eglantine-cirakg/src/video_core/memory_manager.cpp, branch nx2

nx2: leave unwritten sparse blocks without memory when they get zeros

2026-11-01T20:18:01+00:00

Title 1 clears about 1 GB of its sparse regions at boot through the GPU memory manager. Every block the clear touched got 4 MiB of guest memory, which the clear itself then committed on the host, and counted as written, so the texture cache gave the images over it device memory too. A block nobody wrote reads zero already: when the part of a write that lands on it is all zero, leave it as it is. (cherry picked from commit 0d7ff63fd1a373cecff6ba91022c9027e801e53f)

memory_manager: copy big pages without host memory page by page

2026-11-01T04:11:02+00:00

GPU reads and writes of a continuous big page copied through GetPointer() without checking it; the small-page path checks. A big page whose host memory is not there yet (a Switch 2 region backed as it is written, with its backing pending) made that memcpy fault: Mario Kart World crashed in WriteBlockImpl (access violation) on its title screen in two runs of a build on the 10-31 nx2. Such a page now takes the per-page copy, which reports and skips missing pages, as non-continuous big pages already did. Three 10-minute runs after the change, no crash.

nx2: give sparse regions memory only where the GPU writes

2026-10-30T16:27:04+00:00

nx2:gpu 214 regions with the sparse flag (word 7 bit 0x200) are Mario Kart World's terrain texture arrays and layer pools, 170 MB to 5.8 GB, mostly empty. Backed whole they ran the graphics pool dry (26 "could not allocate" per session, later regions without memory), and backing a block at a time made the texture cache rebuild the 1.45 GB arrays each time a block was mapped. A sparse region now gets a device (SMMU) range of its own when it is created, mapped for the GPU at once, so its GPU mapping never changes. Before a GPU-side write lands in it (WriteBlock, Write), the 4 MiB blocks it touches get memory, cleared, behind their device pages. Unbacked device pages read zero and drop writes instead of crashing, and the unmapped-device-access errors are logged 16 times, not once per page. nx2:gpu 215 releases the device range and the blocks after the GPU is done with them. The log reports GPU-only memory in use and its peak. Mario Kart World, 180 s: no allocation failures; ~2.8 GB of terrain blocks get memory instead of 6.5 GB of whole regions, next to 3 GB of ordinary GPU-only regions. Terrain renders as before.

memory_manager: notify the caches once per changed range

2026-10-30T16:27:04+00:00

PageTableOp and BigPageTableOp called ModifyGPUMemory for every page whose entry changed, and each call scans the texture and buffer caches. Mapping one of Mario Kart World's 1.45 GB sparse regions made ~370,000 of those scans. Runs of changed pages are now reported as one range.

video_core: cheaper small GPU reads, buffer usage tracking and compute uniforms

2026-10-29T05:26:52+00:00

- MemoryManager::ReadBlockUnsafe reads within one 4 KiB page through a single translation (as Read does) instead of the page walk: texture and sampler descriptors, several per draw (2.6% -> 0.1% of title 1's GPU thread). - UsageTracker skips ranges it tracked just before (until Reset): draws bind the same multi-megabyte storage buffers again and again, and each Track wrote a word per 4 KiB (2.0% -> 0.7%). - Small compute uniform buffers the GPU has not written are streamed, as graphics ones are; synchronizing them re-protected pages an inline upload had just unprotected, two VirtualProtect calls per dispatch (1.9% -> 0.7%, page protection 3.2% -> 2.3%).

Switch 2: Turing macros, native compute, Ampere shader lowering, native tiling

2026-10-28T02:26:16+00:00

From changes.patch: the TU104/GA10x MME interpreter (shared macro context for the process's channels, also on the compute class), native compute with QMD 3.0 dependency scheduling, the Ampere-to-Maxwell shader lowering, native vertex stream sizes, and the console's GOB swizzle for textures, DMA copies and the host unswizzle shaders. Switch 2 channels turn all of it on at init. The lowering is widened for title 4 (all 32 of its programs lower): general TEX (any dimension/LOD/mask, operand split 1+1/2+1/2+2/4+n, any but consistent descriptor buffer), cube/array, SULD channel masks, 128-bit STG, indexed IPA, VOTE, SHFL with any immediate or predicate output, R2P/P2R byte select, uniform ULOP3/ULEA/UBREV/UFLO.SH/UISETP signed/UMOV URZ, IMNMX cbuf, IADD3 and FFMA uniform operands, FMUL /4 /8 x8, and BSSY/BSYNC/BREAK/WARPSYNC as no-ops. The capture stops at unmapped memory. EDEN_NX2_SHADER_DUMP dumps every native program; tools in C:\dev\nx2work test the lowering offline.

Switch 2: suspend channels on semaphore acquires; fix pushbuffer desyncs

2026-10-28T00:53:15+00:00

NVN2 gates command buffers with semaphores the CPU releases after writing them. Eden read past unmet SEMAPHORED acquires, so it executed stale commands and desynchronised. Channels now suspend on an unmet acquire and resume from the same word once it is met (scheduler keeps the waiting channels, the GPU thread polls them), as in changes.patch. Also: methods split across GPFIFO entries around a called command buffer are set aside and resumed; subchannel 2 defaults to inline-to-memory; unbound subchannels and out-of-range compute/DMA methods are dropped instead of crashing; 4-byte semaphore releases no longer write 16 bytes; Ampere SEM_PAYLOAD_HI/SEM_EXECUTE no longer spin as Kepler acquires; unmapped GPU reads log instead of asserting. Debug aids: EDEN_NX2_PB_CHECK (pushbuffer changed after submit), EDEN_NX2_QUERY_TRACE, a one-time pushbuffer dump.

Switch 2: native memory/images, GPU-accessible attribute, TSG subcontexts, syncpoint shim

2026-10-27T23:30:20+00:00

- gsv 111/112/114/423/429/430/314/315/435 track native memory and images with real ids (meanings from changes.patch); 1510/435 use its block-linear layout. - svc 0x80 sets a real GpuAccessible (0x20) memory attribute (from changes.patch). - Characteristics report TSG subcontexts (flag 22, 64 per TSG at 0xb4); NvRm otherwise allows one subcontext and a second NVN queue fails its setup. - Syncpoint shim: flag 26, address-space ioctl 0xB, and GPU reads of the shim page return guest syncpoint values (scheme from changes.patch). - EDEN_NX2_IPC_TRACE also logs TIPC commands; gdbrsp.py watch/mem improvements.

[*] Remove Fix{Static,Small}VectorADL workaround for GCC 12 (#4386)

2026-10-12T20:21:17+00:00

1. We demand at least, minimum, gcc 14 2. Debian stable boost uses a newer version https://packages.debian.org/trixie/libboost-all-dev, so fix isn't longer required - [x] I have read and followed the [Contribution Guidelines](https://git.eden-emu.dev/eden-emu/eden/src/branch/master/CONTRIBUTING.md#code-contributions). - [x] I have read and followed the [AI Policy](https://git.eden-emu.dev/eden-emu/eden/src/branch/master/docs/policies/AI.md) - [x] I have read and followed the [Coding Guidelines](https://git.eden-emu.dev/eden-emu/eden/src/branch/master/docs/policies/Coding.md) to the best of my ability. ------------------- Reviewed-on: https://git.eden-emu.dev/eden-emu/eden/pulls/4386 Reviewed-by: MaranBr Reviewed-by: crueter

[common/core/dynarmic] Optimize page table allocations (#4219)

2026-09-09T19:35:49+00:00

Reduces the page entries from 32 bytes to 8 and rewrites `VirtualBuffer` to be more efficient in memory usage and specifically for large zero regions. The page table will now only reserve 1GiB instead of 4GiB and of this memory it should only use at most ~8MiB. This PR has the side effect of using Eden on Windows on low memory systems much more plausible since it would previously require ~10GiB of committable memory at front (despite using ~5-6 at most, inadvertently stalling other processes) where as now it should only require around the amount that it'll actually use. Co-authored-by: Lizzie Reviewed-on: https://git.eden-emu.dev/eden-emu/eden/pulls/4219 Reviewed-by: Maufeat Reviewed-by: lizzie Reviewed-by: CamilleLaVey