TL;DR
Get the latest gadgets delivered free with Prime
- Fast, free delivery on millions of items
- Prime Video, Amazon Music and more included
- Member-only deals all year
MindStudio reports that native Linux on an MSI Prestige RTX Spark laptop generated 13.9 tokens per second on GPT-OSS-120B, compared with 65.4 on Windows and 65.1 in WSL. The report links the gap to Linux’s inability to access the laptop’s encrypted power-management firmware; Nvidia has not provided a response or confirmed a fix in the supplied material.
A MindStudio report says native Linux inference on an MSI Prestige RTX Spark laptop reached 13.9 tokens per second on GPT-OSS-120B, far below the 65.4 tokens per second recorded on Windows and 65.1 in Windows Subsystem for Linux (WSL). The report attributes the gap to Linux being unable to access power-management firmware that, it says, keeps the GPU near 780 MHz instead of the roughly 2.2 GHz reported under Windows.
The reported test suggests that Linux can run on the laptop, with GPU acceleration and a working desktop after configuration, but does not achieve comparable inference performance in the tested setup. MindStudio says the GPU clock remained around 790 MHz during a longer Linux benchmark, while Windows held it near 2.2 GHz. Temperatures reportedly stayed between 40°C and 47°C, which the report cites as evidence that heat was not causing the slowdown.
For GPT-OSS-120B, MindStudio recorded 65.4 tokens per second on Windows, 65.1 on WSL and 61 on Nvidia’s DGX Spark desktop. Native Linux on the laptop produced 13.9 tokens per second. In a separate vLLM test with eight concurrent clients, the report gives 127 tokens per second for WSL and 39.5 for native Linux. These are results from the reported configurations, not a broad, independently verified comparison across RTX Spark laptops or Linux distributions.
The report identifies Nvidia’s Platform Controllers and Framework as the suspected source of the difference. It says this Windows-only component manages power delivery between the CPU and GPU, while an encrypted firmware power table is not readable by the Linux setup tested. MindStudio says WSL retains Windows’ driver stack and nearly matches Windows performance. It also describes a separate memory configuration issue: the laptop initially made only 28GB of its 128GB unified memory visible to Linux until the GPU reservation was changed in Windows.
Linux Performance Gap for AI Workloads
The reported results matter to developers considering these laptops for local AI inference. A machine that can run Linux and CUDA may still deliver a much lower throughput in native Linux if its power-management controls are unavailable. That could affect whether developers can use Linux-first tools directly on the hardware or need to keep Windows installed.
WSL appears to offer a practical workaround in MindStudio’s tests: it provides a Linux environment while retaining the Windows driver and reported GPU boost. That is not the same as resolving native Linux support, however. Developers may have different requirements for drivers, system administration and deployment, and the source does not establish whether the same performance gap occurs across other RTX Spark models or workloads.
Products Worth Considering
Apple MacBook Air 15-inch with M5 chip, 24GB RAM, 512GB SSD, Midnight
Logitech MX Mechanical Wireless Illuminated Keyboard, Tactile Quiet, Graphite
EcoFlow Delta 3 Max Portable Power Station – 2048 Wh LiFePO4 Battery, 2400 W Output (3400 W X‑Boost)
As an affiliate, we earn on qualifying purchases.
How the Laptop Runs Linux
The RTX Spark laptop platform appears in systems including MSI Prestige, Asus ProArt and Dell XPS models, according to the report. MindStudio says the laptops use the same GPU design as Nvidia’s DGX Spark desktop but ship with Windows on ARM, while the DGX Spark is sold with Nvidia’s Linux build. The shared CPU-GPU memory, with configurations up to 128GB, is part of the laptops’ appeal for AI development.
Installing Linux on the tested MSI system required workarounds. MindStudio recounts trying Debian and Omarchy, a distribution based on Arch, and describes initial problems with networking, display output and built-in input devices. After using a kernel with Nvidia’s open-source driver and changing the memory allocation from Windows, the report says Linux could run a desktop at 4K and 120Hz with audio, keyboard, touchpad and CUDA acceleration. The account documents one system and configuration; it does not establish that every listed laptop behaves identically.
““Windows produced 65.4 tokens per second” while “native Linux on the laptop” reached “13.9.””
— MindStudio report
Firmware Access and Model Coverage
The supplied material contains no response from Nvidia and does not confirm that the encrypted power table is the definitive cause of the performance gap. That explanation is MindStudio’s account of the system behavior; the source does not provide Nvidia documentation or an independent technical analysis of the firmware.
It is also unclear whether Nvidia plans to add equivalent power-management support to its Linux driver, when that might happen, or whether the restriction applies in the same way to other RTX Spark laptop models. The report’s benchmarks are tied to its stated hardware and software setup. It does not supply enough detail here to establish how results vary with Linux kernel versions, driver releases, model quantization or other inference settings. Its statements about behavior on battery are also incomplete in the supplied text.
Nvidia Linux Driver Support
A direct next step would be for Nvidia to clarify whether its Linux driver can support the platform controller’s power-management functions and whether any change is planned. MindStudio says a fix would depend on Nvidia porting that logic into its Linux driver, but the supplied source does not give a company timetable or commitment.
Until there is a documented driver update or further testing, users should treat the reported figures as a single-source benchmark rather than a guaranteed result for every RTX Spark laptop. Further comparisons across models, Linux configurations and workloads would help establish how widespread the limitation is. For now, the report’s tested alternatives are native Linux with reduced performance or WSL using Windows’ driver stack.
Key Questions
How fast was native Linux in the reported test?
MindStudio recorded 13.9 tokens per second on GPT-OSS-120B with native Linux on its tested MSI Prestige RTX Spark laptop.
How did Windows and WSL compare?
The report gives 65.4 tokens per second for Windows and 65.1 for WSL on the same model test. Those figures are from MindStudio’s setup and are not independently confirmed in the supplied material.
What does the report say causes the slowdown?
It attributes the gap to Linux’s lack of access to a Windows-only power-management component and an encrypted firmware power table. Nvidia has not confirmed that explanation in the source material.
Can Linux run on the RTX Spark laptop?
MindStudio says it got Linux running with GPU acceleration and working laptop hardware after configuration and workarounds. The report describes installation as difficult, and its experience concerns a specific MSI model.
Is there a confirmed fix?
No fix or release date is confirmed in the supplied material. MindStudio says improved native Linux performance would require Nvidia to add the platform power-management logic to its Linux driver.
Source: mindstudio.ai
