TL;DR
Get business pricing on monitors, keyboards and dev gear
- Business-only prices and quantity discounts
- Tax-exempt purchasing
- Multiple users, one account, clear invoices
MindStudio reports that an MSI Prestige RTX Spark laptop running native Linux produced 13.9 tokens per second on GPT-OSS-120B, compared with 65.4 on Windows. The report attributes the gap to Linux lacking access to a Windows-only power-management component and encrypted firmware tables; WSL ran near Windows speed. The measurements and proposed cause come from the report, and Nvidia has not provided a response in the supplied material.
A MindStudio report says native Linux inference on an MSI Prestige RTX Spark laptop ran at 13.9 tokens per second on GPT-OSS-120B, compared with 65.4 on Windows and 65.1 in Windows Subsystem for Linux (WSL). The report links the slowdown to Linux’s inability to use a Windows-only power-management component, which it says leaves the GPU at substantially lower clock speeds when the laptop is plugged in.
MindStudio says the laptop can run Linux with a desktop, CUDA acceleration, and functioning audio, keyboard, and touchpad after configuration. But in its GPT-OSS-120B test, native Linux achieved 13.9 tokens per second. Windows reached 65.4 tokens per second, WSL 65.1, and Nvidia’s DGX Spark desktop 61. Those results are measurements reported by MindStudio, not independently verified figures in the material provided here.
The report says the GPU ran at about 780 to 790 MHz on Linux while connected to wall power, compared with roughly 2.2 GHz on Windows. It attributes the difference to Nvidia Platform Controllers and Framework, a Windows-only component that manages power delivery between the CPU and GPU. MindStudio says the firmware power table is encrypted and that Linux currently cannot read it.
MindStudio also reports a separate memory-configuration obstacle: the machine initially reserved 96GB of its 128GB unified memory for the GPU, leaving Linux able to see only 28GB. The report says changing the allocation in Windows to reserve 4GB for the GPU allowed Linux to see 120GB. This setting was not available through Linux in the reported setup.
Linux Performance Falls Far Behind
The reported gap matters to developers considering these laptops for local AI inference. The RTX Spark machines are marketed around a GPU and unified-memory configuration suited to running large models, while many AI tools, including vLLM and SGLang, are designed primarily for Linux. If the reported power limit applies to other configurations, installing Linux directly could sharply reduce performance despite the hardware being capable of faster results under Windows.
The comparison also shows why WSL may be a practical alternative for now. In MindStudio’s tests, WSL delivered near-Windows throughput while providing a Linux environment and access to the Windows driver stack. That does not establish that WSL will meet every developer’s needs, but it suggests users may not have to choose between Linux-oriented tools and the laptop’s reported full-speed GPU operation.
Popular right now
Dyson Car+Boat Cordless Handheld Vacuum
TCL 65-Inch QM6K QLED Mini-LED 4K HDR Smart Google TV, 144Hz
iRobot Roomba 105 Robot Vacuum with LiDAR Navigation
As an affiliate, we earn on qualifying purchases.
Linux Setup Needs Windows Changes
The RTX Spark platform appears in laptops from manufacturers including MSI, Asus, and Dell. MindStudio says the laptops use the same GPU design as Nvidia’s DGX Spark desktop, but the laptops ship with Windows on ARM, while the DGX Spark ships with Nvidia’s Linux build. The shared GPU design does not, by itself, mean that the laptop’s power controls work the same way under both operating systems.
Installing Linux was not plug-and-play in the reported MSI test. MindStudio describes trying multiple distributions and needing hardware support, installation workarounds, and a Windows-side memory adjustment. After setup, it reported a usable desktop at 4K and 120Hz with CUDA acceleration. Its performance tests then showed the central limitation: Linux could run inference, but at much lower throughput in the plugged-in configuration.
“Windows produced 65.4 tokens per second. WSL … hit 65.1 tokens per second. Native Linux on the laptop … managed only 13.9.”
— MindStudio report
Cause and Scope Need Verification
The supplied material contains one publication’s testing and explanation; it does not include a response from Nvidia, an independent reproduction, or a detailed account of the exact hardware and software versions used. It is also unclear whether the same performance difference occurs on every RTX Spark laptop, Linux distribution, kernel, or driver combination.
The report attributes the slowdown to encrypted firmware power tables and missing Linux support for the platform controller. That explanation is reported rather than confirmed by Nvidia in the material provided. The source text also begins to discuss a change in performance when the laptop is unplugged, but the account is cut off before giving the result. The battery-powered behavior therefore cannot be described from the available information.
Driver Support Is the Open Question
MindStudio says a full fix would require Nvidia to bring the platform controller’s power-management logic into its Linux driver. The supplied source does not state that Nvidia has announced such work or provide a release date. Until a driver or firmware change is documented and tested, the report presents WSL as the current workaround for users seeking Linux tools alongside Windows-level GPU performance.
Further testing across laptop models and Linux configurations would help establish how broadly the reported clock limit applies. Users evaluating the machines should also check whether their intended memory allocation and operating-system setup require changes made from Windows.
Key Questions
How fast was native Linux in the reported test?
MindStudio measured 13.9 tokens per second on GPT-OSS-120B on native Linux, compared with 65.4 on Windows and 65.1 in WSL. These are figures from the report, not independent test results.
What does the report say causes the slowdown?
It attributes the gap to a Windows-only Nvidia power-management component and firmware tables that Linux cannot read in the tested setup. Nvidia has not confirmed that explanation in the supplied material.
Does Linux work on the RTX Spark laptop?
MindStudio says Linux could run on its MSI Prestige test laptop with a desktop, CUDA acceleration, and working basic peripherals after configuration. Its main reported problem was lower plugged-in GPU performance, not an inability to run Linux.
Is WSL a workaround?
In MindStudio’s benchmark, WSL reached 65.1 tokens per second, close to Windows at 65.4. The report says WSL uses the Windows driver stack, but the results do not establish performance across every workload or device.
Has Nvidia announced a Linux fix?
No Nvidia announcement is included in the supplied source. MindStudio says a fix would require Linux driver support for the platform controller’s power-management logic.
Source: mindstudio.ai
