ASUS Zephyrus S (GX502GW) Review
The ASUS Zephyrus S GX502 gaming laptop has some powerful specs in a thin form factor, but this combination usually equals higher temperatures and lower performance, so let’s test it out and find out just how hot it gets and see what can be done to improve it.
I’ve got the highest specced version of the GX502, there’s an Intel i7-9750H CPU, Nvidia RTX 2070 graphics, and that’s the full one, no Max-Q here, along with 32gb of memory running in dual channel. The GX502 is also available with different specs, so expect different results with different configurations. You can find examples and updated prices linked in the description. The Zephyrus design is known for raising the back of the machine up when you open the lid to assist with air flow. As a result, the bottom of the machine is completely flat with no air intakes, so I haven’t bothered testing with my usual cooling pad. Underneath the metal flap that gets pushed down are the two fans, and air seems to be brought in from this gap, and also in through the keyboard. Air is then exhausted from the vents on the back left and right corners, as well as on the back of the machine. Inside we’ve got a few heatpipes, with a couple shared between the processor and graphics.
The ASUS Armory Crate software allows you to change between three different modes, silent, performance and turbo, and I’ve tested all three out here. Basically these modes adjust maximum fan speed, CPU power limits, and control GPU overclocking, as defined here. All testing was also performed with Optimus disabled, so we’re only using the Nvidia graphics. You can also easily swap between these modes by holding the function key and pressing the pressing F5, the key with the fan icon. Thermal testing was completed in an ambient room temperature of 21 degrees Celsius, so expect different results in different environments. At idle both the CPU and GPU were on the cooler side, no issues there. The rest of the results are from combined CPU and GPU workloads, and are meant to represent worst case scenarios as I ran them for extended periods of time. The gaming results towards the upper half of the graph were tested by playing Watch Dogs 2, as I find it to use a good combination of processor and graphics. The stress test results shown on the lower half of the graph are from running the Aida64 CPU stress test with only the stress CPU option checked, and the Heaven GPU benchmark at max settings at the same time to fully load the system.
I’ve only tested silent mode with the stress tests, and the temperatures are the lowest out of all results as this is restricting the power limits harshly, which we’ll see soon. With performance mode enabled things start to heat up, and then with turbo mode we see the increased fan speed reduce the GPU temperatures, though there was never any GPU thermal throttling taking place.
There was, however, intermittent CPU thermal throttling happening. The CPU temperature actually rises slightly in turbo mode and was now constantly thermal throttling at 95 degrees, and this was despite the increase to fan speed as power limits are increased.
Undervolting the CPU made no change in terms of thermals, however I was only able to apply a small undervolt. This is something I’ve noticed with other Zephyrus machines that I’ve tested, they all can’t be undervolted much and I’m not sure why, maybe ASUS are able to buy binned Intel chips that already use less power for their thinner machines, I don’t know.
These are the average clock speeds for the same tests just shown. Check out that super low GPU clock speed in silent mode while under stress test. This is why I didn’t bother testing silent mode in gaming, the game just wasn’t playable in this mode for me, but we’ll see some FPS benchmarks later. When we step up from performance mode to turbo mode we’re seeing a boost to GPU clock speeds, as turbo mode overclocks the graphics. The CPU clock speed also rises as this raises the CPU power limit too, although as we just saw the next limit was thermal throttling. The small CPU undervolt did help reduce the throttling and improved the clock speed, however this is still below the full 4.0GHz all core turbo boost speed of the i7-9750H by around 500MHz.
These are the average TDP values reported by hardware info during these same tests. This explains why the silent profile was useless for gaming, look how hard it’s capping the GPU power. In all other modes it was able to hit its 115 watt limit without issue, as shown by the green bars. In performance mode the CPU power limit raises from a 28 watt PL1 in silent mode to 30 watts, which was the limitation in both stress test and while gaming, as shown by our average of 30 watts during the testing. Although turbo mode boosts PL1 to 45 watts we’re not actually able to hit above 35 watts due to thermal limits being reached.
These are the average CPU clock speeds while under a CPU only workload, so with no GPU load contributing to the thermals we’re seeing better results, with the full 4GHz turbo boost speed being hit at stock simply with turbo mode enabled. It might seem counterintuitive, but as we get higher performance the temperatures actually go down, and this is because the fan speed increases as we step up from silent through to turbo mode. Looking at the power limits, I noticed that while under a CPU only workload PL1 was being set to 90 watts in turbo mode, which explains why prior to undervolting it’s higher than the 45 watt limit I mentioned previously while under combined CPU and GPU stress test.
To try and avoid confusion, here are the power limits that were observed throughout all of these tests, so we so get higher limits being set for us while the GPU is inactive. The 45 watt limit for combined load seems fine anyway, given we don’t even approach it due to thermal limitations, at least in the workloads I’ve tested here.
To demonstrate how this translates into performance I’ve got some Cinebench CPU benchmarks. While under turbo mode the results were actually pretty decent for a 9750H, above average in any case thanks to those higher CPU only power limits on offer.
So how do these different performance modes actually affect games? I’ve tested a few games with these built in options, as well as with my small undervolt to find out. Battlefield 5 was tested in Campaign mode at ultra settings. With silent mode enabled down the bottom we can see that poor performance I mentioned earlier, while the average FPS was still above 60, the dips were terrible making it barely playable. We then see slightly improved performance as we step up through the other options. Far Cry 5 was tested with the built in benchmark at ultra settings. In this test the silent profile didn’t seem to make as big of a difference, otherwise again the frame rates improve as we raise up the settings.
Shadow of the Tomb Raider was tested with the built in benchmark at highest settings. The additional CPU undervolt made no difference to average FPS in this test, and like the other games performance mode wasn’t too far behind, while silent mode was further back - still above 60 FPS for the averages, but as we’ve seen averages aren’t the full story.
If you want to see more gaming benchmarks on this machine, check the card in the top right where I’ve tested 20 games at all setting levels.
As for the external temperatures where you’ll actually be putting your hands, at idle it was around the normal 30 degrees Celsius. While gaming or under stress test with performance mode it was getting quite warm, mid 50s in the center and low 60s right up the back, some of the keys right up the back were quite hot to the touch. With turbo mode enabled and the fans now faster it was a bit cooler now, still quite hot right up the back, though you probably won’t be touching there that often, while the left and right sides were noticeably cooler as this is where air is pulled in.
As for the fan noise produced by the laptop, I’ll let you have a listen to some of these tests. At idle in silent mode the fan was still audible. With the stress tests running with silent mode we can see the machine is still fairly quiet, however as we saw earlier this mode wasn’t too useful for gaming. Performance mode was significantly louder comparatively and a little above the average, while gaming takes it louder still. Overall the ASUS GX502 gaming laptop runs hot. This honestly wasn’t that unexpected, it’s on the thinner side and they’ve crammed a max-p RTX 2070 GPU inside rather than the usual max-q we usually see in this sort of form factor. As a result, we did see extremely impressive gaming performance from this machine, despite the thermal throttling on the CPU identified throughout this testing, so it seems like in most games that GPU power is making up for it.
There were minimal improvements we could make to improve performance, as there are no air intake vents underneath the machine it wasn’t possible to use a cooling pad, and I was only able to add a fairly small undervolt compared to what I’m usually able to do with most other machines, and this is something that I’ve experienced with all Zephyrus laptops that I’ve tested.
Outside of combined CPU and GPU loads such as gaming, we’re seeing nice power limits for the CPU. This results in CPU only workloads scoring quite nicely compared to other machines with the same 9750H CPU.
These differences in performance shown aren’t hard and fast rules, there are different factors which will vary results, primarily the temperature of the room you’re running in, application of thermal paste, and even the specific hardware which comes down to the silicon lottery. You may not be able to undervolt or overclock your hardware the same as me, it depends on the chip and its specific power requirements, so don’t just blindly copy my settings and do some testing to find out where your stable point is for best results. It may be possible to further improve temperatures by swapping the thermal paste, however as this is a review unit that I have to send back I’m not able to change the paste, otherwise the next reviewer will unknowingly report different results due to what I’ve done.
Raising fan speed and undervolting are much easier for most people to do than changing paste anyway, and as we’ve seen these tweaks did help improve performance and temperatures with the GX502. Let me know what you thought about the thermals from the ASUS Zephyrus S GX502 gaming laptop down in the comments.