I recently acquired an Eluktronics MB-15, specifically the i7-8750h model with a GTX 1050ti and an 144hz panel, with the $35 cooling upgrade to Liquid Metal on the CPU and a more traditional compound on the GPU. So far, I’ve been loving it, but I’ve only really been using it for a few days. I didn’t feel comfortable typing up a review for a product I literally just got, so I decided “hey, why don’t I just catalog some of my findings on the thermals?” Well, here they are.
The information in this article was last updated on 5/31/2019. This page is subject to change.
I ran my stress tests per D2Ultima’s recommendations (according to this video. He’s got a pretty awesome, easy to follow video on how to get a testing setup all situated for laptop stressing.) Stressing your laptop’s hardware is a great way for you, as a consumer, to be informed about what you’re actually getting for your money, as well as potentially helping others out in the process! The main reason why stress test results are important is because, for example, some laptops that are marketed as “gaming laptops” actually choke their GPUs and CPUs substantially, sometimes even to the point of running no faster than their lower spec, but same model counterparts. That’s a quick way to send an extra $300 down the drain. Luckily for you, I like doing this kind of thing, and I’d love to do it more in the future with more laptops.
I ran the recommended stress test for 20 minutes with fans pinned to 100% “Turbo mode” with Rev.Center 2.0. My MB-15 out-of-the-box was limiting itself to 35w under combined CPU and GPU load. The reason for that still has me stumped, since it could easily handle 45W and boost up to 60W-ish on occasion for short burst loads. In this configuration, I had it running as fast as it could, and with as much power as it could. As you can see, this laptop can sustain 60W for short yet intensive burst workloads, clocking as high as 3.9 GHz and holding for the first few minutes. However, after the heatsink became “saturated” with heat, and after most benchmarks would usually be over, temperatures climbed and climbed.
By the end of the stress test of 20 minutes (simulating something like an intensive video render, or even Battlefield V in some situations), wattages and frequencies had dropped. The CPU was left bouncing between around 45W to 55W under full load, hanging around 3.5 GHz due to thermal throttling (since this laptop “hardware throttles” or triggers PROCHOT at 90C). This is still pretty good performance, especially given the size of the chassis and the mostly plastic construction. I’m still a tad bit disappointed that it couldn’t sustain higher wattages without overheating. The plastic palmrest might not work as well as a heatsink when compared to the Mech-15 G2’s metallic palmrest. That might be choking things a bit, but don’t quote me on that one; I’m not a laptop engineer by any means.
Besides the CPU, however… Look at this graph!
Outdated Nickelback memes aside, those dips you see are Heaven loading in new assets, and as a result putting the GPU to sleep momentarily. The frequency stays quite high the entire time, especially with how toasty everything was getting on the machine.
As reported by HWInfo64, the GPU clock hit a max of 1784.5 MHz. (The average number is a bit flawed, given the pauses the GPU made while loading in assets, so ignore those.) It dropped only about 30 MHz under sustained load as far as I was able to tell, which isn’t a whole lot in the grand scheme of things. The more careful-eyed of you might have noticed this…
… and gotten a bit suspicious since I haven’t mentioned it yet, but fear not. Here’s what those labels should say.
So, during stress tests, the GPU was only ever limited by “Reliability Voltage.” I’m not exactly sure what that means, but it definitely wasn’t thermal limiting.
I’ll be noting any further things I experience with this laptop performance wise here, as well as going over my findings again in my future review.