Acer Helios 300 2018 Thermal Testing – Undervolting and Overclocking
Laptop thermal testing was completed with an ambient room temperature of 18 degrees Celsius, it's cold here at the moment, as it's winter in Australia, so expect warmer temperatures in a warmer environment at idle. Both the CPU and graphics were quite cool, as shown by the light blue bar at the bottom of the graph working our way up the graph. We start with the gaming results in the green bar, and this was tested playing watchdogs tune and the results aren't too bad. So far, if we manually max out the fans, the temperatures drop back a few degrees on both the CPU and graphics, as shown in yellow back with the fans on stock, but with the CPU under Bolton by minus 0.15 F volts and the GPU overclocked by 200. Megahertz, we see the temperatures increase shown by the orange bar and with the fans maxed out in red, the temperatures drop back a bit so we're getting better temperatures than playing at stock settings in the green bar, better getting better performance. Now, as we'll see later, distress tests were done by running a to 64 and the heav'n benchmark at the same time, in order to attempt to fully utilize split the processor and graphics moving up in the graph and starting with the dark red bar, I started to See both power limit throttling and thermal throttling we'll see how this affected clock speed in the next graph once the fans are maxed out shown by the pink bar just above the temperatures drop back a little, but we actually stopped them flooding.
Although power limit throttling is still, of course, present with the CPU under vault and GPU overclock applied the temperature of the CPU doesn't change, but we'll see in the next graph how this improved performance and the GPU temperature rises a bit from near the clock. Finally, with Berthe CPU on the vault GPU overclock and fans maxed out, as in the dot blue ball, the temperatures drop a bit, but the power limit throttling was still present. These are the average clock speeds for the same temperature tests just shown. You might need to pause and refer back to the previous graph to get the full picture first off starting down the bottom in the gaming results. We can see that just boosting the fan, speed in yellow slightly improved clock speed. This rose much more in the orange and red bars with the CPU wonderful plied, as it helps reduce the power limit throttling taking place, as this particular game uses a fair bit of CPU. The 8750 H has a 3.9 gigahertz or core turbo speed, and we can see in the red bar under stress test, with the under multiplied. We weren't too far from reaching this. In this same test, the graphics core clock speed was averaging just under 1900 megahertz, not bad. Moving up into the stress test results. The clock speeds in the dark red bar are the lowest due to the power limit and thermal throttling, with the fan maxed out in pink.
This doesn't really change anything as we're still hitting the power women once the CPU under volt is applied in the purple bar we're, getting much better performance on the CPU for with the fans maxed out. This appears to increase just a tiny bit, so the power limit is preventing us from reaching the 3.9 gigahertz or Core turbo speed of the 87 58 CPU in this particular stress test workload for games. This will, of course, be less of an issue unless your game is maxing out all cores consistently. So how does this performance boost actually translate into games in the games tested? The exact same windows updates, game updates and Nvidia drivers were installed, so there shouldn't be any changes other than the CPU under bolting and graphics overclocking. The same 0.15 EV volt under bolt to the CPU was applied as before, along with a 200 megahertz GPU core overclocked and 100 megahertz GPU memory, overclocked pub G was tested using the replay feature and at Ultra settings we're just seeing a little 5 percent boost to The average frame rates, but just a 1.5 percent improvement at very low settings, far cry 5 was tested with the built in benchmark and out ultra settings. There was also a 5 improvement to the average framerate and around the same boost at low settings Rainbow. Six siege was also tested with the bill and benchmark and at ultra settings there was just a 4 improvement to average frame rates.
Although the 1 lows didn't really change here. However, all the results were quite close at all levels, so we're, seeing a little improvement with the CP wonder, vaulting and GPU overclocking applied, although it depends on the particular game in question and setting levels in use. These are the clock speeds I got while just running CPU only stress tests without any GP load power limit throttling was always present in this test. Even without the GPU load, and even with the CPU Wonderbolt applied intel x2, you showed it sitting on a 45 watt. Tdp in a full multi core stress test, and I wasn't able to change this by modifying the values. Err next to you so I'm, guessing it's, defined at a low level and can't be changed. The laptop has a 180 watt power, brick similar to many other laptops. With similar specs that I've tested, it would be nice if there was a way to only limit the power, if required, say under full CPU and GPU load, rather than just arbitrarily cut it off of 45 watts. But this just seems to be how all laptops work. So I can only assume there's a good reason for it. I'Ve got some Cinebench CPU benchmarks here, and we can see that we get a nice boost in performance with the minus 0.15, though volt under volt applied at the top of the graph. As mentioned power limit. Throttling was present in all CPU. Only stress tests, so is still not getting full performance here.
This is about what I expected, based on other 8750 H laptops that I've tested, ideally with no throttling. The CPU should be able to pass 1200 points, so we're not too far behind the mark wants under vaulted single core workloads are the same regardless as no throttling takes place there. So less threaded workloads. Many games, for instance, will likely just be fine and get full performance, but if you need full performance in multi core workloads such as video exporting or games that actually support multi core well, for example, then you may want to look at under vaulting the CPU or Even if you just want to try and reduce temperatures, it can help. Here are the GPU in the clock speeds, while under a graphical, only stress test, aces, predator sense software lets you apply GPU over clocks easily in two different levels, known as pasta and Tober the faster profile over clocks, the Phoebe you call by 45 megahertz and the Memory by 50 megahertz, while the turbo profile doubles this to 90 megahertz on the core and 100 megahertz on the memory, I was able to get a little further improvement by manually overclocking it with MSI Afterburner, as shown in red, Fitness will probably vary between laptops, as It depends on the particular chip, as for the external temperatures will actually be put in your hands at idle. The body of the laptop was sitting in the low 30s fairly cool while gaming.
This increases to the mid 40s toward the center of the keyboard and high 40s towards the back. This was just a little warmer than running my stress test and with the same test running, but with the CPU under volted and fins maxed out, we can see an improvement of a few degrees. As for the fan, noise produced by the laptop I'll lay you ever. Listen to some of these tests at idle. It was almost silent early just inaudible, while gaming with the water fan profile. It was about as loud as most gaming laptops, I've tested under the combined CPU and GPU stress test. This row is just a little and then finally, with the fans maxed out, it was fairly loud. You'Ve got the option of controlling the fan, speeds on the CPU or graphics independently, through Asus prejudice and software, so that should help in finding a good balance between temperatures and fan noise. Overall, the performance was about what I expected for a laptop like this. With the 87 58 CPU just to be clear, the power limit throttling isn't an issue unique to the Helios 300 I've, seen this in pretty much every eye: 78758 laptop I've tested, but as covered here there are steps we can take to mitigate this and improve performance. These differences in performance shown aren't hard and fast rules. There are different factors which will vary results, primarily the temperature you're running in application of thermal paste, and given the specific hardware which comes down to the silicon lottery, you may not be able to undergo all overclock.
Your Hardware, the same as me, it depends on the chip and it's specific power requirements, so don't just blindly copy my settings and do some testing to find out what your stable point is for the best results with under faulting alone. That was still minor thermal throttling under stress test with the fans in stock, but this was removed by maxing out the fans and remember I'm testing in a fairly cool room at 18 degrees Celsius. So in a warmer environment, thermal throttling may be more of a challenge. While you could probably improve the temperatures by swapping out the thermal paste that's, not something I can test in a review unit. If I go ahead and remove the stock thermal paste and replace my own, I can't put the old paste back, so the next reviewer would experience something different from what you'd actually see with the product and unknowingly report incorrect information. Due to what I've done. That said, pal limit throttling was more of an issue from my testing anyway, on defaulting, on the other hand, isn't physically intrusive and as we've seen, it did improve the performance and temperatures in this particular unit with no downsides once you've got a stable under vol it's. A great way to get back some performance if you're running too hot or facing car limit throttling like we were here. Let me know how much of a performance boost you've found by under bolting hardware and what you thought of the improvements here and don't forget to subscribe for the full review of the Acer predator Helios 300 gaming laptop as well as future tech.