There are some changes to the Automatic Test Mode in this pre-release.
If all goes well, this will also be the final 0.11.0.0 release, but I need people to test this with Ryzen 9000 CPUs.

Automatic Test Mode:

• The Automatic Test Mode should now support Ryzen 9000 (and 8000)!
• Added a new setVoltageOnlyForTestedCore config option, which when active will only set the currently tested core to a negative Curve Optimizer value. All the other cores will be set to 0, which avoids any possible crashes due to other cores being too low
• Added the creation of a System Restore Point when the Automatic Test Mode is being started. This helps recovering from a corrupted Windows installation (ask me how I know)
  This can be turned off in the config if you don't care about that
• Added a new Minimum option for the voltage starting values. These will set the Curve Optimizer values to -30 for Ryzen 5000 and -50 for Ryzen 7000 and above (and will throw an error for Intel)
• The default value for the starting values is now CurrentValues - and it does it says, it uses the currently set values as the starting point, which is actually the same behavior as the previous Default value, just more clear
• You can now provide the Curve Optimizer starting values separated with a | as well, which makes it easier to use the values provided at the end of a test
• There is now a 120 second waiting time after a crash before the testing is resumed, to avoid Windows registering this as a "failed" boot if the computer crashes immediately after starting the test again.
  Otherwise you might run into a situation where the Recovery Screen is being brought up during the reboot, and no testing can proceed.
  The waiting can be changed in the config file, and it can also be skipped by pressing a key
• The scheduled task for the auto-resume of the testing should now be more protected against simply vanishing during a crash (although it sometimes still seems to happen)
• The .automode file should not corrupt as often anymore during a crash

Misc changes:

• Switched to using ryzen-smu-cli for setting the Curve Optimizer values
• Added a new Ryzen.AutomaticTestMode.Start.ini config preset, which makes use of setVoltageOnlyForTestedCore = 1 and the Minimum starting values
• Added checks for Visual C++ libraries and for .NET 8
• The .automode file is now in JSON format
• Updated the included BoostTester tool to support a --core-repeat <n> argument, which allows for longer durations per core
• You can now use auto for the mode in y-cruncher, which will use the automatically selected binary that y-cruncher selects
• Added a new CorePairs option for the core testing order. This will form "pairs" for cores, so e.g. 0-1, 0-2, 0-n, up to n-0, n-1, n-(n-1)
• Added a Run Multiconfig CoreCycler.bat batch file that allows you to chain multiple config files, it will look for multiconfig-x.ini files (see the file comments for caveats)
• You can now define a single FFT size to Prime95 (so e.g. 720 instead of 720-720)

Multiple bug fixes:

• Fixed a bug where the core that crashed could be added multiple times to the testing sequence, making the last in the sequence to be skipped
• Fixed a bug where stopOnError caused problems in combination with the automatic test mode
• Added a check if one of the CCDs for Ryzen is disabled to prevent a bug
• Fixed a bug where a new log file would be started if the core that crashed the computer was core 0
• Fixed a bug with resuming the threads, which sometimes caused the script to abort
• Fixed an issue where setting the CO values would take longer than expected if y-cruncher was used
• Sometimes Aida64 didn't exit correctly when terminating the script. This should now work better
• Fixed a bug when registering the scheduled task
• Made the error checks for Visual C++ and .NET more precise
• Added functionality that checks if the System Restore is enabled, and if not, asks to enable it
• Added additional error check when creating the Event Log source
• Added multiple tries to get and set the Curve Optimizer values, because ZenStates-Core can seem to fail from time to time
• Waiting for one second before restarting the stress test program now, which hopefully should avoid errors if the allocated memory could not be freed fast enough
• Added some explanation for the CPULOAD error message
• Tried to avoid some errors when trying to get the stress test process

There are some changes to the Automatic Test Mode in this pre-release.
If all goes well, this will also be the final 0.11.0.0 release, but I need people to test this with Ryzen 9000 CPUs.

Automatic Test Mode:

• The Automatic Test Mode should now support Ryzen 9000 (and 8000)!
• Added a new setVoltageOnlyForTestedCore config option, which when active will only set the currently tested core to a negative Curve Optimizer value. All the other cores will be set to 0, which avoids any possible crashes due to other cores being too low
• Added the creation of a System Restore Point when the Automatic Test Mode is being started. This helps recovering from a corrupted Windows installation (ask me how I know)
  This can be turned off in the config if you don't care about that
• Added a new Minimum option for the voltage starting values. These will set the Curve Optimizer values to -30 for Ryzen 5000 and -50 for Ryzen 7000 and above (and will throw an error for Intel)
• The default value for the starting values is now CurrentValues - and it does it says, it uses the currently set values as the starting point, which is actually the same behavior as the previous Default value, just more clear
• You can now provide the Curve Optimizer starting values separated with a | as well, which makes it easier to use the values provided at the end of a test
• There is now a 120 second waiting time after a crash before the testing is resumed, to avoid Windows registering this as a "failed" boot if the computer crashes immediately after starting the test again.
  Otherwise you might run into a situation where the Recovery Screen is being brought up during the reboot, and no testing can proceed.
  The waiting can be changed in the config file, and it can also be skipped by pressing a key
• The scheduled task for the auto-resume of the testing should now be more protected against simply vanishing during a crash (although it sometimes still seems to happen)
• The .automode file should not corrupt as often anymore during a crash

Misc changes:

• Switched to using ryzen-smu-cli for setting the Curve Optimizer values
• Added a new Ryzen.AutomaticTestMode.Start.ini config preset, which makes use of setVoltageOnlyForTestedCore = 1 and the Minimum starting values
• Added checks for Visual C++ libraries and for .NET 8
• The .automode file is now in JSON format

Multiple bug fixes:

• Fixed a bug where the core that crashed could be added multiple times to the testing sequence, making the last in the sequence to be skipped
• Fixed a bug where stopOnError caused problems in combination with the automatic test mode
• Added a check if one of the CCDs for Ryzen is disabled to prevent a bug
• Fixed a bug where a new log file would be started if the core that crashed the computer was core 0
• Fixed a bug with resuming the threads, which sometimes caused the script to abort
• Fixed an issue where setting the CO values would take longer than expected if y-cruncher was used
• Sometimes Aida64 didn't exit correctly when terminating the script. This should now work better
• Fixed a bug when registering the scheduled task

Test version that includes ryzen-smu-cli for setting the Curve Optimizer values.
Should work with Ryzen 9000.

Not yet fully ready.

Version 0.10.0.0 offers a new "Automatic Test Mode" For Ryzen 5000 and 7000 and Intel processors up to 14th gen, which will automatically adjust the undervolt value upwards if a core has thrown an error.
To activate it you need to enable the enableAutomaticAdjustment setting in the config.ini. Additionally you can also enable the enableResumeAfterUnexpectedExit setting, which will try to resume the script if an unexpected crash of the computer has happened. Although for this to work correctly you would need to enable the automatic logon for Windows, otherwise the script cannot automatically restart.

More info can be found here:
https://learn.microsoft.com/en-us/sysinternals/downloads/autologon
https://learn.microsoft.com/en-us/troubleshoot/windows-server/user-profiles-and-logon/turn-on-automatic-logon

Changelog:

Final:

• The script now tries to find the stress test process more thoroughly (fixes #72)
• Now checking if the Windows Event Log service is running (fixes #101)
• Renamed WriteVerbose to WriteVerboseText and WriteDebug to WriteDebugText to prevent possible collisions with the built-in methods
• Disabled the "QuickEdit Mode" feature of the Windows command line terminal, which should prevent accidental script pauses
• You can now use "shortcuts" for y-cruncher (for example "19-ZN2" instead of "19-ZN2 ~ Kagari")

alpha5:

• Fixed a bug that prevented the automatic test mode to resume the test if the crash happened on core 0 (fixes #89)

alpha4:

• Fixed a bug with Linpack's path AGAIN
• The Automatic Test Mode unfortunately doesn't work with Ryzen 9000

alpha3:

• Added config presets for Automatic Test Mode
• Fixed a bug where Linpack wouldn't start from a directory including a space (fixes #84)
• Fixed a bug where resuming after a crash with a "default" coresToTest order would result in an index out of bounds error (fixes #85)
• Also it didn't resume on the crashed core
• The config file's content is now put into the log file for easier debugging
• Fixed a bug where the log file for y-cruncher was actually using the naming scheme for Linpack (under unclear circumstances, but it happened at least once)
• Fixed a bug when the update check doesn't return any output at all (fixes #83)

alpha2:

• Fixed a bug with the detection of the starting values (see #79)

alpha1:

• Added an automated test mode with automatic Curve Optimizer / voltage offset adjustment on error
• Added the possibility to automatically resume the script after a reboot via a Scheduled Task
• A WHEA warning/error can now be treated as a "real" error if the APIC ID matches the tested core
• See the default.config.ini for the new settings

alpha5:

• Fixed a bug that prevented the automatic test mode to resume the test if the crash happened on core 0 (fixes #89)

alpha4:

• Fixed a bug with Linpack's path AGAIN
• The Automatic Test Mode unfortunately doesn't work with Ryzen 9000

alpha3:

• Added config presets for Automatic Test Mode
• Fixed a bug where Linpack wouldn't start from a directory including a space (fixes #84)
• Fixed a bug where resuming after a crash with a "default" coresToTest order would result in an index out of bounds error (fixes #85)
• Also it didn't resume on the crashed core
• The config file's content is now put into the log file for easier debugging
• Fixed a bug where the log file for y-cruncher was actually using the naming scheme for Linpack (under unclear circumstances, but it happened at least once)
• Fixed a bug when the update check doesn't return any output at all (fixes #83)

alpha2:

• Fixed a bug with the detection of the starting values (see #79)

alpha1:

• Added an automated test mode with automatic Curve Optimizer / voltage offset adjustment on error
• Added the possibility to automatically resume the script after a reboot via a Scheduled Task
• A WHEA warning/error can now be treated as a "real" error if the APIC ID matches the tested core
• See the default.config.ini for the new settings

• Added an automated test mode with automatic Curve Optimizer / voltage offset adjustment on error
• Added the possibility to automatically resume the script after a reboot via a Scheduled Task
• A WHEA warning/error can now be treated as a "real" error if the APIC ID matches the tested core
• See the default.config.ini for the new settings
• Fixed a bug with the detection of the starting values (see #79)
• Added config presets for Automatic Test Mode
• Fixed a bug where Linpack wouldn't start from a directory including a space (fixes #84)
• Fixed a bug where resuming after a crash with a "default" coresToTest order would result in an index out of bounds error (fixes #85)
• Also it didn't resume on the crashed core
• The config file's content is now put into the log file for easier debugging
• Fixed a bug where the log file for y-cruncher was actually using the naming scheme for Linpack (under unclear circumstances, but it happened at least once)
• Fixed a bug when the update check doesn't return any output at all (fixes #83)
• Fixed a bug with Linpack's path AGAIN
• The Automatic Test Mode unfortunately doesn't work with Ryzen 9000

• Added an automated test mode with automatic Curve Optimizer / voltage offset adjustment on error
• Added the possibility to automatically resume the script after a reboot via a Scheduled Task
• A WHEA warning/error can now be treated as a "real" error if the APIC ID matches the tested core
• See the default.config.ini for the new settings
• Fixed a bug with the detection of the starting values (see #79)
• Added config presets for Automatic Test Mode
• Fixed a bug where Linpack wouldn't start from a directory including a space (fixes #84)
• Fixed a bug where resuming after a crash with a "default" coresToTest order would result in an index out of bounds error (fixes #85)
• Also it didn't resume on the crashed core
• The config file's content is now put into the log file for easier debugging
• Fixed a bug where the log file for y-cruncher was actually using the naming scheme for Linpack (under unclear circumstances, but it happened at least once)
• Fixed a bug when the update check doesn't return any output at all (fixes #83)

Fixed a bug with the detection of the starting values (see #79)

• Added an automated test mode with automatic Curve Optimizer / voltage offset adjustment on error
• Added the possibility to automatically resume the script after a reboot via a Scheduled Task
• A WHEA warning/error can now be treated as a "real" error if the APIC ID matches the tested core
• See the default.config.ini for the new settings

Minor bug fixed when only a single core was selected and a random test order was chosen.
Also fixed a bug if no valid core to test could be selected.