriver/Execution/src/Driver/Dispatch/DispatchLookup.cpp		keeps the dispatch-cache mutex held while validating a cache entry
APS5_NO_UNREGISTERED_DROP	presence	core/libs/prx/libSceAgcDriver/Graphics/src/Texture.cpp		stores a storage image's pending results even when its memory is no longer registered, instead of dropping them
APS5_NO_VALIDATE_CACHE	presence	core/libs/prx/libSceAgcDriver/Graphics/src/Draw.cpp		validates every draw instead of serving the validation memo
APS5_NO_VALIDATE_SKIP	presence	core/libs/prx/libSceAgcDriver/Execution/src/Driver/Memory/WriteEvidence.cpp		syncs every pending write instead of skipping validation on written-element evidence
APS5_NO_VOUCHED_VALIDATION	presence	core/libs/prx/libSceAgcDriver/Execution/src/Driver/Dispatch/VariantValidation.cpp		validates a cached draw or dispatch through its per-run checks instead of its table block's vouch
APS5_NO_WAIT_OVERLAP_SUBMIT	presence	core/libs/prx/libSceAgcDriver/Execution/src/Driver/Synchronization/WaitMemory.cpp		disables the unlocked wait that submits recorded work while overlapping writes are open
APS5_NO_WORDWISE_CAPTURE	presence	core/libs/prx/libSceAgcDriver/Execution/src/ShaderMemory.cpp		fetches every captured page whole instead of reading pending pages word by word
APS5_NO_WORKER_AFFINITY	presence	core/libs/prx/libSceAgcDriver/Execution/src/Driver/Queues/WorkerAffinity.cpp		leaves the queue workers and presenter unpinned (affinity mask 0)
APS5_NO_WRITER_KEYED_EVIDENCE	presence	core/libs/prx/libSceAgcDriver/Execution/src/Driver/Memory/WriteEvidence.cpp		records written-element evidence without keying it by writer
APS5_NO_WRITE_EVIDENCE	presence	core/libs/prx/libSceAgcDriver/Execution/src/Driver/Memory/WriteEvidence.cpp		disables the written-element evidence that lets a read skip the flush hook
APS5_NO_WRITE_ONLY_PASSES	presence	core/libs/prx/libSceAgcDriver/Graphics/src/Draw.cpp		counts a buffer a draw only stores to as read and written, so the next draw storing to it begins a new render pass
APS5_NO_WRITTEN_SHADOW	presence	core/libs/prx/libSceAgcDriver/Graphics/src/GuestBufferMemory.cpp		binds written staging elements in place instead of through the written shadow
APS5_PACKET_EPOCH	presence	core/libs/prx/libSceAgcDriver/Execution/src/Driver/Synchronization/WaitStatistics.cpp		makes queue workers bump the collect epoch before every packet
APS5_PIPELINED_DRAWS	integer	core/libs/prx/libSceAgcDriver/Execution/src/Driver/Draw/DrawPipeline.cpp	8	depth of the queue-0 draw pipeline: 0 records draws inline, 1 selects the default depth of 8, and any other value is the depth capped at 256
APS5_PIPELINE_DISPATCH	presence	core/libs/prx/libSceAgcDriver/Execution/src/Driver/Dispatch/Dispatch.cpp		records an eligible dispatch as an item in the queue-0 pipeline instead of running it inline
APS5_PIPELINE_DMA	presence	core/libs/prx/libSceAgcDriver/Execution/src/Driver/Packets/PacketMemory.cpp		records a queue-0 DMA data packet in the draw pipeline instead of executing it inline
APS5_PIPELINE_DRAIN_FILLS	presence	core/libs/prx/libSceAgcDriver/Execution/src/Driver/Memory/BufferFill.cpp		runs queue-0 buffer fills through the draw pipeline to keep their order; any value stores them in place instead
APS5_PIPELINE_DRAIN_LABELS	presence	core/libs/prx/libSceAgcDriver/Execution/src/Driver/Packets/PacketExecution.cpp		answers a queue-0 wait from the draw pipeline's own pending labels while it is busy; any value drains the pipeline at the wait
APS5_PIPELINE_DRAIN_SUBMISSIONS	presence	core/libs/prx/libSceAgcDriver/Execution/src/Driver/Packets/PacketExecution.cpp		lets a submission end without submitting its recorded work, so it joins a later submit with the pipeline's queue; any value drains the pipeline and submits at the end of the submission
APS5_PIPELINE_JOURNAL_ENTRIES	integer	core/libs/prx/libSceAgcDriver/Graphics/src/PipelineJournal.cpp	1024	number of journal entries loaded at most; a longer journal is truncated to this count
APS5_PIPELINE_JOURNAL_MAX_MIB	integer	core/libs/prx/libSceAgcDriver/Graphics/src/PipelineJournal.cpp	1024	size cap in MiB on the pipeline journal; later pipelines are not recorded once it is reached
APS5_PIPELINE_JOURNAL_THREADS	integer	core/libs/prx/libSceAgcDriver/Graphics/src/PipelineJournal.cpp		threads that replay the journal's pipelines, 0 disables the replay; unset uses half the hardware threads, at least one
APS5_PIPELINE_STATS	string	core/libs/prx/libSceAgcDriver/Execution/src/VulkanDevice.cpp		captures and prints driver statistics for each new pipeline, requiring VK_KHR_pipeline_executable_properties; 1 enables
APS5_PIPELINE_STORE_ENTRIES	integer	core/libs/prx/libSceAgcDriver/Graphics/src/Pipeline.cpp	1024	LRU bound on the entries of the pipeline store; an entry is evicted only once no recorded draw holds it
APS5_PM4_RING	integer	core/libs/prx/libSceAgcDriver/Execution/src/Driver/Queues/SubmissionRing.cpp	16	number of submissions the PM4 ring keeps for dumps, from 0 to 4096; 0 disables the ring
APS5_POLL_REAP	presence	core/libs/prx/libSceAgcDriver/Execution/src/Driver/Synchronization/WaitMemory.cpp		reaps completions on every poll tick
APS5_POLL_TRY_EACH	presence	core/libs/prx/libSceAgcDriver/Execution/src/Driver/Synchronization/WaitMemory.cpp		tries the poll service on every tick instead of at the poll interval
APS5_PREBUILD_THREADS	integer	core/libs/prx/libSceAgcDriver/Execution/src/Driver/Shaders/ShaderRegistry.cpp	1	number of threads in the one-thread pool that prebuilds registration-form shader libraries
APS5_PRESENT_POLL_FENCE	presence	core/libs/prx/libSceAgcDriver/Execution/src/VulkanDevice.cpp		polls the present fence every 50 us instead of blocking on it
APS5_PROBE_DISPATCH	string	core/libs/prx/libSceAgcDriver/Execution/src/Driver/Dispatch/Dispatch.cpp		probes one dispatch at a hex program address (optionally :index) with debug output
APS5_PROFILE_DRAW	presence	core/libs/prx/libSceAgcDriver/Execution/src/Driver/Dispatch/Dispatch.cpp		prints the per-phase driver and draw timing totals
APS5_PROFILE_GPU	presence	core/libs/prx/libSceAgcDriver/Graphics/src/Recorder.cpp		records the GPU time of recorded work by key from batch timestamps
APS5_REAP_EACH_PACKET	presence	core/libs/prx/libSceAgcDriver/Execution/src/Driver/Synchronization/DeferredLabels.cpp		reaps deferred labels after every packet
APS5_RELEASE_ON_UNLOCK	presence	core/libs/prx/libSceAgcDriver/Graphics/src/Recorder.cpp		destroys deferred batches on the unlocking thread instead of a release thread
APS5_RELEASE_QUEUE_MAX	integer	core/libs/prx/libSceAgcDriver/Graphics/src/Recorder.cpp	256	maximum deferred batches handed to the release thread before a thread destroys its own
APS5_RELEASE_UNDER_LOCK	presence	core/libs/prx/libSceAgcDriver/Graphics/src/Recorder.cpp		destroys deferred batches in finish() under the lock instead of on a release thread
APS5_RESOURCE_CACHE_ENTRIES	integer	core/libs/prx/libSceAgcDriver/Graphics/src/ShaderResources.cpp	1024	maximum entries in the per-device shader-resource cache
APS5_RESOURCE_CACHE_TABLE_ELEMENTS	integer	core/libs/prx/libSceAgcDriver/Graphics/src/ShaderResources.cpp	65536	number of image-table elements the shader-resource cache keeps (0 keeps the default)
APS5_SAMPLE_COMMITTER	string	core/libs/prx/libSceAgcDriver/Execution/src/WorkerSampler.cpp		samples queue 0's draw committer every millisecond and writes its stacks to a file (Windows only)
APS5_SAMPLE_THREADS	string	core/libs/prx/libSceAgcDriver/Execution/src/WorkerSampler.cpp		samples every thread of the process every 2 ms and writes each busy thread's hottest frames to a file (Windows only)
APS5_SAMPLE_WORKER	string	core/libs/prx/libSceAgcDriver/Execution/src/WorkerSampler.cpp		samples the calling queue worker every millisecond and writes its stacks to a file (Windows only)
APS5_SHADOW_VERIFY	presence	core/libs/prx/libSceAgcDriver/Graphics/src/UnitShadow.cpp		verifies that the shadow-published bytes equal the guest bytes
APS5_SHARED_COLLECT_MEMO	presence	core/libs/prx/libSceAgcDriver/Execution/src/GuestMemory.cpp		uses the tracker's shared collect-memo ring under its mutex instead of the per-thread ring
APS5_SHARED_MEMORY_COUNTERS	presence	core/libs/prx/libSceAgcDriver/Execution/src/GuestMemory.cpp		accumulates the memory counters into one shared set with atomic RMWs
APS5_SKIP_INVALID_INDIRECT_REGISTERS	presence	core/libs/prx/libSceAgcDriver/Execution/src/Driver/Packets/PacketExecution.cpp		skips the entries whose offset dword is no register offset in a SET_*_REG_INDIRECT list read again where its packet executes, and writes its valid entries, instead of stopping
APS5_SKIP_REFUSED_DRAWS	presence	core/libs/prx/libSceAgcDriver/Execution/src/Driver/Diagnostics.cpp		skips a draw whose graphics stage the recompiler refuses, reporting each refused shader once, instead of failing the queue
APS5_SKIP_REQUEST_DUMPS	integer	core/libs/prx/libSceAgcDriver/Execution/src/Driver/Shaders/ShaderDiagnostics.cpp	8	caps RecompileRequest dumps per shader-skip reason
APS5_SKIP_SHADER	integer	core/libs/prx/libSceAgcDriver/Execution/src/Driver/Diagnostics.cpp	0	hex program address of a shader to skip preparing
APS5_SNAPSHOT_CHECK_FATAL	presence	core/libs/prx/libSceAgcDriver/Graphics/src/GuestBufferMemory.cpp		throws on the first snapshot that differs from registered memory instead of only counting them
APS5_SPANNED_WAVE	presence	core/libs/prx/libSceAgcDriver/Execution/src/VulkanDevice.cpp		spans a single-wave wave64 compute program over two full 32-lane subgroups
APS5_SPIRV_OPT	string	core/libs/prx/libSceAgcDriver/tests/SpirvOptimizer.cpp		selects the SPIR-V optimization pipeline: full restores the performance pipeline, none skips optimization
APS5_SRGB_ATTACHMENT_PROXY	string	core/libs/prx/libSceAgcDriver/Graphics/src/Texture.cpp		forces the sRGB color-attachment proxy format; 1 forces it
APS5_SRGB_SHADER_DECODE	presence	core/libs/prx/libSceAgcDriver/Graphics/src/TextureFormat.cpp		1 forces shader sRGB decode for the sRGB 8-bit formats; copy does the same, except that 8 and 8_8 sRGB sample an RGBA8 sRGB copy when the device can blit to RGBA8
APS5_STAGING_POOL_MIB	integer	core/libs/prx/libSceAgcDriver/Graphics/src/BufferPool.cpp		size in MiB of the device-local staging buffer pool; 0 disables the device tier
APS5_STALE_EVICT	presence	core/libs/prx/libSceAgcDriver/Graphics/src/Texture.cpp		also evicts a proved-present dead image's pending layers on an access
APS5_STALL_WATCHDOG_SECS	integer	core/libs/prx/libSceAgcDriver/Execution/src/Driver/Queues/SubmissionRing.cpp	0	seconds a queue may go without progress before the submission rings are dumped; 0 disables the watchdog, at most 86400
APS5_STORE_AT_FLIP	string	core/libs/prx/libSceAgcDriver/Execution/src/VulkanDevice.cpp		stores every pending storage-image result at each display flip; only 1 is accepted
APS5_STORE_AT_LABEL	presence	core/libs/prx/libSceAgcDriver/Execution/src/VulkanDevice.cpp		flushes every pending storage-image result when a label is written on the GPU
APS5_SUBMISSION_EPOCH_EACH	presence	core/libs/prx/libSceAgcDriver/Execution/src/Driver/Packets/PacketExecution.cpp		bumps the guest-memory collect epoch on every submission instead of only when the submission is newer than the last bump
APS5_SUBMIT_AT_END	presence	core/libs/prx/libSceAgcDriver/Execution/src/Driver/Packets/PacketExecution.cpp		submits recorded work at the end of a submission even when a worker is queued
APS5_SUSPEND_DRAIN	presence	core/libs/prx/libSceAgcDriver/Execution/src/Driver/Packets/PacketExecution.cpp		drains the device at a suspend point
APS5_SYNC_COMPLETION_DRAWS	presence	core/libs/prx/libSceAgcDriver/Graphics/src/Draw.cpp		keeps every draw with completion work synchronous
APS5_SYNC_DISPATCH	presence	core/libs/prx/libSceAgcDriver/Execution/src/VulkanDevice.cpp		waits for every dispatch instead of batching it
APS5_SYNC_DRAWS	presence	core/libs/prx/libSceAgcDriver/Graphics/src/Draw.cpp		makes every draw synchronous
APS5_SYNC_FLIP	presence	core/libs/prx/libSceAgcDriver/Execution/src/Driver/Packets/PacketMemory.cpp		makes flips synchronous (no frames in flight)
APS5_SYNC_LEASE_DISPATCH	presence	core/libs/prx/libSceAgcDriver/Graphics/src/GuestBufferMemory.cpp		waits for leased draws and dispatches at once instead of deferring them
APS5_SYNC_SAMPLE_DUMPS	presence	core/libs/prx/libSceAgcDriver/Execution/src/VulkanDevice.cpp		takes the CPU path for occlusion-counter sample dumps, as APS5_NO_GPU_SAMPLE_DUMPS does
APS5_SYNC_SHADER_PREPARE	presence	core/libs/prx/libSceAgcDriver/Execution/src/Driver/Shaders/ShaderRegistry.cpp		prepares registered shaders on the registering thread instead of in the background
APS5_SYNC_TEXTURE_UPLOAD	presence	core/libs/prx/libSceAgcDriver/Graphics/src/Texture.cpp		restores the waiting batch for every texture upload
APS5_TEMPLATE_AUDIT_MS	integer	core/libs/prx/libSceAgcDriver/Graphics/src/ShaderResources.cpp	1000	milliseconds after which a cached shader-resource object takes the full walk instead of the fast proof; 0 turns the audit off
APS5_TEXTURE_CACHE_MIB	integer	core/libs/prx/libSceAgcDriver/Graphics/src/ShaderResources.cpp	0	forced byte budget in MiB for the storage and sampled texture caches; 0 selects the budget automatically
APS5_THREAD_NICE	integer	core/libs/prx/libSceAgcDriver/Execution/src/ThreadPriority.cpp	-10	nice value the raised queue workers and presenter ask rtkit for (clamped to -20..19)
APS5_THREAD_PRIORITY	string	core/libs/prx/libSceAgcDriver/Execution/src/ThreadPriority.cpp	off	raises the queue workers and presenter above the title's threads: off, high or rt
APS5_THREAD_RT	integer	core/libs/prx/libSceAgcDriver/Execution/src/ThreadPriority.cpp	10	realtime priority the raised threads ask rtkit for (clamped to 1..99)
APS5_THREAD_RTTIME_US	integer	core/libs/prx/libSceAgcDriver/Execution/src/ThreadPriority.cpp	200000	hard RLIMIT_RTTIME the realtime plan lowers to, in microseconds
APS5_TRACE_BARRIERS	integer	core/libs/prx/libSceAgcDriver/Graphics/src/Recorder.cpp		prints that many simulated hazard barriers before stopping
APS5_TRACE_BENCH	string	core/libs/prx/libSceAgcDriver/Execution/include/PerformanceTrace.hpp		1 with APS5_TRACE_SUMMARY_EVERY also rewrites aps5-bench-<commit>-<pid>.json.gz with every interval's frame times; any other value fails the first present
APS5_TRACE_BIGBUF	presence	core/libs/prx/libSceAgcDriver/Graphics/src/GuestBufferMemory.cpp		lists every copied guest region of at least 1 MiB
APS5_TRACE_CAPSYNC	presence	core/libs/prx/libSceAgcDriver/Execution/src/Driver/Memory/WriteEvidence.cpp		prints every unlocked hook wait
APS5_TRACE_CONTEXT_STATE	presence	core/libs/prx/libSceAgcDriver/Execution/src/Pm4.cpp		logs context-state operations and the context registers written
APS5_TRACE_COPY_HLE	presence	core/libs/prx/libSceAgcDriver/Execution/src/Driver/Memory/BufferCopy.cpp		traces the CPU copy HLE's refusals
APS5_TRACE_DCC_KEYS	presence	core/libs/prx/libSceAgcDriver/Execution/src/DisplayBuffer.cpp		traces stores of uncompressed DCC keys
APS5_TRACE_DIR	string	core/libs/prx/libSceAgcDriver/Execution/src/PerformanceTrace.cpp	.	directory for the trace and summary files
APS5_TRACE_DISPATCH_CACHE	presence	core/libs/prx/libSceAgcDriver/Execution/src/Driver/Dispatch/Dispatch.cpp		reports dispatch-cache entries whose key changed
APS5_TRACE_DISPATCH_IO	string	core/libs/prx/libSceAgcDriver/Execution/src/Driver/Dispatch/Dispatch.cpp		prints every dispatch's resources; 2 also lists the small-buffer words
APS5_TRACE_DRAWS	presence	core/libs/prx/libSceAgcDriver/Execution/src/Driver/Packets/PacketExecution.cpp		prints one line per draw with its phases and inputs
APS5_TRACE_DRAW_CACHE	presence	core/libs/prx/libSceAgcDriver/Execution/src/Driver/Draw/DrawCacheStatistics.cpp		reports draw-cache entries whose key changed
APS5_TRACE_DRAW_PROGRAMS	presence	core/libs/prx/libSceAgcDriver/Execution/src/Driver/Draw/Draw.cpp		prints one line per recorded draw with the stage, role and code address of each program
APS5_TRACE_EVENTS	integer	core/libs/prx/libSceAgcDriver/Execution/src/PerformanceTrace.cpp		number of events a windowed trace keeps in its buffer
APS5_TRACE_FILL_COVER	presence	core/libs/prx/libSceAgcDriver/Graphics/src/Texture.cpp		prints the first fills that meet images without covering them
APS5_TRACE_FLUSH	presence	core/libs/prx/libSceAgcDriver/Graphics/src/Texture.cpp		names what forces pending storage-image results into guest memory
APS5_TRACE_FRAMES	string	core/libs/prx/libSceAgcDriver/Execution/include/PerformanceTrace.hpp		records <count> presents from present <start> (whole numbers, start:count)
APS5_TRACE_GPU	presence	core/libs/prx/libSceAgcDriver/Execution/src/DisplayBuffer.cpp		prints GPU wait, submission and scanout events, the execute lines, and waits longer than 200 ms (read in DisplayBuffer, PacketExecution, QueueWorker, Submission and WaitMemory)
APS5_TRACE_HOST_IMPORT	presence	core/libs/prx/libSceAgcDriver/Graphics/src/GuestBufferMemory.cpp		traces host-import mapping failures
APS5_TRACE_INDIRECT_DRAWS	presence	core/libs/prx/libSceAgcDriver/Execution/src/Driver/Draw/DrawPrecheck.cpp		prints indirect draw packets that reach the precheck
APS5_TRACE_LATE_LABELS	presence	core/libs/prx/libSceAgcDriver/Execution/src/Driver/Synchronization/WaitMemory.cpp		prints how late labels were trusted or refused
APS5_TRACE_PREPARATION_WAIT	presence	core/libs/prx/libSceAgcDriver/Execution/src/Driver/Shaders/ShaderRegistry.cpp		prints where a demanded shader preparation sat in the queue
APS5_TRACE_QUERY	presence	core/libs/prx/libSceAgcDriver/Execution/src/GuestMemory.cpp		prints every 16th page query with its cost
APS5_TRACE_RESOURCE_KEY_MISSES	presence	core/libs/prx/libSceAgcDriver/Graphics/src/ShaderResources.cpp		reports shader-resource cache key misses
APS5_TRACE_SECONDS	string	core/libs/prx/libSceAgcDriver/Execution/include/PerformanceTrace.hpp		records from the first present <start> seconds after the first present for <length> seconds (whole numbers, start:length)
APS5_TRACE_SHADER_REGS	string	core/libs/prx/libSceAgcDriver/Execution/src/Driver/Shaders/ShaderRegistry.cpp		prints a shader's registers for one hex program address, or all
APS5_TRACE_SHADOW	integer	core/libs/prx/libSceAgcDriver/Graphics/src/UnitShadow.cpp		prints the first n shadow publishes, the first 64 seeds, every evict and retire
APS5_TRACE_SKIPS	presence	core/libs/prx/libSceAgcDriver/Execution/src/Driver/Shaders/ShaderDiagnostics.cpp		reports each distinct shader skip once
APS5_TRACE_STAGING	presence	core/libs/prx/libSceAgcDriver/Graphics/src/GuestBufferMemory.cpp		prints every distinct staged range once, with its size and whether an atomic element lies in it
APS5_TRACE_SUMMARY_EVERY	integer	core/libs/prx/libSceAgcDriver/Execution/include/PerformanceTrace.hpp		seconds between periodic trace summaries, emitted on the next present
APS5_TRACE_TEXTURES	presence	core/libs/prx/libSceAgcDriver/Graphics/src/ShaderResources.cpp		prints a sampled texture's bytes and how many are nonzero
APS5_TRACE_UNREADABLE	presence	core/libs/prx/libSceAgcDriver/Execution/src/GuestMemory.cpp		names the code that checked an inaccessible range
APS5_TRACE_UPLOAD	presence	core/libs/prx/libSceAgcDriver/Graphics/src/Texture.cpp		names why a storage image is uploaded again
APS5_UNIT_SHADOW_MIB	integer	core/libs/prx/libSceAgcDriver/Graphics/src/UnitShadow.cpp	1024	budget in MiB of the unit shadow's slabs
APS5_UNIT_SHADOW_SLAB_MIB	integer	core/libs/prx/libSceAgcDriver/Graphics/src/UnitShadow.cpp	8	size in MiB of one unit-shadow slab
APS5_VALIDATE_LEGACY	presence	core/libs/prx/libSceAgcDriver/Execution/src/Driver/Memory/WriteEvidence.cpp		uses the legacy snapshot and full-compare path for pending-write validation
APS5_VALIDATE_SYNC_SKIP	presence	core/libs/prx/libSceAgcDriver/Execution/src/Driver/Memory/WriteEvidence.cpp		reads skipped dwords through the hook too and counts a mismatch
APS5_VERIFY_BDA_TABLES	presence	core/libs/prx/libSceAgcDriver/Graphics/src/GuestBufferMemory.cpp		checks a reused BDA page table against the one just built
APS5_VERIFY_DATA_HITS	presence	core/libs/prx/libSceAgcDriver/Execution/src/Driver/Dispatch/DispatchCache.cpp		recomputes a dispatch data hit and aborts on a disagreement
APS5_VERIFY_DRAW_DATA_HITS	presence	core/libs/prx/libSceAgcDriver/Execution/src/Driver/Draw/DrawCache.cpp		recomputes a draw data hit and aborts on a disagreement
APS5_VERIFY_DRAW_ENTRIES	presence	core/libs/prx/libSceAgcDriver/Execution/src/Driver/Draw/DrawCache.cpp		re-decodes every draw-cache hit through the full path and checks it
APS5_VERIFY_DRAW_RECIPE	presence	core/libs/prx/libSceAgcDriver/Execution/src/Driver/Draw/DrawCache.cpp		re-decodes a draw recipe instead of using the cache and checks the result
APS5_VERIFY_DRAW_SNAPSHOTS	presence	core/libs/prx/libSceAgcDriver/Graphics/src/ShaderResources.cpp		re-reads a reused draw snapshot and counts the ones that differ
APS5_VERIFY_KNOWN_VALUES	presence	core/libs/prx/libSceAgcDriver/Execution/src/Driver/Memory/BufferCopy.cpp		serves known values through the hook too and counts mismatches
APS5_VERIFY_PROOFS	presence	core/libs/prx/libSceAgcDriver/Graphics/src/ShaderResources.cpp		runs the full walk beside every own-object refresh proof and aborts on a different decision
APS5_VERIFY_RECIPE	presence	core/libs/prx/libSceAgcDriver/Execution/src/VulkanDevice.cpp		makes the ordinary prepare path beside a recipe hit and compares the answers
APS5_VERIFY_REGISTER_KEYS	presence	core/libs/prx/libSceAgcDriver/Execution/src/Driver/Draw/DrawRegisterKey.cpp		checks the incremental register fingerprint against the full state
APS5_VERIFY_SET_LAYOUTS	presence	core/libs/prx/libSceAgcDriver/Graphics/src/Draw.cpp		checks the draw's set layouts against those described from the shaders
APS5_WAIT_LOCK	presence	core/libs/prx/libSceAgcDriver/Execution/src/Driver/Synchronization/WaitMemory.cpp		always takes the GPU mutex on a wait (no unlocked label lookup)
APS5_WAIT_LOCK_ENTRY	presence	core/libs/prx/libSceAgcDriver/Execution/src/Driver/Synchronization/WaitMemory.cpp		takes the GPU mutex blocking on the entry instead of trying it
APS5_WATCH_MEMORY	string	core/libs/prx/libSceAgcDriver/Execution/src/VulkanDevice.cpp		watches named guest ranges (<hex address>:<hex bytes>[,...]) across dispatches; needs APS5_SYNC_DISPATCH=1
APS5_WORKER_AFFINITY	presence	core/libs/prx/libSceAgcDriver/Execution/src/Driver/Queues/WorkerAffinity.cpp		pins the queue workers and presenter to the performant cores (needed to enable it on non-Linux hosts)
APS5_WORKER_AFFINITY_MASK	integer	core/libs/prx/libSceAgcDriver/Execution/src/Driver/Queues/WorkerAffinity.cpp		explicit CPU affinity mask for the queue workers and presenter
APS5_WORKING_SET_MIB	integer	core/libs/prx/libSceAgcDriver/Execution/src/VulkanDevice.cpp		sets the process working-set minimum to this many MiB (Windows only)
APS5_WRITE_EVIDENCE_AFTER	integer	core/libs/prx/libSceAgcDriver/Execution/src/Driver/Memory/WriteEvidence.cpp	4	consecutive unchanged observations before a dword's written-element evidence is trusted
APS5_WRITE_EVIDENCE_MAX_KIB	integer	core/libs/prx/libSceAgcDriver/Execution/src/Driver/Memory/WriteEvidence.cpp	16	largest written range in KiB that the evidence may serve
APS5_WRITE_EVIDENCE_SAMPLE	integer	core/libs/prx/libSceAgcDriver/Execution/src/Driver/Memory/WriteEvidence.cpp	16	reads one in this many pending writes to sample them
APS5_WRITE_WATCH_IMPORTS	string	core/libs/prx/libSceAgcDriver/Graphics/src/GuestBufferMemory.cpp		how host imports are write-tracked: probe, watch or unwatch
APS5_WRITTEN_SHADOW_MAX_KIB	integer	core/libs/prx/libSceAgcDriver/Graphics/src/GuestBufferMemory.cpp	2048	upper bound in KiB of the written staging window
APS5_WRITTEN_SHADOW_MIN_KIB	integer	core/libs/prx/libSceAgcDriver/Graphics/src/GuestBufferMemory.cpp	16	lower bound in KiB of the written staging window

libSceAjm.native

Name	Kind	Source	Default	Description
APS5_TRACE_AJM	presence	core/libs/prx/libSceAjm.native/src/Ajm.cpp		logs every job the title submits and what the library writes back

libSceAudioOut

Name	Kind	Source	Default	Description
APS5_TRACE_AUDIOOUT	presence	core/libs/prx/libSceAudioOut/src/AudioOut.cpp		logs each audio-out port opened with its type, sample count, frequency and format
APS5_TRACE_AUDIOOUT2	presence	core/libs/prx/libSceAudioOut/src/AudioOut2Context.cpp		prints the title's AudioOut2 context and port parameters and stream activity

libSceFiber

Name	Kind	Source	Default	Description
APS5_TRACE_FIBER	presence	core/libs/prx/libSceFiber/Export.cpp		logs fiber initialization, switches and returns to the thread

libSceFont

Name	Kind	Source	Default	Description
ANYPS5_SYSTEM_FONTS	string	core/libs/prx/libSceFont/src/SystemFont.cpp		directory holding the system fonts, defaulting to anyps5-fonts beside the executable

libSceNpEntitlementAccess

Name	Kind	Source	Default	Description
ANYPS5_ENTITLEMENTS	string	core/libs/prx/libSceNpEntitlementAccess/Export.cpp		path of the entitlements ini listing owned add-ons, defaulting to anyps5-entitlements.ini beside the executable

libSceNpTrophy2

Name	Kind	Source	Default	Description
APS5_TRACE_UDS	presence	core/libs/prx/libSceNpTrophy2/src/TitleTrophies.cpp		prints the UDS events contexts post and the stat changes trophy rules apply, and the trophy files a lookup does not find (read in NpUniversalDataSystem/Event and NpTrophy2/TitleTrophies)

libScePad

Name	Kind	Source	Default	Description
APS5_PAD_INPUT_FILE	string	core/libs/prx/libScePad/src/PadState.cpp		path of the file polled for injected pad button and stick commands

libScePlayGo

Name	Kind	Source	Default	Description
APS5_PLAYGO_UNLISTED_CHUNKS	string	core/libs/prx/libScePlayGo/Export.cpp		answer for a chunk ID the playgo manifest does not list: installed or absent; unset throws instead of answering

libSceSaveData.native

Name	Kind	Source	Default	Description
APS5_SAVEDATA_TRACE	presence	core/libs/prx/libSceSaveData.native/Export.cpp	3	logs every save-data operation instead of the first three

libSceSystemService

Name	Kind	Source	Default	Description
ANYPS5_LANGUAGE	integer	core/libs/prx/libSceSystemService/Export.cpp	1	console language number returned by sceSystemServiceParamGetInt, defaulting to English US

libSceVideoOut

Name	Kind	Source	Default	Description
ANYPS5_INPUT_CONFIG	string	core/libs/prx/libSceVideoOut/src/InputMapping.cpp		path of the input mapping file that binds host keys and mouse to guest pad actions
ANYPS5_REQUIRE_DISPLAY	presence	core/libs/prx/libSceVideoOut/tests/Flip.cpp		makes video-out tests fail instead of skipping when no display or Vulkan device is available
APS5_NO_PAD_INPUT	presence	core/libs/prx/libSceVideoOut/src/PadInput.cpp		ignores keyboard and mouse input so a measurement run is not advanced by a stray press
APS5_VIDEOOUT_TELEMETRY_DIR	string	core/libs/prx/libSceVideoOut/src/Telemetry.cpp		directory where videoout.jsonl is appended with the pad and present telemetry lines; nothing is opened when unset or empty

libSceVideodec2

Name	Kind	Source	Default	Description
APS5_TRACE_VIDEODEC	presence	core/libs/prx/libSceVideodec2/Export.cpp		logs decoder memory queries, queue allocation and the first few decodes

libc

Name	Kind	Source	Default	Description
APS5_CRASH_INDIRECT_MEMORY	presence	core/libs/prx/libc/src/specifics/windows/CrashReport.cpp		dumps the words the faulting registers point at in the Windows crash report
APS5_HANG_DUMP_SECS	integer	core/libs/prx/libc/src/specifics/windows/CrashReport.cpp		seconds after which a watchdog dumps all threads and aborts
APS5_HEAP_CACHE_MIB	integer	core/libs/prx/libc/src/GuestHeap.cpp	4096	size in MiB of the cache of freed large blocks before the oldest are decommitted
APS5_HELPER_AFFINITY_MASK	integer	core/libs/prx/libc/include/CpuTopology.hpp		hex core mask that helper threads (deferred release, mirror refresh) are pinned to
APS5_LOG_TIMESTAMPS	presence	core/libs/prx/libc/include/general/LogMacros.hpp		prefixes each log line with the milliseconds since the first log
APS5_NO_PROCMAP_QUERY	presence	core/libs/prx/libc/include/specifics/linux/HostMapping.hpp		skips the /proc/self/maps PROCMAP_QUERY ioctl and scans the maps file instead
APS5_NO_SSE4A_EMULATION	presence	core/libs/prx/libc/src/specifics/windows/CrashReport.cpp		disables emulation of AMD-only SSE4a instructions, making such a fault fatal
APS5_NO_WRITE_WATCH	presence	core/libs/prx/libc/src/GuestArena.cpp		disables write watching so guest buffers are fully re-read on every use
APS5_PIN_WAIT_MS	integer	core/libs/prx/libc/src/GuestAllocations.cpp	60000	milliseconds to wait for a pinned allocation lease before failing
APS5_TRACE_AFFINITY	presence	core/libs/prx/libc/include/CpuTopology.hpp		prints a line naming the role, tid and mask of every thread pinning
APS5_TRACE_CALLS	string	core/libs/prx/libc/src/specifics/x86_64/CallTrace.cpp		comma-separated libraries (or all) whose guest calls are traced
APS5_TRACE_EXIT	presence	core/libs/prx/libc/src/Process.cpp		prints a backtrace of the caller of exit, _Exit and the process-exit path
APS5_TRACE_SSE4A	presence	core/libs/prx/libc/src/specifics/windows/CrashReport.cpp		prints each emulated SSE4a instruction once per rip
APS5_WATCH_ADDR	integer	core/libs/prx/libc/src/specifics/windows/CrashReport.cpp		absolute hex address whose writes are reported after its page is committed
APS5_WATCH_WRITE	integer	core/libs/prx/libc/src/specifics/windows/CrashReport.cpp		guest ELF vaddr whose 8-byte writes are reported by single-stepping the protected page

libkernel

Name	Kind	Source	Default	Description
ANYPS5_LOADEXEC_ARGV_SKIP	string	core/libs/prx/libkernel/System/src/Process.cpp		on Windows, set to 1 to drop the first process argument and clear the variable
APS5_COARSE_TIMED_WAITS	presence	core/libs/prx/libkernel/Time/TimedWait.cpp		on Windows, rounds timed waits up to the scheduler tick
APS5_JOB_AFFINITY	presence	core/libs/prx/libkernel/Pthread/src/Thread.cpp		pins the title's job worker threads to the efficiency cores of a hybrid CPU
APS5_JOB_AFFINITY_MASK	integer	core/libs/prx/libkernel/Pthread/src/Thread.cpp		hex core mask that job worker threads are pinned to
APS5_NO_JOB_AFFINITY	presence	core/libs/prx/libkernel/Pthread/src/Thread.cpp		disables pinning of job worker threads
APS5_NO_SLEEP_WAKE	presence	core/libs/prx/libkernel/Time/SleepWake.cpp		keeps every sleep of 1 ms or less in full instead of ending it at a GPU batch, batch completion action or flip completion
APS5_NO_TIMER_RESOLUTION	presence	core/libs/prx/libkernel/Time/Time.cpp		keeps the default 15.6 ms scheduler tick instead of raising it to 0.5 ms on Windows
APS5_TIME_SCALE	integer	core/libs/prx/libkernel/Time/Time.cpp	1	factor applied to the guest monotonic clocks and TSC; sleeps still take real time
APS5_TRACE_APR	presence	core/libs/prx/libkernel/Apr/src/Apr.cpp		logs every APR file read with its path, offset, size and destination
APS5_TRACE_MEMORY	presence	core/libs/prx/libkernel/DirectMemory/DirectMemory.cpp		logs direct-memory map, remap, protect, unmap and physical allocation operations
APS5_TRACE_SYNC	presence	core/libs/prx/libkernel/EventFlag/src/EventFlag.cpp		logs event-flag create, set and wait operations
APS5_TRACE_TIMERS	presence	core/libs/prx/libkernel/Time/Time.cpp		runs a timed-wait selftest at startup and prints mean, p50 and max latencies
APS5_TRACE_USLEEP	presence	core/libs/prx/libkernel/Time/Time.cpp		reports every 2000 usleep calls the guest call sites that sleep most
APS5_TRACE_WAITS	presence	core/libs/prx/libkernel/Time/Time.cpp		aggregates blocking waits per guest call site and prints the longest-waiting sites every 3 s