# Cemu detailed changelog for 1.10.0f
# Patreon release date: 2017-09-25
# Public release date:  2017-10-02

# New in 1.10.0f:

general: Made RDTSC more robust against aggressive Out-of-order execution on some AMD CPU models

# New in 1.10.0e:

GX2: Fixed texture corruption issue in GX2CopySurface
coreinit: Fixed a bug in thread scheduler that could lead to worse performance
AX: Maximum audio volume is now a bit louder
general: Removed 'disable audio' debug option    

# New in 1.10.0d:

general: Fixed a race condition in RDTSC timer code
AX: Slightly optimized audio code

# New in 1.10.0c:

FS: Fixed a bug introduced by 1.10.0 that caused the creation of save game directories to fail
GX2: Fixed a crash that occurred when the streamout cache ran out of memory

# New in 1.10.0b:

general: Fixed a bug where 'Use RDTSC' wasn't always enabled per default

# New in 1.10.0:

general: Added support for RDTSC if invariant mode is supported by CPU
general: Added -mlc <path> command line parameter to customize the path of the mlc01 directory
general: Moved some experimental settings from the debug menu to the options menu
general: Cemu will now use the local timezone for Cafe OS calendar time
general: Adjusted some game profiles (cycle counter no longer enforced, emulateSinglePrecision=true removed)

CPU/JIT: emulateSinglePrecision is now enabled by default

GX2: Fixed secondary instance step value being incorrectly set in GX2SetFetchShader
GX2: Added 'vsyncFrequency' property to graphic pack rules.txt (see example_pack.zip for details)
GX2: Fixed a crash that occurred when a shader link error was encountered
GX2: Adjust maximum mip level for rescaled textures
GX2: Slightly optimized some frequently used GX2 API
GX2: Submitting GPU7 commands on the wrong PPC core will no longer crash Cemu

coreinit: Added API MEMSetAllocModeForExpHeap(), MEMFreeToExpHeap()

Input: Fixed a bug where keyboard special characters where always detected as pressed
Input: Improvements to input profile system

AX: Rewritten audio code. Compared to the old audio backend the following has improved:
        Correct channel and volume mixing
        Correct implementation of voice protection (no more playback offset randomly running out of bounds)
        Generally improved audio timing (less stutter and crackle)
        Correct implementation of aux stage
        Accurate emulation of all AX API