MSI AB / RTSS development news thread

Thanks

 

Official MSI AB v4.6.0 and RTSS v7.2.1 are uploading and will be available for download shortly. Full changes list includes:

MSI Afterburner 4.6.0 build 14820:

- Added NVIDIA Turing GPU architecture support:
o Added voltage control for reference design NVIDIA GeForce RTX 20x0 series graphics cards
o Advanced GPU Boost control for NVIDIA GeForce RTX 20x0 series graphics cards. Extended voltage/frequency curve editor on GeForce RTX 20x0 family graphics cards allows you to tune additional piecewise power/frequency floor and temperature/frequency floor curves. Control points on those new curves allow you to control GPU Boost power and thermal throttling algorithms more precisely than traditional power limit and thermal limit sliders
o Hardware abstraction layer has been revamped to provide support for multiple independent fans per GPU due to introducing dual fan design on reference design NVIDIA GeForce RTX 20x0 series graphics cards and due to introducing native dual fan control in NVAPI. Both fans of NVIDIA GeForce RTX 20x0 can be monitored independently in hardware monitoring module now and can be controlled synchronically in manual mode
o Added NVIDIA Scanner technology support
- Added Monolithic Power Systems MP2884A and MP2888A voltage controllers support
- Added core, memory, auxiliary PEXVDD and auxiliary 1V8 voltage control for custom design MSI RTX 2080Ti Lightning series graphics cards with MP2888A + MP2884A +uP1816 voltage regulators
- Added VRM and VRM2 temperature monitoring for custom design MSI RTX 2080Ti Lightning series graphics cards
- Added AMD Vega 20 GPU architecture support:
o Added Overdrive 8 technology support for AMD RADEON VII series graphics cards
o Added extended thermal monitoring for RADEON VII series graphics cards. Extended thermal monitoring includes dual-channel GPU temperature monitoring (edge and junction temperatures), memory temperature monitoring and quad-channel VRM temperature monitoring (core, SOC and two memory VRM temperatures)
o GPU usage monitoring filtering algorithms are reenabled for AMD ADL based GPU usage monitoring implementation to minimize fluctuations in monitored GPU usage on AMD RADEON VII series graphics cards. Please take a note that filtering algorithms make GPU usage graph more inert and smooth, so you may disable them via configuration file if necessary
- Added fan control support for AMD Adrenalin 2019 drivers family. Please take a note that AMD Adrenalin 2019 drivers adjust fan speed in PWM duty cycle mode but fan speed monitoring is provided in RPM only, which means that you cannot read back exact programmed fan speed in percents. Due to this limitation, MSI Afterburner is reporting fan speed as percent of maximum RPM instead of real fan duty cycle when default AMD auto fan control is in use. So monitored fan speed can slightly mismatch programmed fan speed curve if you're using default AMD automatic fan control
- Added new MSI Gaming Z skin by Drerex Design
- Improved hardware monitoring module:
o Added thermal offset for CPU temperature monitoring on AMD Ryzen 7 2700X processors
o Similar CPU temperatures are now cloned for all cores on AMD CPUs to provide more convenient and unified On-Screen Display layout between Intel and AMD CPUs
o “Pagefile usage” graph in hardware monitoring module has been renamed to “Commit charge”
o Added group operation notifier to the caption of active hardware monitoring graphs list in “Monitoring” tab in the application properties. The notifier is informing you that you can apply a setting to all or to multiple selected hardware monitoring graphs when you hold group operation key (i.e. <Shift> or <Ctrl>) and hover mouse cursor over the setting that can be applied to multiple hardware monitoring graphs at once (e.g. “Show in On-Screen Display”)
o Added “Select by group” and “Select by type” commands to context menu of active hardware monitoring graphs list in “Monitoring” tab in the application properties. New commands allow selecting the graphs of exactly the same type (e.g. CPU1-CPU8 usage graphs on eight core CPU) or the graphs with the same OSD/LCD group name (for example, GPU clock, GPU power, GPU usage and the rest graphs with “GPU” group name). Such commands allow you to show similar items (e.g. all CPU usages) in On-Screen Display with a single click or rename a group for all child graphs in the single pass
o Multi group edit mode is automatically selected now for group name edit field when you enable "Override group name" option while holding <Ctrl> key. You may still manually toggle this mode with <Ctrl>+<M> hotkey anytime when editing the group name
o On-Screen Display item type (i.e. text, graph or combination of both) is now displayed in “Properties” column of active hardware monitoring graphs list in “Monitoring” tab in the application properties
o Added programmable hotkey for clearing monitoring history
o Added “Apply” button to application properties window. Now you may iteratively customize monitoring module without closing the properties to see the result on each step. Please take a note that some properties (e.g. application language) cannot be applied until you completely close properties window by design of application
o Application properties dialog window is now displayed with topmost style when the properties are activated from detached monitoring window and “Always on top” mode is enabled for it
o Now “Mark minimum” and “Mark maximum” commands are available in the context menu of monitoring window even when a fragment of history is not selected. In this case the commands result in setting tracking marker in a global extremum position. You may still use <Shift> key and mouse cursor to select a fragment of history then use the context menu to set tracker marker in local extremum position
o Now you may press <F11> key in monitoring window to capture print friendly screenshot of hardware monitoring history graphs. The screenshot includes MSI Afterburner watermark by default, you may customize it by replacing overlay.png file located in .\Graphics folder. This feature can be useful for hardware reviewers
o Now you may hold <Del> key then click unwanted graphs in hardware monitoring history viewer mode to hide them. This feature can be useful before capturing print friendly screenshot, if you want to exclude unnecessary graphs from it
o Improved On-Screen Display layout editor:
- System reserved variable slots 250-255 are now displayed as system variables in the layout editor
- Added embedded graphs support. Embedded graphs can be displayed directly in the text in a separate column, it can be convenient for displaying per-core CPU usages
- Added embedded graph width adjustment variable to the “Graph” property node
- Added new “Graph placement” property node. You can use it to place each graph on top of text, on bottom or make it embedded in in text column
- Added new variables to “Separators” property node. Now you may define independent separators for the graphs located on top of text part, for the text part and for the graphs located on bottom of text part.
- Slightly altered interpretation of prolog and epilog separators. Please take a note that prolog separator is no longer automatically appended with new line symbol, so you must specify it manually when necessary and epilog separator is now displayed below the bottom group of graphs
- The list of hardware monitoring items in “Add item(s)” dialog window is alphabetically sorted now
- Added macroses for embedding any data from monitoring module into OSD text (e.g. prolog). You may use macroses to display anything available in the monitoring module, e.g. %CPU temperature%
- Improved On-Screen Display hypertext formatting performance due to more optimal macro replace strategy
- Revamped “modern web” layout. New layout is demonstrating new text formatting features including big custom framerate counter, custom embedded background and On-Screed Display header, custom benchmark and system time indicators and embedded CPU and GPU usage graphs
- On-Screen Display layouts are saved in monitoring profiles now
o Improved HwInfo plugin:
- Added UPS load, power, input voltage and charge level data sources to default plugin configuration
- Improved voltage/frequency editor:
o Voltage/frequency editor window is now available on AMD GPUs. On AMD platform it allows you to edit core frequencies and voltages independently per each P-State:
- Similar to NVIDIA GPUs, you may either adjust each point frequency or voltage independently with both mouse and keyboard (<Up> / <Down> or <Ctrl>+<Up> / <Ctrl>+<Down> keys), hold <Shift> key and drag single point to apply similar frequency or voltage offset to all P-states or hold <Alt> key and drag single point to add the same offset to each P-State frequency or voltage offset
- You may press <PageUp> / <PageDown> keys to switch between frequency or voltage keyboard input focus for selected P-State
- Read only P-States are currently not tracked, which means that you can try to adjust the frequency and voltage for locked states but changes will be ignored when applying the curve
- P-States locking functionality is not implemented yet
o Now you may press <Tab> / <Shift>+<Tab> keys when a point is selected to switch to the next / previous point. This feature can be useful if you’re fine-tuning the offsets from keyboard
o Now you may press <Enter> key when a point is selected to type in desired frequency or voltage offset for it
o Now you may press <Shift>+<Enter> keys when a point is selected to type in absolute desired frequency or voltage explicitly and let the editor to calculate the offset automatically
o Added undo/redo support:
- Now voltage/frequency curve editor is saving up to 1024 last states of the curve during editing and allows you to undo/redo changes with <Ctrl>+<Z> or <Ctrl>+<Y> / <Ctrl>+<Shift>+<Z> keys
- Undo history is discarded when you apply the curve
- Number of recorded undo/redo steps is displayed in square brackets in the editor window caption
o Slightly increased default voltage range for voltage/frequency curve. Now you may lock voltage to lower values if you’re using voltage/frequency editor to downvolt and reduce power consumption on your NVIDIA GPU. Power users may customize the range further, the voltage and frequency limits are fully adjustable via the configuration file
o Now you may press <F5> key in the editor window to reread the curve from hardware. This feature can be useful when adjusting the curve in changing thermal environment
o Voltage and frequency axes are dynamically scaled now, so the captions do not overlap when window is too small
- Improved hardware control shared memory interface. During the past years, external applications like MSI Remote Server were using this interface for tuning GPU hardware settings remotely from external applications. The improvements are intended to allow connecting external stress testing and automatic overclocking related applications to MSI Afterburner via this interface:
o Now voltage/frequency curve on NVIDIA Pascal and newer NVIDIA GPU architectures is accessible via hardware control shared memory interface
o New hardware control shared memory interface command allows MSI Afterburner to load hardware settings from external application without immediately applying new settings to GPU
o Added notification message, allowing external applications to notify MSI Afterburner about new command written to hardware control shared memory. Without the notification, MSI Afterburner is executing external commands on each hardware polling iteration like before. Please refer to SDK and MACMSharedMemorySample source code to see notification message usage example
o Added hardware identification info to GPU entries in hardware control shared memory. Hardware identification info allows external applications to reconcile own enumerated devices with logical GPUs enumerated by MSI Afterburner
o Now hardware control shared memory is refreshed on delayed fan speed readback events
- New bundled MSI Overclocking Scanner application in now included in MSI Afterburner distributive:
o MSI Overclocking Scanner is currently supported on NVIDIA GTX 10x0 and NVIDIA RTX 20x0 series graphics cards under 64-bit operating systems only. On such systems you may activate the scanner with dedicated button from the main application window or directly from voltage/frequency curve editor window
o MSI Overclocking Scanner is powered by NVIDIA Scanner technology, which is using proprietary algorithms to quickly and reliably test manually overclocked GPU stability or find the maximum stable GPU overclocking in automatic mode with a single click. The scanner is using embedded NVIDIA test load to stress GPU. The scanner provides you two functional modes:
- In test mode MSI Overclocking Scanner is stress-testing your manual GPU overclocking settings during approximately 5 minutes. The result is returned as GPU stability confidence level (0% - unstable, 100% - stable)
- In scan mode MSI Overclocking Scanner is stress-testing and slowly increasing clocks on voltage/frequency curve points and this way automatically detecting the maximum stable GPU overclocking. The result is returned as modified voltage/frequency curve and average GPU overclocking in MHz
- Added framerate limiter compatibility settings group to “On-Screen Display” tab to the application properties. The settings include programmable hotkeys for globally disabling, enabling or toggling framerate limiting techniques of RivaTuner Statistics Server. The settings are intended to provide a way to dynamically enable or disable scanline sync technology based framarate limiting without altering application profiles, however you may use it to toggle traditional framerate limter as well
- Now CPU topology info (i.e. each logical CPU mapping to physical CPU package and core) is displayed in system info window
- Fixed kernel mode hardware access interface fix for AMD Vega GPUs. HBM temperature, GPU power and GPU voltage are no longer returning bogus values when kernel mode hardware access interface is enabled on AMD Vega series graphics cards
- Fixed low-level GPU usage monitoring for AMD Polaris GPU family. Software SMC registers offset is no longer hardcoded, offset detection algorithm is more future proof now. Please take a note that it will not fix zero GPU load issue when "Enable unified GPU
usage monitoring" option is enabled. This option implementation is based on low-level DirectX performance counters, which require OS and display driver WDDM versions to match. Current AMD drivers are WDDM 2.5 so this performance counter is only working when you're under WDDM 2.5 OS (Windows October Update). If you're under older OS version, this performance counter won't work unless you downgrade to WDDM 2.4 driver (18.9.3 or older)
- Hardcoded voltage/frequency curve clock multiplier implementation has been replaced with heuristic multiplier detection in order to provide unified voltage/frequency curve control implementation for NVIDIA Pascal and newer NVIDIA GPU architectures
- Improved realtime voltage/frequency curve editor GUI scaling. Now GUI is scaled property when adjusting skin scaling with open voltage/frequency curve editor window
- Improved report window, fan curve editor window, monitoring window and voltage/frequency curve editor scaling for scaling ratios greater than 100%
- Now mandatory MSI Afterburner installation option is greyed out and cannot be unchecked during installation
- Now 64-bit Visual C++ 2008 runtimes are also installed by MSI Afterburner installer. 64-bit runtimes are required for proper functionality of bundled MSI Overclocking Scanner application
- Multilanguage user interface library is DPI aware now. Main window skinned interface is no longer being scaled by OS and no longer looks blurred by default when greater than 100% DPI is selected. Now main skinned window interface can be manually zoomed with
skin scaling slider in “User interface” tab, and the properties are automatically scaled by OS according to selected DPI
- Added new skin scaling mode for power users. New mode provides better scaling performance at the expense of some scaling image quality loss
- Skin scaling implementation is now asynchronous, so increasing skin scaling ratio doesn’t increase GUI update time and doesn’t decrease GUI response time
- Skin scaling slider range has been extended from 75%-150% to 75%-200%
- RivaTuner Statistics Server has been upgraded to v7.2.1

 

RivaTuner Statistics Server v7.2.1 build 20237:


- Fixed issue with the latest NVIDIA Adaptive Shading update of Woflenstein II: The New Colossus locking on startup. Please take a note that now Woflenstein II: The New Colossus is presenting frames from compute queue in NVIDIA codepath similar to AMD codepath. Due to that reason, On-Screen Display rendering is performance expensive and disabled by default in such environment.Experienced users, understanding and accepting On-Screen Display related performance penalty, may reenable it with PresentFromCompute profile switch
- Fixed black screen issue in Vulkan applications on NVIDIA Kepler GPU family under 4xx.xx series drivers family
- Fixed multhithreaded active busy-wait loop based frame capture implementation, which could prevent desktop videocapture from working in the previous version
- Improved compatibility with multithreaded Direct3D1x applications, using multiple DXGI swapchains and concurrently accessing them from different threads (e.g. Microsoft Photos UWP application)
- Improved On-Screen Display hypertext formatting implementation:
o Variable slot 250 is no longer used to define own On-Screen Display statistics format, now it is allowing the client applications to redefine default text format. Such implementation allows the client applications like MSI Afterburner to edit default text color and size directly inside the On-Screen Display layout editor
o Improved subpixel precision for raster fonts provides better font scaling quality and improves rendering quality for small superscript/subscript-styled text
o Improved font baseline positioning. Now subscript-styled text is aligned on the baseline of maximum sized font symbol in the line instead of the baseline for 100% font size
o Improved font scaling implementation for large fonts. Now requested font size is scaled from the closest reference prerendered font size instead of 100% font size
o Changed set of prerendered font sizes from 40%, 50%, 80% and 100% to 50%, 100% and 200%. Prerendered font sizes can be used as is without scaling and with the best possible rendering quality, any different font sizes are interpolated
o Added background fill flag for embedded objects. Background fill is intended to simplify maximum value visualization with barchart-styled embedded graphs
o Now it is allowed to use zero width or height in bar embedding tags to specify whole text extent
o Embedded objects are now affected by text size and alignment tags
o Added backspace symbols support. Backspaces are intended to be used in conjunction with bar embedding tags. Inserting backspace symbol immediately after the bar embedding tag results in returning the cursor position back to embedded bar origin. Such implementation provides simplified way of embedding custom background in the hypertext. Please take a note that backspace sequences are currently not supported, so the cursor cannot be moved beyond the previous position
o Added new tag for displaying benchmark time
- Added suspending support to VTable hook handlers. Suspending is used to improve reference counter tracking implementation and make it more future proof
- Graphs are no longer displayed in On-Screen Display when Vector2D On-Screen Display rendering mode is selected
- Slightly refactored Direct3D and Vulkan On-Screen Display rendering codepaths due to implementation of optional offscreen rendering mode. This mode is currently disabled by default, it is reserved for future versions as a codebase for asynchronous On-Screen Display rendering implementation
- Multilanguage user interface library is DPI aware now. Main window skinned interface is no longer being scaled by OS and no longer looks blurred by default when greater than 100% DPI is selected. Now main skinned window interface can be manually zoomed with skin scaling slider in “User interface” tab, and the properties are automatically scaled by OS according to selected DPI
- Added new skin scaling mode for power users. New mode provides better scaling performance at the expense of some scaling image quality loss
- Skin scaling implementation is now asynchronous, so increasing skin scaling ratio doesn’t increase GUI update time and doesn’ t decrease GUI response time
- Skin scaling slider range has been extended from 75%-150% to 75%-200%
- Some profile settings are treated as power user oriented and they are locked by default at global profile level now. Such settings include “Enable Custom Direct3D support” compatibility option and Vector2D On-Screen Display rendering mode selection. Context help was not recommending enabling those compatibility related options globally, however many users tended to enable them blindly then blamed the product for inappropriate performance or reduced 3D applications compatibility. Such settings are now ignored and locked for enabling in global profile unless you manually unlock them
- Added “Enable framerate limiter” option to general application properties. This option allow controlling RivaTuner Statistics Server’s framerate limiting support globally, so you may temporarily disable it if necessary without modifying generic framerate limiter or
scanline sync related settings in the profiles. The client applications like MSI Afterburner can also distantly control this option via hotkeys
- Now RivaTuner Statistics Server validates DirectX runtimes installation at startup and displays warning message if required DirectX runtime components are missing. Power users may disable this warning message as well as any previously existing diagnostic warning messages displayed at RivaTuner Statistics Server startup and enable silent startup mode via the configuration file if necessary
- Added user extendable exclusions list for desktop windows notification performed during installation and uninstallation of system CBT hooks. Currently the only exclusion in the list is Microsoft Outlook process, which previously tended to display help popup window on closing RivaTuner Statistics Server
- Now RivaTuner Statistics Server hides own frametime history overlay when framerate or frametime is displayed by the client On-Screen Display application
- Now RivaTuner Statistics Server hides own benchmark statistics when benchmark time, minimum, average, maximum, 1% low, 0.1% low framerate is displayed by the client On-Screen Display application
- Added On-Screen Display compatibility profile for Splinter Cell Chaos Theory
- Updated profiles list​

 

Yep, I'll take some break now.

 

No, RTSS is not intended for it. No need to push this idea further please.

 

No, there is no "known bug" with that. GPU Boost basics: offset is static, base clock is not.

 

It moves up/down depending on temperature

 

Some development news:
We're discussing vector of development and possible new features for the next version of Afterburner with MSI now. No promises yet, but it is quite possible that we'll focus on implementation of async fan control for reference design RTX 20x0 series cards. To avoid confusion I'll explicitly mention that we're talking about reference design dual async fan solutions only (reference resign 20x0 fan controllers), so it won't work on third party custom design proprietary async fan control solutions (such as EVGA iCX cards).

 

RTSS version?