What’s actually slowing this PC down?
Pick the symptom - the matching free tool is one click away.
Yes—on a compatible hybrid-graphics computer, an AI application can use the discrete GPU while a display connected to the integrated GPU handles scanout. The key is to distinguish the GPU doing the compute from the GPU physically driving the display. Choose the discrete GPU for the AI application, then check the computer’s wiring or mux support separately if you need to change which GPU owns a display output.
Compute and display output are separate decisions
A computer can render or compute on one GPU and send the resulting frames to another GPU’s display pipeline. NVIDIA’s Optimus Developer Guide describes this arrangement when a display is connected to the integrated GPU (IGP): the NVIDIA GPU renders, then transfers the finished frames to the IGP for scanout. NVIDIA also says a CUDA application can find and create a context on the NVIDIA GPU even when it is not the display device.
As an Amazon Associate I earn from qualifying purchases.
That does not mean a software preference can change the electrical route of an HDMI, DisplayPort, USB-C, or built-in panel connection. A GPU preference asks which adapter an application should use; the computer’s motherboard, port wiring, and any display mux determine which GPU drives a particular output.
Choose the discrete GPU for the AI application
Windows
On Windows, set a per-application graphics preference where the option is available. Microsoft’s Surface Book 3 guidance gives this example path: Settings > System > Display > Graphics Settings. Add or select the desktop application, choose Options, then select the high-performance GPU if it identifies the expected discrete GPU. Labels and availability can vary by PC and Windows version; the cited path is Surface-specific.
#1 Best Overall
- AI Performance: 767 AI TOPS
- OC mode: 2632 MHz (OC mode)/ 2602 MHz (Default mode)
- Powered by the NVIDIA Blackwell architecture and DLSS 4
- Axial-tech fan design features a smaller fan hub that facilitates longer blades and a barrier ring that increases downward air pressure
- A 2.5-slot design maximizes compatibility and cooling efficiency for superior performance in small chassis
Microsoft describes DXGI_GPU_PREFERENCE_HIGH_PERFORMANCE as a preference for the highest-performing GPU, such as a discrete or external GPU, and DXGI_GPU_PREFERENCE_MINIMUM_POWER as a preference for the minimum-powered GPU, such as an iGPU. These are preferences, not guarantees that every application will use a particular adapter. See Microsoft’s DXGI GPU preference documentation.
Linux with NVIDIA PRIME
For Linux systems using NVIDIA PRIME render offload, use the desktop’s discrete-GPU launch option when available, or identify and select a GPU with switcherooctl. NVIDIA’s PRIME Render Offload guide documents switcherooctl list for listing available GPUs and their indices, and switcherooctl launch -g <index> <command> for starting a process on a selected GPU.
Rank #2
- Powered by the NVIDIA Blackwell architecture and DLSS 4
- Powered by GeForce RTX 5070 Ti
- Integrated with 16GB GDDR7 256bit memory interface
- PCIe 5.0
- WINDFORCE cooling system
NVIDIA also documents API-specific environment variables for its Intel-plus-NVIDIA example: __NV_PRIME_RENDER_OFFLOAD=1 enables NVIDIA render offload; __GLX_VENDOR_LIBRARY_NAME=nvidia selects NVIDIA for an OpenGL context; and __VK_LAYER_NV_optimus=NVIDIA_only selects NVIDIA for Vulkan device selection. The right mechanism depends on the application and graphics API, so do not treat one variable combination as universal. For CUDA AI workloads, check the framework’s device report rather than relying on a graphics-renderer check.
Find out which GPU owns the display connection
Check the computer maker’s documentation or graphics-control software to determine whether the internal panel and each external port are connected to the iGPU or discrete GPU. NVIDIA’s Optimus guide notes that display heads may be connected to either GPU. If a display is wired to the iGPU, the dGPU can render and pass frames to the iGPU for scanout. If a port is wired directly to the dGPU, selecting the iGPU for an application does not reroute that port’s physical signal.
Rank #3
- Powered by the NVIDIA Blackwell architecture and DLSS 4
- Powered by GeForce RTX 5060
- Integrated with 8GB GDDR7 128bit memory interface
- PCIe 5.0
- WINDFORCE cooling system
Some laptops include a hardware display mux that can switch the internal panel between GPUs. Whether it can be controlled by Windows depends on platform and driver support; a BIOS option, NVIDIA Control Panel setting, or Windows preference should not be assumed to exist on every computer.
When Windows can switch the internal panel
Microsoft’s Automatic Display Switch (ADS) support is optional. Microsoft documents it beginning with Windows 11 version 24H2 update 2025.01D (WDDM 3.2), but Windows version alone does not establish that a laptop supports it. The feature requires coordinated support from Windows, the platform’s ACPI/mux firmware, and both GPU drivers. The initial ADS implementation applies to the laptop’s internal panel, not external connectors. See Microsoft’s Automatic Display Switch documentation.
Rank #4
- Powered by the NVIDIA Blackwell architecture and DLSS 4. System Requirements: Minimum 850W PSU with 16-pin 12V-2x6 (12VHPWR) connector required. Verify before purchasing.
- Military-grade components deliver rock-solid power and longer lifespan for ultimate durability. Compatibility: 348mm (13.7") length, 3.6 slots, 4.3 lbs. Confirm case clearance and slot spacing. GPU bracket included.
- Protective PCB coating helps protect against short circuits caused by moisture, dust, or debris
- 3.6-slot design with massive fin array optimized for airflow from three Axial-tech fans
- Phase-change GPU thermal pad helps ensure optimal thermal performance and longevity, outlasting traditional thermal paste for graphics cards under heavy loads
Verify the AI workload is using the intended GPU
- Confirm both adapters are detected. Check the operating system’s device list and confirm their drivers are loaded. NVIDIA’s Linux guide demonstrates
lspcifor identifying Intel and NVIDIA adapters. - Set the application preference or launch method. Use Windows Graphics settings or the Linux desktop/PRIME method appropriate to the application.
- Check while the workload is running. On Windows, open Task Manager and inspect the application’s GPU Engine column. On Linux, verify CUDA device selection in the AI framework itself. NVIDIA’s PRIME guide also demonstrates
glxinfo -Bfor checking an OpenGL renderer and a runtime power-state check under/sys/bus/pci/devices/.../power/runtime_status; those graphics checks alone do not prove that an AI process selected CUDA. - Investigate display routing separately. If the question is which GPU drives a panel or port, consult the computer’s connection documentation and mux controls. Changing the AI application’s adapter preference does not change that wiring.
Check AI software requirements independently
GPU selection and display routing do not establish that a particular AI application will run on a given GPU. Check that application’s requirements for supported GPU architecture, driver, framework, and memory. The NVIDIA CUDA documentation cited here establishes device discovery behavior, not the requirements of every AI tool. If the software specifically requires CUDA, confirm that the discrete GPU and its installed software stack meet the application’s stated requirements.
Quick Recap
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
