EmuCR Feeds
Email Us

EmuCR: SharpEmu SharpEmu Git (2026/11/17) is compiled. An experimental PlayStation 5 emulator for Windows, Linux and macOS.

SharpEmu Changelog:
* [AGC/Loader/GPU/Kernel] Super Meat Boy 3D: AGC exports, red-zone patching, GPU-write protection before stores and two missing imports (#1146)
* fix tests
* [Diagnostics] Report dispatch replays and their misses
* [GPU] Replay a dispatch's bound state for the next dispatch of its pipeline
* [GPU] Compare a stage's inputs without its thread limits
* [Diagnostics] Report bounded device-address preparations and held slice uses
* [GPU] Hold a clean read slice use on its registered buffer
* [GPU] Skip the device-address sweep for programs whose reads are all bounded
* [GPU] Carry a plan's bounded device-address reads into its program info
* [ShaderCompiler] Mark a plan whose device-address reads are all bounded
* [ShaderCompiler] Bound a computed device-address read offset by its operations and access guard
* [ShaderCompiler] Read a level's runs in their static order
* [GPU] Read a resolve's remembered clean page without the read checks
* [GPU] Remember a resolve's last clean page for direct reads
* [ShaderCompiler] Order a materialization level's reads by base and offset once
* [Diagnostics] Report why a host refuses a draw replay
* [GPU] Replay a draw that repeats the previous item's resources
* [GPU] Capture a draw's bindings for a replay of the next item
* [Diagnostics] Log the command queue depth and slice lookahead per render worker
* [Diagnostics] Count metadata changes by call site
* [GPU] Repeat draw targets that resolve to the same attachments
* [GPU] Expose the image request comparison
* [GPU] Reset a draw record's resolve and copy fields only after those kinds
* [GPU] Read the slice lookahead per render worker from the environment
* [GPU] Hand the rest of a stolen slice to the render workers
* [GPU] Queue a slice's rest chunks through one helper
* [GPU] Stop a render worker slice at a repeated image lookup failure
* [GPU] Keep images used this frame when a render worker resolves an overlap
* [GPU] Read guest memory without the store lock when the page has no pending store
* [GPU] Skip zeroing the native color clear scan buffer
* [GPU] Take the write lock only when a native color clear registration changes
* [GPU] Share the native color clear registration check
* [GPU] Check stage memo freshness with one lock per stage
* [GPU] Validate a stage's remembered textures under one image cache lock
* [ShaderPrewarm] Leave the load address out of a compute record's identity

Download: SharpEmu Git (2026/11/17)
Source: Here



Random Related Topic Refresh Related Topic

Random Related Topic Loading...

0 Comments

Post a Comment

Can't post a comment? Try This!