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
A MindStudio report measured 13.9 tokens per second for GPT-OSS-120B on native Linux on an MSI Prestige RTX Spark laptop, compared with 65.4 on Windows and 65.1 in WSL. The report attributes the gap to low Linux GPU clock speeds and a Windows-only Nvidia power-management component; Nvidia has not provided a response in the supplied material.
A MindStudio report says native Linux ran GPT-OSS-120B at 13.9 tokens per second on an MSI Prestige RTX Spark laptop, compared with 65.4 on Windows and 65.1 in Windows Subsystem for Linux (WSL). The report links the slowdown to Linux GPU clocks staying near 790 MHz, well below the roughly 2.2 GHz recorded on Windows, and says a Windows-only Nvidia power-management component is involved.
The measurements in the report show a large performance gap on the tested laptop. Windows reached 65.4 tokens per second on GPT-OSS-120B, while WSL reached 65.1. MindStudio reported 13.9 tokens per second for native Linux, and 61 for Nvidia’s DGX Spark desktop system. These are results from the report’s tests, not independent verification or a general benchmark of every RTX Spark laptop or workload.
MindStudio also described a separate test using vLLM with eight concurrent clients: WSL reached 127 tokens per second and native Linux reached 39.5. In a longer run, the report said Windows completed the task in about eight minutes, with the GPU near 2.2 GHz, while Linux took 37 minutes at roughly 790 MHz. Recorded temperatures of 40 to 47°C were cited as evidence against thermal throttling as the cause in that test.
The report says Linux can be installed and used on the MSI laptop, including with Nvidia’s open kernel driver, but setup required workarounds. After configuration, it reported a working desktop, CUDA acceleration, 4K output at 120Hz, and functioning audio and input devices. The article identifies power management, rather than the basic ability to boot Linux, as the main performance problem.
Linux Users Face a Performance Gap
The results matter to developers who buy these laptops for local AI work and prefer Linux-based tools. In MindStudio’s tests, native Linux delivered a small fraction of the Windows and WSL throughput on the same laptop, potentially lengthening model runs and reducing the practicality of using Linux directly for inference. The figures apply to the workloads and setup described in the report; they do not establish that every application will see the same slowdown.
For now, the report presents WSL as a workaround: it runs a Linux environment on top of Windows and, in the reported tests, retained performance close to Windows. That may suit developers whose tools run in WSL, but it is not the same as running Linux natively. The distinction is relevant for users who want a Linux-only installation, need particular hardware access, or want to avoid relying on Windows for the driver and power-management stack.
Products Worth Considering
Apple MacBook Air 13-inch with M5 chip, 16GB memory, 512GB SSD, and 12MP Center Stage camera
Apple MacBook Air 15-inch with M5 chip, 16GB Memory, 512GB SSD, and Liquid Retina Display
Samsung Galaxy Tab A11+ 11-Inch Tablet, 6GB RAM, 128GB Storage, Expandable, Gray
As an affiliate, we earn on qualifying purchases.
A Windows-Based RTX Spark Platform
The RTX Spark laptop platform appears in systems including the MSI Prestige, Asus ProArt and Dell XPS, according to the source article. These laptops pair Nvidia GPU hardware with Windows on ARM and unified memory shared by the CPU and GPU, with configurations described as reaching up to 128GB. Nvidia’s DGX Spark desktop, by contrast, ships with Nvidia’s own Linux build.
MindStudio reports that the MSI unit initially exposed only 28GB of its 128GB unified memory to Linux because the firmware reserved 96GB for GPU use. The article says the allocation could be adjusted in Windows; setting the GPU reservation to 4GB allowed Linux to see 120GB. The report also describes installation challenges involving Wi-Fi, display output and laptop-specific support, which required alternative hardware or software configurations.
According to MindStudio, the power issue involves the Nvidia Platform Controllers and Framework, which manages power delivery between the CPU and GPU. The report says its power table is encrypted and can be read by the Windows driver, while Linux lacks an equivalent mechanism. That explanation is the source’s account of the cause, rather than a publicly documented Nvidia confirmation in the material provided.
“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 Linux Fix Remain Unconfirmed
The supplied material does not include a response from Nvidia, independent replication of the benchmarks, or a confirmed timeline for a Linux driver change. The explanation involving encrypted firmware and the Platform Controllers and Framework comes from MindStudio’s reporting; the source does not provide a public Nvidia statement confirming that diagnosis.
The report’s excerpt also ends while describing how performance changes when the laptop is unplugged. It is therefore not possible from the supplied material to state the complete battery-versus-wall-power results. Testing details such as software versions, full system configuration and whether other RTX Spark laptop models behave the same way are also not fully established here.
Nvidia Driver Support Is the Next Test
The next significant development would be an Nvidia statement, driver release or documented Linux support for the platform’s power controls. MindStudio says a fix would require bringing the power-management logic into Nvidia’s Linux driver, but the supplied report gives no release date or confirmation that such work is underway.
Until a fix or further testing is available, Linux users considering these laptops can treat the reported figures as a warning about native inference performance, not as a guarantee for every device or workload. Additional benchmarks across models, power states and software versions would help establish how broad the issue is and whether later driver changes alter the results.
Key Questions
How fast was native Linux in the reported test?
MindStudio measured 13.9 tokens per second on GPT-OSS-120B on an MSI Prestige RTX Spark laptop running native Linux. The same report recorded 65.4 on Windows and 65.1 in WSL.
Does Linux run on the RTX Spark laptop?
According to MindStudio, Linux can boot and provide a desktop, GPU acceleration and working laptop features after configuration. The report describes installation hurdles, including device support and memory allocation, so it was not a plug-and-play setup.
What does the report say causes the slowdown?
MindStudio attributes it to Linux GPU clocks staying around 790 MHz rather than reaching roughly 2.2 GHz as in its Windows test. It says a Windows-only Nvidia power-management component handles encrypted power-table information that Linux cannot currently access. Nvidia confirmation is not included in the supplied material.
Is WSL a workaround?
In the reported tests, WSL reached performance close to Windows: 65.1 tokens per second versus 65.4 on GPT-OSS-120B. WSL runs Linux tools within Windows and uses the Windows driver stack, so it is an alternative to native Linux rather than a fix for native Linux support.
Has Nvidia announced a fix?
No Nvidia fix or timetable is confirmed in the supplied source material. MindStudio says Linux support for the platform’s power-management logic would be needed, but that is the report’s assessment, not a confirmed Nvidia commitment.
Source: mindstudio.ai
