In real PS3 (it seems), when a thread with a higher priority than the caller is signaled and that there is available space on the running queue for the other hardware thread to start It prioritizes signaled thread caller's hardware thread switches instantly to the new thread code while signaling to the other hardware thread to execute the caller's code.
Resulting in a delay to the caller after such thread is signaled
Fix and optimize sys_ppu_thread_yield
Fix LV2 syscalls with timeout bug. (use ppu_thread::cancel_sleep instead)
Move timeout notification out of mutex scope
Allow g_waiting timeouts to be awaked in scope
Hitherto the statistics have been exclusively thread-specific.
Other improvements:
* Fixed container management so a collision of a new element with an older element of the record will become impossible.
* Added thread name to thread-specific information printing.
* Fixed condition to abort SPU block statistics collection, now matches SPU LLVM Profiler's.
* Fix possible division by 0 by checking `samples`.
* Update asmjit dependency (aarch64 branch)
* Disable USE_DISCORD_RPC by default
* Dump some JIT objects in rpcs3 cache dir
* Add SIGILL handler for all platforms
* Fix resetting zeroing denormals in thread pool
* Refactor most v128:: utils into global gv_** functions
* Refactor PPU interpreter (incomplete), remove "precise"
* - Instruction specializations with multiple accuracy flags
* - Adjust calling convention for speed
* - Removed precise/fast setting, replaced with static
* - Started refactoring interpreters for building at runtime JIT
* (I got tired of poor compiler optimizations)
* - Expose some accuracy settings (SAT, NJ, VNAN, FPCC)
* - Add exec_bytes PPU thread variable (akin to cycle count)
* PPU LLVM: fix VCTUXS+VCTSXS instruction NaN results
* SPU interpreter: remove "precise" for now (extremely non-portable)
* - As with PPU, settings changed to static/dynamic for interpreters.
* - Precise options will be implemented later
* Fix termination after fatal error dialog
