Alienware Area 51m Thermal Testing, Overclocking and Undervolting
So what sort of temperatures we are looking at with this desktop replacement In this testing Ill be taking a detailed look at thermals and seeing how much we can improve performance and temps with some simple changes.. The Alienware Area 51m is available in different configurations. Ive got the top end model with Intels 8 core overclockable desktop CPU, the i9 9900K Nvidia RTX, 2080 graphics and 64GB of memory running in dual channel.. You can find examples of different spec options, as well as prices linked in the description. On the bottom of the laptop. There are air intake vents towards the back, and most of the air is exhausted from the two corners on the back, though, there are also smaller vents on both left and right sides towards the back.. As for the heatpipes, well, as I saw at CES earlier this year, theyre pretty massive as youd expect for hardware like this.. The Alienware control center allows you to run the laptop at stock settings or also pick from two built in overclocking profiles, 1 and 2, and these are the settings that each profile gives us.. None of the profiles perform undervolting. They basically just boost clock. Speeds. Thermal testing was completed in an ambient room. Temperature of 21 degrees Celsius so expect different results in different environments.. All of these tests 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.. All tests were also run with the fans at maximum speed. Lets start with the stress test results at stock down the bottom. The CPU was thermal throttling at 99 degrees Celsius, as shown by the blue bar.. Simply by applying a 0.15v undervolt to the CPU, we were able to drop the temperatures back by 12 degrees. When we enable the overclock 1 profile in combination with the undervolt. The CPU temperatures rise, a couple of degrees and the GPU also goes up by 4. If you recall earlier this profile, also overclocks the GPU, which equals more heat. Stepping up to the overclock 2 profile, which boosts the 8 core boost, speed by 100MHz, were back to thermal throttling at 99 degrees., As the higher tier overclock requires more power. I had to drop the undervolt to 0.075v to keep it stable in this workload and then, when we add in the cooling pad, were only seeing a small improvement to the GPU temperature.. The gaming results are a bit better comparatively in terms of temperatures, not quite as intense as a synthetic stress.
Test. Honestly, once we add the undervolt, even the overclock 1 profile, isnt actually running too hot at all, so personally, Id probably use it somewhere around this level. To keep it from getting too hot, as youll soon see, were not really getting much extra performance from the extra heat.. In any case, the cooling pad did at least make a bigger difference here compared to the stress tests.. These are the average clock speeds for the same tests, just shown. Youll notice that theres not actually all that much difference in terms of the CPU average over all 8 cores, maybe a 350MHz difference from the lowest result to the best result. Rather than tweaking for maximum Performance with these specs, I think it will be more important to make changes to keep it running. Cool instead. In the stress test at stock were hitting thermal throttling, so just a lowly 4.5GHz all core speed on the CPU. Undervolting allows the CPU to hit the stock 4.7GHz all core turbo boost, speed of the 9900K and then with the overclock 1 profile with the Undervolt were able to hit the newly defined 4.8GHz, all core limit., Its worth, remembering OC1 doesnt undervolt by default. But basically, if I didnt undervolt, the overclock profiles were useless in this workload, as we were already thermal throttling at stock with the fans maxed out., Applying overclocks isnt, going to help us out without first dropping those temperatures, hence the undervolting being essential to properly utilize.
The overclock 1 and 2 profiles in this stress test, though this would not be the case in instances like gaming, where thermal throttling is not hit, though, of course depends on the game.. The GPU also sees a nice clock, speed bump here due to the overclock. This profile applies and then a little extra boost with overclock 2, which further raises the clock speed. With the second overclock profile. The all core turbo boost speed is raised to 4.9GHz, however, were only getting a little extra speed here, as I had to lower the undervolt to keep it. Stable. Were hitting thermal throttling here, so this isnt, all that useful and nor would be adding more voltage and the cooling pad didnt really help this out.. The game tests saw similar clock speed changes, though we were able to reach the full 4.9GHz. All core turbo boost speed with the overclock 2 profile here with 2000MHz on the 2080.. These are the average TDP values during these same tests.. The take aways are that were using a fair bit less power when undervolting is applied, which is why we see lower temperatures. Less power equals less heat, which equals more performance when thermal throttling is hit.. Otherwise, the RTX 2080 was pretty crazy, spiking up to 190 watts at times.. These are the average CPU clock speeds, while under a CPU only workload. At stock, we are able to hit the full 4.7GHz all core turbo boost speed of the 9900K under this workload.
So no change once undervolted.. If anything, this test is too weak for this tier of hardware, and maybe I should have used stress FPU given were hitting 5GHz with my manual overclock at the bottom.. We can also see how much of a difference to thermals undervolting makes all theses. Tests with higher speeds are running cooler than the stock out of the box results, which makes me wonder why Alienware didnt incorporate stable undervolt settings with the overclock 1 and 2 profiles.. My guess is generally thats, not something you want to do with such a high powered CPU, although in my testing so far its been good., The temps basically match the CPU TDP as more power equals more heat. To demonstrate how this translates into performance. Ive got some Cinebench CPU benchmarks. Here. In this test were not really seeing that much of a performance boost a couple hundred points depending on the customizations being made.. In this workload I couldnt get overclock 2 working even with a lower undervolt, so I ended up giving up. So this profile really doesnt seem useful for multicore load, as weve seen earlier, but may be more useful in games that dont thrash the CPU.. So how do these performance boosts actually translate into games? Ive tested with the exact same Windows, Nvidia and game updates. Installed. Far Cry New Dawn was tested using the built in benchmark at ultra settings. Theres a bit going on here. So let me explain.
Ive got the laptop tested at stock in the purple bars with the overclock 1 profile in the green bars. Then, with the overclock 2 profile in the red bars.. Basically, we can see the overclock 1 profile is giving us the best results, as by default. The overclock 2 profile seems to push it too far and cause higher levels of throttling still better than stock results. But again, I think the overclock 1 profile is a good sweet spot, especially if you undervolt it., If youre after more gaming benchmarks from the Alienware 51m check the card in the top right, where Ive tested 15 different games at 4K, 1440p and 1080p resolutions.. As for the external temperatures, where youll actually be putting your hands at idle, it was sitting at about the typical 30 degree celsius average., With the stress tests going. The keyboard area is actually still quite cool and only getting to 40 up the back.. In the same test with the fan maxed out, it was a little cooler, only a little warm to the touch and quite impressive when you consider that in this test, the CPU is thermal throttling at 99 degrees.. As for the fan noise produced by the laptop Ill, let you have a listen to some of these tests. At idle. The fans were only just audible. With balanced mode set. It was definitely quieter compared to most other laptops tested. So you do have the option of sacrificing performance and thermals for quieter.
Noise levels., Even with the fans at maximum while it is loud, Ive, certainly had louder machines with much weaker specs that still ran just as hot take the Aorus 15 for example.. You might be asking why Ive tested different volume levels of fans without also testing the thermals well buckle in because have I got a story for you.? Essentially the Alienware control center software is trash, or at least that was my experience while using the latest version. During my testing., You need it installed to use the overclock profiles and control fan speed.. Basically, with the software installed during both stress test and while gaming, the GPU power limit, would drop every few minutes. You can see in the screenshot. The maximum power was 195 watts, but the minimum was about 40, while its currently sitting at 50, with a '0MHz clock speed.. This means that it thought it was operating at 100, though the performance would suffer during these periods.. You can see the historical GPU load bar is essentially the same, but the clock speed cycles up and down during these periods.. Yes, both power bricks were plugged in. I even tried a clean install of Windows, but the problem persisted as soon as I reinstalled the Alienware software again.. The only way to prevent this from happening was to remove the alienware oc controls and control center software.. By doing this, the fans were stuck at what ever I last configured them to which in my case was maximum.
. I also was never able to reinstall the oc controls, no matter what I did, all sorts of errors kept coming up, no matter what order. I did things in and I dont have time to keep doing fresh installs of Windows., Basically after wasting over a day troubleshooting. In the end, I decided it was easier for me to manually, apply the settings of the overclocking profiles with MSI Afterburner and Intel XTU and just leave the fans at maximum so that I wouldnt have these dips in performance.. I couldnt find very many instances of this issue happening to others, so it could just be my unit.. There was one guy who posted online, who seemed to think it was an incompatibility with having MSI Afterburner installed. However, I confirmed this was not the case after my clean Windows, install. Just to be clear, it wasnt a reset. I tried that too then after I downloaded a fresh 1903 ISO and installed it from scratch.. This was one hundred percent, some sort of software bug in the control center. If I swapped between the different profiles, while in a low performance period, it would reset the state and the power limits would boost back up for a while straight away.. So, while it should be possible for them to fix this, this issue severely impacted performance in my testing to the point where I had to rage quit trying to use it. However, I believe the data represented here accurately shows what youd actually get from using the software overclock presets normally, as I manually set the same settings that those use.
Overall Im pretty impressed with the performance of this machine software issues. Aside.. Yes, out of the box with stock settings, it does run hot and thermal throttles under a combined, CPU and GPU stress test. However, with some simple undervolting, it was possible to improve this quite a bit.. You really do need to undervolt it, though, to get the most out of it, as we saw due to the thermal throttling the stock, all core turbo boost, speed, wasnt being hit under stress test.. This means enabling the overclock profiles is essentially pointless on the CPU side. Until we can bring the temperatures under control, though some extra GPU performance was still possible from these.. Combining the settings with undervolting, though, gave excellent performance without the temperatures being too crazy, so it would have been great if Alienware could have incorporated this into their default profiles. Maybe next time. A cooling pad didnt, help out too much here. Even while thermal throttling clock speed boosts were not really seen, though, interestingly, the fans werent crazy, loud either. Dont. Get me wrong. 55 decibels. Maxed out is on the louder side, but it was possible to lower the fan, speed at the expense of thermals and when you compare it to other machines, Ive tested that reach the same noise levels with far weaker specs. I think its fair.. This crazy machine is powered by two power bricks, which are 330 watts and 180 watts, and I didnt find the battery to drain during any of my testing.
Even with the overclocks.. These differences in performance shown arent hard and fast rules. There are different factors which will vary results, primarily the temperature of the room, youre 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 dont 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 Im not able to change the paste, otherwise, the next reviewer will unknowingly report different results. Due to what Ive, done. Additional undervolting, boosting fan, speed or using a cooling pad are much easier for most people to do anyway and as weve seen, these tweaks did help improve performance of the Alienware 51m.. Let me know what you thought of the 51ms thermals down in the comments and Im, sorry that it didnt catch on fire like so many of you were hoping..