Chromaport

Laptops with Multiple GPUs

Many modern laptops have two GPUs:

  • Integrated GPU (iGPU) - built into the CPU and optimized for low power consumption.
  • Discrete GPU (dGPU) - a dedicated NVIDIA, AMD or Intel GPU designed for higher graphics performance.

Chromaport can work with both, but laptops with multiple GPUs can behave differently depending on how the manufacturer routes the internal and external displays.

Understanding the GPU and display topology can help you get the best performance and lowest latency from a Chromaport session.

How laptop GPUs work

A typical laptop might have an integrated GPU and a discrete GPU:

Laptop
├── Integrated GPU
│   └── Internal display
│
└── Discrete GPU
    └── 3D rendering / compute

The discrete GPU is normally much faster for gaming, 3D applications, video processing, and other GPU-intensive workloads.

However, having a discrete GPU does not necessarily mean that the laptop's internal display is connected directly to it.

This distinction is particularly important for Chromaport because there are three separate things to consider:

  1. The GPU rendering the application or game.
  2. The GPU connected to the display being captured.
  3. The GPU used by Chromaport for capture and video encoding.

These can all be the same GPU, or they can be different GPUs.

Hybrid graphics

Most laptops use some form of hybrid graphics. In this configuration, the integrated GPU may be responsible for the laptop's internal display while the discrete GPU renders demanding applications.

                 ┌───────────────┐
                 │ NVIDIA dGPU   │
                 │ Game rendering│
                 └───────┬───────┘
                         │
                         │ rendered frames
                         ▼
                 ┌───────────────┐
                 │ Intel iGPU    │
                 │ Display       │
                 └───────┬───────┘
                         │
                         ▼
                  Laptop display

The discrete GPU can therefore be doing the actual 3D rendering even though the display itself is connected to the integrated GPU.

This configuration is commonly used because the integrated GPU consumes less power when the discrete GPU is not needed.

Laptop manufacturers and GPU vendors may use different names for these configurations, including:

  • NVIDIA Optimus
  • NVIDIA Advanced Optimus
  • AMD Switchable Graphics
  • Hybrid Graphics
  • Dynamic Graphics
  • MSHybrid

Why GPU selection matters for Chromaport

The discrete GPU is generally much more powerful than the integrated GPU. It also typically has a more capable dedicated hardware video encoder. This makes it attractive for both running demanding applications and encoding the video stream sent by Chromaport.

However, selecting the discrete GPU for Chromaport is not always the fastest configuration.

When the display is connected to a different GPU than the one performing capture or encoding, frames may need to be transferred between GPUs. These transfers consume memory bandwidth and can add latency.

Example: game on the discrete GPU, display on the integrated GPU

Consider a laptop where the game runs on the discrete GPU but the internal display is connected to the integrated GPU:

Game
  │
  ▼
dGPU
  │
  │ rendered frame
  ▼
iGPU
  │
  ▼
Display
  │
  │ captured frame
  ▼
dGPU
  │
  │ encode
  ▼
Network

Conceptually, the rendered frame can cross from the discrete GPU to the integrated GPU for display, and the captured image may then need to become available to the discrete GPU again for processing or encoding.

The exact path depends on the laptop's hardware topology, Windows graphics stack, capture API, and driver implementation. Windows and the GPU driver can optimize parts of this process, so the diagram is a simplified representation rather than a guaranteed path on every laptop.

Nevertheless, additional GPU-to-GPU transfers can increase latency and consume bandwidth.

Why the discrete GPU can still be preferable

Discrete GPUs generally provide significantly more graphics performance and often have a more capable hardware video encoder than integrated GPUs.

For high-resolution or high-frame-rate streaming, this can make the discrete GPU substantially better at encoding the Chromaport stream.

As a result, there is a trade-off between keeping capture and encoding close to the display and using the more powerful GPU for encoding.

Which GPU should I choose for Chromaport?

There is no universal answer for laptops with multiple GPUs. The best choice depends primarily on which GPU drives the display being captured.

Display connected directly to the discrete GPU

If the display being captured is connected directly to the discrete GPU, selecting the discrete GPU for Chromaport is generally the preferred configuration.

Game
  │
  ▼
dGPU
  │
  ├── Display
  │
  └── Chromaport
          │
          ▼
       Encoder
          │
          ▼
       Network

In this configuration, rendering, display output, capture, and encoding can remain on the same GPU, minimizing unnecessary GPU-to-GPU transfers.

Display connected to the integrated GPU

If the display being captured is connected to the integrated GPU, the situation is more complicated.

If the game runs on the discrete GPU and Chromaport also runs on the discrete GPU, the frame may need to cross the GPU boundary during the capture process.

Game
  │
  ▼
dGPU
  │
  ▼
iGPU
  │
  ▼
Display
  │
  ▼
Chromaport
  │
  ▼
dGPU
  │
  ▼
Encode → Network

In this configuration, choosing the discrete GPU gives Chromaport access to its more powerful encoding hardware, but may introduce additional GPU-to-GPU transfers.

Choosing the integrated GPU instead can keep capture closer to the display, potentially reducing transfers, but may result in less capable encoding performance.

The best configuration therefore depends on the specific laptop and workload.

Recommended approach

As a general rule:

  • If the captured display is connected to the discrete GPU, use the discrete GPU for Chromaport.
  • If the captured display is connected to the integrated GPU, either GPU may be preferable. The discrete GPU can provide better encoding performance, while the integrated GPU may reduce GPU-to-GPU transfers.
  • If your laptop supports a MUX switch, using discrete-GPU-only mode can provide the cleanest topology for maximum performance and low latency.

If you are troubleshooting latency or performance, test the available configurations rather than assuming that the more powerful GPU is always the optimal choice.

Selecting the GPU used by Chromaport

Windows allows you to choose which GPU an application should prefer.

Windows Graphics settings

  1. Open Windows Settings.
  2. Go to System → Display → Graphics.
  3. Find Chromaport in the application list.
  4. Select Options.
  5. Choose High performance or Power saving.
  6. Click Save.
  7. Restart Chromaport.

Depending on your Windows version, the exact wording may differ slightly.

High performance normally selects the laptop's discrete GPU, while Power saving normally selects the integrated GPU.

You can use the same settings for applications you run through Chromaport, such as games or 3D applications.

Selecting the discrete GPU in BIOS or the laptop control application

Many gaming and performance-oriented laptops provide a hardware-level option to change the GPU mode.

Depending on the manufacturer, this may be called:

  • Discrete GPU
  • dGPU only
  • Dedicated GPU
  • NVIDIA GPU only
  • Ultimate
  • Discrete Graphics
  • MUX Switch

The exact options depend on the laptop.

Hybrid mode

iGPU ────────── Internal display
  │
  └── Windows desktop

dGPU ────────── Games / applications

Hybrid mode generally provides better battery life because the discrete GPU can be powered down when it is not needed.

Discrete GPU mode

dGPU ────────── Internal display
  │
  ├── Games
  ├── Applications
  └── Chromaport

In discrete-GPU mode, the display is connected directly to the discrete GPU. This can provide a more direct path for rendering, capture, and encoding.

Why can discrete GPU mode improve performance?

In hybrid mode, an application can render on the discrete GPU while the integrated GPU remains responsible for driving the display.

Game
  │
  ▼
dGPU
  │
  │ rendered frame
  ▼
iGPU
  │
  ▼
Display

This requires the rendered image to cross the GPU boundary before it reaches the display. Depending on the hardware and workload, this can introduce additional latency and memory traffic.

With a MUX switch or discrete-GPU-only mode, the path can instead be:

Game
  │
  ▼
dGPU
  │
  ▼
Display

This can reduce frame-copying and GPU-to-GPU transfers.

It can be particularly beneficial for:

  • High-refresh-rate gaming
  • Low-latency workloads
  • High-resolution displays
  • GPU-intensive applications
  • Remote streaming and capture

The exact performance improvement depends on the laptop and workload.

Trade-off

Discrete-GPU mode normally consumes more power.

If battery life is important, hybrid mode may be preferable.

If the laptop is plugged in and maximum performance and low latency are the priority, discrete-GPU mode may provide better results.

External monitors can change GPU routing

Connecting an external monitor can significantly change how a laptop's GPUs are used.

The external display may be physically connected to either the integrated GPU or the discrete GPU.

For example, a laptop might have:

Internal display
       │
       ▼
     iGPU

External HDMI
       │
       ▼
     dGPU

In this configuration, Windows may use the iGPU for the laptop's internal display while the external monitor is driven directly by the dGPU.

Another laptop may have both displays connected through the iGPU:

Internal display ──┐
                   ├── iGPU
External display ──┘

dGPU → rendering

The exact topology is determined by the laptop's hardware design.

Why external monitor routing matters for Chromaport

Chromaport captures a display, not simply "the fastest GPU". The GPU associated with the display you select can therefore affect the capture path.

For example:

GPU 0: Intel Graphics
    └── Laptop display

GPU 1: NVIDIA RTX
    └── External display

If Chromaport captures the external display, it may be interacting with a different GPU than when it captures the internal laptop display.

This can affect:

  • Capture performance
  • GPU utilization
  • Encoding performance
  • Frame latency
  • Which GPU is active during a session

How to determine which GPU drives a display

Windows provides several ways to inspect your display and GPU configuration.

Task Manager

Open Task Manager → Performance and look at the available GPUs.

Then open Processes and enable:

  • GPU
  • GPU engine

The GPU engine column can show which GPU an application is using.

Chromaport.exe          GPU 1 - 3D
Game.exe                GPU 1 - 3D
Desktop Window Manager  GPU 0 - 3D

This can help identify whether an application is using the integrated or discrete GPU.

Windows Display settings

Go to Settings → System → Display → Advanced display.

Select the display you want to inspect.

Windows can provide information about the selected display, including its refresh rate and display adapter.

Connecting a monitor directly to the discrete GPU

If your laptop's HDMI, DisplayPort, USB-C, or Thunderbolt output is wired directly to the discrete GPU, connecting an external monitor may automatically cause that monitor to use the dGPU.

                    Laptop

             ┌─────────────────┐
             │ NVIDIA dGPU     │
             └───────┬─────────┘
                     │
                  HDMI/DP
                     │
                     ▼
              External monitor


             ┌─────────────────┐
             │ Intel iGPU      │
             └───────┬─────────┘
                     │
                     ▼
               Internal display

This can be useful for Chromaport because the external monitor may have a direct path to the discrete GPU.

However, the physical port used matters. On some laptops, different USB-C, HDMI, or DisplayPort outputs can be connected to different GPUs.

Check your laptop manufacturer's documentation if you need to determine the exact topology.

USB-C and Thunderbolt displays

USB-C display connections can be more complicated.

A USB-C port may be connected to:

  • The integrated GPU
  • The discrete GPU
  • Both GPUs through a switching mechanism
  • A Thunderbolt controller that is itself connected to a particular GPU

As a result, two USB-C ports on the same laptop can behave differently.

If you are troubleshooting Chromaport performance, try another display output if your laptop provides one.

NVIDIA Optimus and Advanced Optimus

NVIDIA laptops can use different display-routing technologies.

Optimus

With traditional Optimus, the integrated GPU commonly remains responsible for the internal display while the NVIDIA GPU renders demanding applications.

Application
     │
     ▼
 NVIDIA GPU
     │
     ▼
 Intel GPU
     │
     ▼
Internal display

Advanced Optimus

Some laptops include Advanced Optimus, which allows the display connection to be dynamically switched between the integrated and discrete GPUs.

This can provide a compromise between battery life and performance.

When maximum performance is required, the internal display can be connected directly to the NVIDIA GPU.

MUX switches

A MUX switch is hardware that controls which GPU is connected to the laptop display.

A laptop with a MUX switch may offer modes such as:

ModeDisplayTypical advantage
HybridiGPUBetter battery life
DiscretedGPUMaximum performance
AutomaticChanges dynamicallyBalance between both

If your laptop supports a MUX switch, enabling discrete-GPU mode can be worth trying when using Chromaport for high-performance workloads.

The setting is usually available in the laptop manufacturer's control application or BIOS/UEFI.

Recommended configuration for Chromaport

Gaming or GPU-intensive workloads

  1. Connect the laptop to AC power.
  2. Check which GPU is connected to the display you want to capture.
  3. If the display is connected directly to the discrete GPU, set Chromaport toHigh performance in Windows Graphics settings.
  4. Set the game or application to High performance as well.
  5. If available, try Discrete GPU / dGPU-only / Ultimate mode for maximum performance and low latency.
  6. If using an external monitor, try connecting it to a display output that is directly connected to the discrete GPU.
  7. Restart applications after changing GPU settings.

If the captured display is connected to the integrated GPU, test bothHigh performance and Power saving for Chromaport if latency or capture performance is not as expected.

Battery-focused use

Keep the laptop in hybrid mode and allow Windows to select the appropriate GPU automatically.

This generally provides better battery life while still allowing demanding applications to use the discrete GPU.

If Chromaport performance is lower than expected

If you have a powerful NVIDIA or AMD GPU but Chromaport performance is unexpectedly low, check the following:

1. Which GPU drives the captured display?

First determine whether the display you are capturing is connected to the integrated GPU or the discrete GPU.

2. Is Chromaport using the expected GPU?

Check:

Windows Settings → System → Display → Graphics → Chromaport → Options

Try High performance or Power saving depending on the display topology.

3. Is the application using the expected GPU?

Check the application's GPU engine in Task Manager.

4. Is the laptop in hybrid mode?

If your laptop has a MUX switch, try Discrete GPU mode.

5. Is an external monitor connected?

The external monitor may be connected directly to the discrete GPU, potentially changing the optimal configuration.

6. Are you using the correct display output?

On some laptops, HDMI, DisplayPort, and USB-C ports can be connected to different GPUs.

Try another port if available.

7. Compare latency and GPU utilization

If both GPUs are available for Chromaport, compare the configurations while running the same workload. Look at capture latency, frame rate, GPU utilization, and encoding performance.

The discrete GPU is not automatically the best choice when the captured display is connected to the integrated GPU.

Understanding the GPU topology

The most important thing to remember is that GPU selection and display routing are two different things.

A laptop can have a topology where the integrated GPU drives the internal display while the discrete GPU handles demanding applications:

             ┌─────────────┐
             │   CPU       │
             │   + iGPU    │
             └──────┬──────┘
                    │
              Internal display
                    │
                    │
             ┌──────┴──────┐
             │   Windows   │
             └──────┬──────┘
                    │
             ┌──────┴──────┐
             │   dGPU      │
             └─────────────┘

Or the external display may be connected directly to the discrete GPU:

       iGPU ───── Internal display

       dGPU ───── External display
         │
         └──────── Games / applications

Or, with a MUX switch:

              ┌─────────┐
iGPU ─────────►│         │
              │   MUX   ├──── Display
dGPU ─────────►│         │
              └─────────┘

There is no single configuration that applies to every laptop.

For Chromaport, the important factors are:

  • Which GPU renders the application.
  • Which GPU drives the display being captured.
  • Which GPU Chromaport uses for capture and encoding.
  • Whether frames need to cross between GPUs.

When possible, keeping rendering, display output, capture, and encoding on the same GPU provides the simplest and potentially lowest-latency path. On hybrid laptops, however, this is not always possible, and the optimal configuration depends on the specific hardware and workload.