SharpEmu Git (2026/10/08) is compiled. An experimental PlayStation 5 emulator for Windows, Linux and macOS.

SharpEmu Changelog:
* Unity fixes 1016p2 (#1062)
* Revert the 18 Silent Hill rendering and performance merges (#1065)
* [GPU] Sweep CPU writes at memory visibility points instead of before every device-address program (#1033)
* [GPU] Add internal render resolution scaling (#1038)
* [GPU/HLE] Cut per-draw and per-call managed allocations on hot paths (#1054)
* [GPU] Restore scene visibility and half-wave shuffles (#1025)
* [GPU] Avoid command-worker readbacks for clean bytes on GPU-dirty pages (#1059)
* [GPU] Read guest buffers back from unified memory without a copy (#1058)
* [macOS] Fix the Rosetta GC wedge, a CPU-flip buffer-wait deadlock and frozen host movies (#1051)
* [GPU] Restore half-float arithmetic and image readback state (#1024)
* [HLE/Cpu] Remove host lock contention from POSIX memory queries, thread resumes and timed waits (#1031)
* [SpirvTranslator] Share one EXEC test across a run of vector ALU writes (#1040)
* [SpirvTranslator/Vk] Cheaper guest buffer loads and f16 conversion, gated on device probes (#1035)
* [SpirvTranslator] Cut emulated-wave64 and private-LDS overhead in translated shaders (#1034)
* [GPU/Vk] Keep guest readbacks of GPU-written memory off the command worker (#1032)
* [SpirvTranslator] Keep the packed-half vector registers in one variable per register (#1041)
* [GPU] Cut per-draw command-worker cost in resource materialization and pixel input setup (#1039)
* [GPU] Keep sampled depth attachments in one rendering scope (#1037)
* [GPU] Draw non-indexed indirect draws from their arguments on the GPU (#1036)
* [Cpu] Copy page-sized guest memcpys on the managed path (#1055)

Download: SharpEmu Git (2026/10/08)
Source: Here