Rtss download

On my new Windows 10 desktop I wanted an on-screen display (OSD) of system performance overlay while gaming - similar those benchmarking videos one finds on YouTube. While I don’t overclock my CPU or GPU, I still do want to know the utilization and temperatures, as these give me a hint as to whether the system is thermal throttled or bottlenecked.Starting from knowing absolutely nothing, I started by trying a few tools recommended by various netizens. I started with MSI Afterburner, but removed it as I don’t like the skins, is missing temperature data for my chipset, and I’d prefer to avoid accidentally blowing up my system!In reality, all I needed was two pieces of free software:RivaTuner Statistics Server (RTSS) 7.3.0 Beta 6, which although is installed with MSI Afterburner, is actually a separate program.HWiNFO64 6.34-4300 which has support for my B550 / Ryzen 5000 combination.The usual disclaimers - don’t download from untrusted sources, and don’t blindly follow what random stranges say (I am, of course, referring to myself).There are three ways to setup the OSD, and all of them need RTSS running.HWiNFO sending data to RTSS, orRTSS using internal HAL data source (HWiNFO not required), orRTSS getting data from HWiFO.1. HWiNFO sending data to RTSSIn HWiNFO, Config > OSD (RTSS):Check Enable hotkey for toggling, and set the hotkey, I use Control-Shift-F12 Select the stats you want sent to RTSS, and for each, Check Show value in OSD and any other options (I like Units in superscript)Define the Position (Line and Column)Under the Custom tab, you can rename the elements, e.g. rename GPU Core Load to GPU - just to be consistent with the statistic called CPU which is actually the CPU temperature from the motherboard. Note that in RTSS, the OverlayEditor.dll should be disabled (more on that later).EntryLineColumnShow LabelColourTotal CPU Usage11PinkPhysical Memory Used12PinkCPU (Temperature)13YPinkGPU Core Load21GreenGPU Memory Allocated22GreenGPU (Temperature)23YGreenWhen running a DirectX game, hit the hotkey, and you’ll get something like this: The downside is you don’t have the flexibility to design your own layout. So, on to Option 2...2. RTSS using internal HAL data sourceIn RTSS - the

In this section, we describe a few methods and tools you can use to start and stop Real-Time (RTSS) processes, including: RTX64 Task Manager - Task Manager allows you to view, start, or end any active RTSS process or Windows process linked to RTX64 on your system, whether started by Task Manager or with the RtssRun command line utility. Windows processes linked to RTX64 are un-managed processes that make calls to the Real-time Interface (RTAPI) or managed processes that use the IntervalZero.RTX64 interface to communicate with RTSS processes. RtssRun - RtssRun can be used to run an RTSS application from a Windows command prompt. RtssKill - You can use the RtssKill utility to forcibly terminate a particular RTSS process. Running RTSS Processes You can run RTSS processes using one of these methods: Double-clicking on the binary in Windows Explorer Starting a task from the RTX64 Task Manager Using RtssRun from a command prompt on your target system The following instructions are only for using RtssRun from a command prompt. For more information on starting and stopping a task from the RTX64 Task Manager, see Task Manager. starting a task with RtssRun: Open a Command Prompt Type RtssRun, and then type the full path to the RTSS binary. For example, to run the SRTM binary provided with RTX64: RtssRun “C:\Program Files\IntervalZero\RTX64\bin\srtm.rtss” If RtssRun is successful, a tone will begin and the RTX Server console will open and display text. Upon completion, the SRTM process prints a summary and histogram of latency results. For more information on SRTM and other samples, see the RTX64 Program Examples. Stopping RTSS Processes If you must terminate a RTSS process, you can stop an RTSS application using RTX64 Task Manager or RtssKill from a command prompt on your target system: NOTE: You should only kill a

Statistic you want. You can use HWiNFO exclusively or mix data sources, e.g. here I use both HWiNFO and the Internal HAL: Back in the editor, hit Insert to add one layer per statistic to display. Layout everything the way you want it to be displayed, including graphs! To do this:Double click the newly inserted Text layer and in the Layer properties dialog.Hit the + button under Hypertext.In the dropdown (that says Framerate by default), select what statistic to show: Previously we displayed text by checking Add current value macro, but to display the statistic as a charts, check Add embedded graph instead. Select use custom template then hit ... to edit the graph properties, e.g. choose a graph style (bar chart, filled area chart or line chart) and set other properties: By overlaying charts for CPU usage, RAM usage percent and GPU1 usage one on top of the next, I landed up with this: You can use the menu options to set the z-order (i.e. “bring-to-front” or “send-to-back”) so that in my case, CPU is the top-most, followed by GPU, followed by RAM. The bottom most chart has a semi-transparent purple-ish background color set as well.ConclusionTo summarize the options:If all you need is simple statistics without temperatures, Option 1. RTSS using internal HAL data source is the most resource efficient, but a bit of work to layout.Option 2. HWiNFO sending data to RTSS is very simple, absolutely no fuss (with fewer options), though it may be less pretty.But, me, I prefer Option 3. RTSS getting data from HWiFO because it gives me a more complete set of statistics, and if you take the time to design the layout, you’ll get graphs to boot! However, Option 3 does consume the most memory. I did a non-scientific test of RAM usage by RTSS (RTSS, RTSS Encoder Server and RTSS Hooks Loader) and HWiNFO64. Strictly, since the statistics displayed were different, the results are not directly comparable, but for illustration:OptionRTSS processesHWiNFO64 process1. Internal HAL7.8 + 1.1 + 2.8 MB2. HWiNFO sending to RTSS4.4 + 1.1 + 2.8 MB25.3 MB3. RTSS getting from

RTX64 and Windows only. When running Maestro, of course, you must always use the former configuration, which is the default.After the restart, open the RTX64 Control Panel to configure RTX64 as follows:Configure the RTSS Subsystem > Change internal system behavior: Set startup type to Automatic. Ensure the HAL timer period is 100us, and the default thread time quantum is 0. Check Free the Stack on TerminateThread call and Use priority inversion.Configure the RTSS Subsystem > Configure Memory Allocation Behavior: Select Request from Windows. This is the default in RTX64 3.4 for Win 10, but you'll need to change it in RTX64 4.5 for Win 11. Leave the other settings unchanged. Note that this is contrary to IntervalZero's recommendation for deployed applications. However, Maestro has very modest memory requirements, allocates all memory at startup, and does not require any additional memory over the course of its operation.Configure the RTSS Subsystem > Configure the RTSS Watchdog Timer: Disable the watchdog timer.Configure the RTSS Subsystem > Configure Exception Support: Check Use Structured Exception Handling and Terminate the Faulting Process.Configure the RTSS Subsystem > Configure Power Management Settings: Ensure that Windows idle detection is disabled. In addition, click on Windows Power Management and verify that the system is configured with the RTX64-Recommended power plan, which should be created when RTX64 is installed. This power plan disables hibernation, sleep, and other power-saving features that could impact real-time performance. For more information, see the page Real-Time Subsystem > Hardware Considerations > Causes and Management of Interrupt