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’s testing of an MSI Prestige RTX Spark laptop found GPT-OSS-120B inference at 13.9 tokens per second under native Linux, compared with 65.4 on Windows and 65.1 in WSL. The report attributes the gap to GPU clocks staying near 780 MHz on plugged-in Linux, with encrypted power-management firmware that Linux cannot currently read. Nvidia has not provided a response or a timetable for a Linux fix in the supplied material.
Native Linux inference on an MSI Prestige RTX Spark laptop ran at 13.9 tokens per second in MindStudio’s test of GPT-OSS-120B, compared with 65.4 on Windows and 65.1 in Windows Subsystem for Linux (WSL). The report links the performance gap to the GPU running at about 780 MHz on plugged-in Linux, far below its roughly 2.2 GHz Windows clock, and says encrypted power-management firmware prevents Linux from accessing the needed settings.
MindStudio said the laptop could boot Linux and, after configuration, run a desktop with GPU acceleration, 4K output at 120Hz, audio, keyboard and touchpad. The setup was not plug-and-play: the report describes installation and hardware-support problems that required workarounds, and says the laptop’s memory allocation had to be changed through Windows before Linux could access most of its unified memory.
In the reported GPT-OSS-120B test, Windows reached 65.4 tokens per second, WSL reached 65.1, and Nvidia’s DGX Spark desktop reached 61. Native Linux on the laptop managed 13.9. MindStudio also reported a separate vLLM test with eight concurrent clients: WSL reached 127 tokens per second, while native Linux reached 39.5. The report gives no full benchmark methodology in the supplied material, so the figures should be read as results from its tests, not as independently verified performance guarantees.
MindStudio said a longer run took about eight minutes on Windows and 37 minutes on Linux. It reported GPU clocks near 2.2 GHz under Windows and around 790 MHz during the Linux run, with temperatures between 40 and 47°C. The report says those temperatures suggest the slowdown was not caused by thermal throttling, but instead by the GPU not being allowed to run at full speed.
Linux Developers Face a Speed Penalty
The results matter to AI developers considering these laptops because many inference and model-serving tools, including vLLM and SGLang, are commonly used in Linux environments. The measurements suggest that installing Linux directly could sharply reduce throughput on this hardware, despite the system’s ability to run Linux and access the GPU.
For users who need Linux tools but want performance close to Windows, MindStudio identifies WSL as a practical workaround: it runs a Linux environment while retaining the Windows driver stack. In the reported GPT-OSS-120B test, its throughput was nearly the same as Windows. That is a result from this particular laptop and workload, not proof that WSL will match native Linux across all software or configurations.
Products Worth Considering
Apple MacBook Air 13.6-inch with M5 Chip, 16GB Memory, 512GB SSD, Sky Blue
Samsung Galaxy Tab A11+ (6 GB RAM, 128 GB storage, expandable) – Gray
Bose SoundLink Flex Bluetooth Speaker (2nd Gen) – Portable Outdoor Speaker with Hi‑Fi Audio, Waterproof and Dustproof, USB‑C, Up to 12 Hours
As an affiliate, we earn on qualifying purchases.
Why the Windows Driver Matters
The RTX Spark laptops described in the report include systems from MSI, Asus and Dell. MindStudio says the machines use a GPU design related to Nvidia’s DGX Spark, but ship with Windows on ARM, while the DGX Spark desktop is supplied with Nvidia’s Linux build. That difference makes the laptop’s platform and driver support relevant to Linux performance.
According to MindStudio, a Windows-only component called the Nvidia Platform Controllers and Framework manages power delivery between the CPU and GPU. The report says its power table is encrypted and only the Windows driver can decrypt it, leaving Linux without an equivalent way to manage GPU power. It also describes a separate memory-configuration issue: the laptop initially reserved 96GB of unified memory for the GPU, and the setting to alter that split was available through Windows.
““Windows produced 65.4 tokens per second” while “native Linux on the laptop” reached “13.9.””
— MindStudio report
Linux Power Support Remains Unclear
The supplied source material does not include a response from Nvidia, confirmation from the laptop makers, or a schedule for adding Linux support for the platform controller’s power-management functions. It is not clear whether a driver update, firmware change or another configuration would resolve the clock restriction.
The report also notes that Linux performance changes with the laptop’s power source, but the supplied text cuts off before giving the unplugged result. The exact behavior on battery therefore cannot be established from the available material. The benchmark results are attributed to MindStudio; independent replication and details such as software versions and test settings are not included in the source text provided.
Driver Changes Would Set the Pace
MindStudio says a lasting fix depends on Nvidia porting the platform controller’s power-management logic into its Linux driver. Until Nvidia or the laptop manufacturers publish more information, the status and timing of any fix remain unknown. Linux users weighing a purchase or dual-boot setup can treat WSL as the report’s current performance-preserving option, while recognizing that its results may differ by workload and system configuration.
Key Questions
How fast was native Linux in the reported test?
MindStudio measured 13.9 tokens per second running GPT-OSS-120B on native Linux on the tested MSI Prestige RTX Spark laptop. Windows reached 65.4 tokens per second in the same reported comparison.
What does the report say causes the slowdown?
It attributes the gap to the GPU running at about 780 MHz on plugged-in Linux, compared with roughly 2.2 GHz on Windows. MindStudio says encrypted power-management firmware is accessible through the Windows driver but lacks a Linux equivalent.
Does Linux run on the RTX Spark laptop?
Yes. MindStudio reports that Linux booted and ran a desktop with GPU acceleration and working core laptop features after configuration and workarounds. It says installation and hardware setup were not straightforward.
Is WSL a workaround?
In MindStudio’s GPT-OSS-120B test, WSL reached 65.1 tokens per second, close to Windows at 65.4. WSL runs Linux within Windows and uses the Windows driver stack, but that result applies to the reported test rather than every workload.
Has Nvidia announced a Linux fix?
No Nvidia response or release timetable appears in the supplied source material. MindStudio says a fix would require Nvidia to bring the platform controller’s power-management logic to its Linux driver.
Source: mindstudio.ai
