We conducted our tests using a Windows 11 Pro desktop PC with the following specifications:
Each thermal compound was applied identically, and Cinebench was used to stress test the CPU under identical conditions. We measured temperature, power consumption, and clock speeds to evaluate thermal performance.
The temperature graph includes Noctua NT-H2 as a baseline for comparison, alongside the five PTM7950 samples. PTM7950 AliExpress No. 4 (brown line) consistently recorded the highest temperatures throughout the test. In contrast, AliExpress No. 1 (green) and No. 3 (pink) maintained the lowest overall temperatures, closely followed by the eBay sample (purple).
AliExpress No. 2 (white) and Noctua NT-H2 (yellow) performed similarly, landing in the middle range of the group. All samples demonstrated relatively stable thermal performance, but the spread between the lowest and highest readings highlights the potential variability depending on the source.
Across all tested thermal compounds, including the PTM7950 samples and Noctua NT-H2, CPU package power consumption remained nearly identical. The lines on the graph overlap closely, with all configurations hovering around 65W throughout most of the test duration.
While there were some minor deviations—particularly during transitions such as the start of the second Cinebench pass—these were brief and not unique to any one sample. Overall, no meaningful difference in power draw was observed between any of the PTM7950 sources or the Noctua baseline.
The core clock speed graph shows all samples generally maintained similar average frequencies throughout the test. Noctua NT-H2, PTM7950 AliExpress No. 2 (white), and No. 1 (green) recorded the highest and most consistent core clock averages, especially during peak load near the second Cinebench pass.
PTM7950 eBay (purple), AliExpress No. 3 (pink), and No. 4 (brown) trailed slightly behind, with AliExpress No. 4 showing the lowest sustained clock speeds of the group.
While there are some fluctuations in all samples, the differences were modest overall and largely followed the temperature patterns observed earlier.
CPU utilisation remained consistently high across most of the PTM7950 samples and the Noctua NT-H2 baseline. All compounds maintained close to 100% usage throughout the test, with only minor and temporary dips occurring during transitional moments like the second Cinebench pass.
Notably, the Noctua NT-H2 sample showed the most pronounced drop during this interval, whereas PTM7950 AliExpress No. 4 (brown) maintained nearly flat utilisation throughout. The remaining samples—eBay, and AliExpress No. 1, 2, and 3—clustered tightly together, exhibiting brief and relatively shallow drops.
Overall, CPU utilisation differences were small and likely not impactful under typical conditions.
PTM7950 samples showed colour variations, with the eBay and AliExpress No. 3 samples having an identical greenish-grey hue, while the AliExpress No. 1, No. 2 and No. 4 samples exhibited a consistent light grey shade.
To better understand whether the visual differences between PTM7950 samples have any correlation with performance, we compared the thermal results of one greenish-grey sample (eBay) and one light-grey sample (AliExpress No. 1).
The graph reveals that the greenish-grey (eBay) and light-grey (AliExpress No. 1) samples performed similarly overall, with AliExpress No. 1 holding a very slight edge in sustained thermal behaviour. While the difference was small—typically within 1°C—it was consistent across the duration of the test.
This suggests that colour variation alone is not a definitive indicator of performance, though the light-grey sample in this specific case edged out slightly better results.
This test highlights noticeable differences in thermal performance between various PTM7950 samples, even among those that look identical. While some samples performed very similarly to each other—and to Noctua NT-H2—others showed a measurable variation in temperature under sustained load.
Despite the visual similarities, performance varied both across colour categories and among samples of the same shade. CPU power consumption, clock speed, and utilisation remained mostly consistent, with temperature being the primary differentiator.
These results reflect short-term performance under controlled conditions and represent only a small selection of what is available on the market. We will be revisiting the eBay sample in a follow-up blog post after two months of use to see how its performance holds up over time.
Below are links to the tested PTM7950 sources. These are not affiliate links—this post is purely for informational purposes. We can’t guarantee the links will remain functional or that the price will be the same. We purchased the samples below to satisfy our own curiosity. 😊
We are not suggesting you purchase any of them; the links are provided only if you’re interested in seeing the listings for reference.
https://www.ebay.co.uk/itm/225680389112
https://www.aliexpress.com/item/1005005079093393.html
https://www.aliexpress.com/item/1005005960728133.html
https://www.aliexpress.com/item/1005004870180057.html
Our Priority is Helping You, Not Making Sales
Your technical issues matter to us. We’re here to listen and guide you through your options, not to sell you unnecessary solutions.
Confused by terms like VPN and SSD? We break down the tech talk into simple language—or dive deep if you prefer—ensuring you understand each step.
When tech troubles strike, every second counts. We promise swift and efficient replies, reducing your downtime and easing your stress—because we’re always ready to help.
Ready to Resolve Your Tech Troubles?
Don’t let technical issues slow you down. Our team is ready to provide you with fast, effective solutions tailored to your needs. Contact us to discuss your issues and discover how we can help you get back up and running smoothly.
