The Metabox P775TM-G is a pretty crazy laptop, featuring RTX 2080 graphics and a desktop Intel i9-9900K CPU, making a great desktop replacement or portable workstation, but just how well does it perform and what temperatures are we looking at? Let’s find out! Let’s start with the specs, for the CPU you’ve got the choice of a 9700K or 9900K, as mentioned I have the 9900K here.
I’ve got 32gb of DDR4-2666 memory running in dual channel, but it supports up to 64gb. For the graphics there’s Nvidia’s latest RTX 2080, but you can also pick the 2060 or 2070, and this powers the 17.
3” 1080p 144Hz screen, but you can also upgrade to 1440p at 120Hz, or 4K 60Hz. There’s also gigabit ethernet, 802.11ac WiFi and bluetooth, with a number of WiFi card upgrades to choose from. The lid and interior of the laptop are both a matte black and solid plastic, all of the edges were smooth and no sharp corners.
The dimensions of the laptop are 41.8cm in width, 29.5cm in depth, and about 4.1cm in height, so fairly large but expected considering the high end specs inside. The starting weight is listed as 3.9KG, but this will vary based on hardware selection and mine came in at 4.
1KG. As my scales only go up to 5KG I’ve had to weigh the 330 watt power brick and cables separately, which were about 1.4KG, so we’re looking at almost 5.6KG all up. As mentioned the screen is 17.
3 inches. You’ve got the option of 1080p 144Hz which I’ve got here, 1440p at 120Hz which might be a better choice for gamers considering the specs, and 4K at 60Hz which has better colour gamut, and all options have Nvidia’s G-Sync.
I’ve measured the current colour gamut using the Spyder 5 Pro, and my results returned 96% of sRGB, 67% of NTSC and 72% of AdobeRGB, so alright results, but this will vary depending on which panel you end up selecting.
At 100% brightness I measured the panel at 313 nits in the center, and with a 810:1 contrast ratio, so about average brightness for a laptop. There was a fair bit of backlight bleed in my unit, I’ve taken a long exposure photo in a dark room as a worst case backlight bleed test and it was occasionally noticeable during darker scenes, though this will vary between laptop and panel.
There was an average amount of screen flex while intentionally moving it, but as a fairly thick plastic it did feel sturdy, and having the hinges out towards the far corners helped with stability. Despite it weighing over 4KG it can only just be opened with one finger, as the hinge is fairly stiff and most of the weight is up towards the back, all that cooling I expect, but it was stable using it on my lap, though not really recommended given the weight and heat it can generate.
Above the display in the center is the camera. Like most other clevo units it’s got a decent 1080p camera and the microphone sounds pretty good too, although it does pick up a bit of its own fan noise.
The keyboard was pretty good to type with, no problems there. It’s got backlighting which can be customized in three separate zones and there are a few built in effects. Here’s how typing on the keyboard sounds to give you an idea of what to expect.
There was almost no keyboard flex at all even while pushing down hard, it was extremely solid. Above the keyboard is the power button in the center, as well as some LED indicators for caps lock and more.
Right up the back below the screen are the two 2 watt speakers. They sound really good for laptop speakers, fairly loud yet still clear with some bass from the subwoofer found underneath. The speakers also still work alright with the lid closed.
The touchpad was smooth and worked well. It uses synaptics drivers and has physical left and right click buttons as well as your usual gestures in Windows, and there’s also a fingerprint scanner in the top left corner.
Fingerprints show up on the hard matte plastic but they were easy to wipe off as it’s smooth. On the left there’s a gigabit ethernet port, two USB 3.1 Gen2 Type-C ports, and the first has Thunderbolt 3 support, two USB 3.
1 Gen 1 Type-A ports and SD card slot. On the right there are four 3.5mm audio jacks including headphone, microphone, line-in and S/PDIF, two more USB 3.1 Gen 1 Type-A ports and Kensington lock. The back has large air exhaust vents on the corners, so no hot air blowing on your hands, then in the middle there’s a HDMI 2.
0 output, two mini DisplayPort 1.4 outputs and the power input. The front is plastic with status LEDs towards the right. Underneath has air intake vents towards the back, subwoofer in the center towards the front, and rubber feet to rise it up a bit and help airflow, and they did a good job of preventing movement.
The battery can be removed easily without tools, otherwise the rest needs a phillips head screwdriver. The panel next to the battery has the two 2.5 inch drive bays and one of the two M.2 slots. Behind the final panel we get access to the second M.
2 slot, heatsinks and two memory slots. There are apparently 4 slots in total, so the other two are probably hiding on the back. Powering the laptop is an 8 cell 82 Watt hour battery, and with a full charge and just watching YouTube videos with the screen on half brightness, keyboard lighting off and background apps disabled, I was able to use it for 2 hours and 6 minutes.
Not great, but honestly not unexpected considering the powerful hardware, and it was using the RTX 2080 graphics the whole time, no Nvidia Optimus here. While playing the Witcher 3 with medium settings and Nvidia’s battery boost set to 30 FPS the battery lasted for 1 hour and 13 minutes, and the frame rate didn’t drop at any point.
Thermal testing was completed with an ambient room temperature of 25 degrees Celsius, and there is a single heatpipe between the CPU and GPU, so a change in temperature of one may affect the other. Starting at the bottom of the graph, at idle it was fairly warm even with the fans running.
The gaming tests were done by playing Watch Dogs 2, as I find it to use a good amount of combined CPU and graphics. Once the graphics are overclocked by 70MHz, CPU overclocked with power limit boosted and a -0.
1v CPU undervolt applied and fans maxed out, the GPU temperature stays the same but the CPU temperature rises and thermal throttles on 98 degrees Celsius. The stress tests were done with Aida64 and the heaven benchmark run at the same time in order to try and fully utilize both the processor and graphics, and this is resulting in thermal throttling on the CPU at 98 degrees Celsius in all instances except one, at stock with the fan maxed out.
At the top with a cooling pad in use the graphics temperature dropped the most, but the CPU was still thermal throttling, we’ll see how that improves performance in the next graph. These are the clock speeds for the same tests just shown.
Again starting down the bottom with the gaming results we’re getting a 500MHz improvement to the CPU with the power limit boosted and undervolt applied. The CPU overclock isn’t doing anything though as this is still below the 4.
7GHz all core turbo speed of the 9900K, as we hit thermal throttling before this is ever reached. Similar results in the stress tests, with the best results from the cooling pad, although not really that big of a difference in performance.
Here are the clock speeds from a CPU only stress test without any CPU load. At stock it was maxing out at around 3.8GHz on all 8 cores due to power limit throttling. With a -0.1v undervolt applied to the CPU and power limit boosted we’re almost able to reach the stock 4.
7GHz all core turbo speed of the i9-9900K. I couldn’t get any higher as this was now hitting the 98 degree Celsius thermal throttling, so again overclocking in a CPU only workload doesn’t seem too useful given we never pass the stock speeds here.
Despite the thermal throttling, these clock speeds on 8 CPU cores are still very nice in a laptop. To demonstrate how this translates into performance I’ve got some Cinebench CPU benchmarks here. I’ve included the i7-8750H down the bottom for comparison as it’s a pretty common good laptop CPU at the moment.
I’ve also included the results of a 9900K at stock in a desktop PC to get an idea of how the laptop stacks up, and even best case from the laptop it’s not quite able to achieve what the same CPU is capable of in a PC with proper cooling.
Here are the GPU only clock speeds while under a graphical only stress test, and I was only able to apply a 70MHz core overclock before experiencing crashes, and this didn’t really affect temperatures, just a 1 degree difference, we’ll see how this changes game performance a bit later.
I’ve used my thermal camera to check out the areas where you’ll actually be putting your hands, and at idle it was in the mid to high 30s. While gaming it was approaching 50 in the center and it felt warm to the touch, and then slightly cooler with the stress tests running.
As for the fan noise produced by the laptop I’ll let you have a listen to some of these tests. At idle it was fairly silent, only just audible. While under stress test it was a little louder than most other gaming laptops I’ve tested, and similar while gaming.
Then with the fan maxed out it was quite loud but only a little more than before. Finally let’s take a look at some gaming benchmarks, although I’ve got a 1080p screen here I have also attached an external monitor in order to test 1440p, as I think the specs we’ve got are too high to only focus on 1080p.
These games were tested at stock, we’ll look at overclocking later. Battlefield 5 was tested in campaign mode, and I’ve got RTX results shown in green. At 1080p it was still mostly playable with RTX on at ultra settings, not too far below a 60 FPS average, but we can get around double frame rate with RTX off, shown by the purple bars.
At 1440p RTX needed medium settings to play well, although in my personal opinion the game both looks and performs better at ultra settings with RTX off. Far Cry 5 was tested with the built in benchmark.
At 1080p the results were very nice, easily over 100 FPS possible even at ultra settings and 144 at low. Stepping up to 1440p the results are still quite nice, above 60 FPS 1% low at ultra settings with high settings averaging around the 100 FPS mark.
Assassin’s Creed Odyssey was also tested with the built in benchmark, and at 1080p we’re just able to average above 60 FPS with max settings, though 100 and above was possible with medium. At 1440p the results are still quite good for this game, with 60 FPS achievable with very high settings, not bad considering the game doesn’t really need a high frame rate to play.
Fortnite was tested with the replay feature, and at 1080p even at epic settings I was seeing very high frame rates, even the 1% lows are well above 100 FPS. Even at 1440p with the highest epic settings it’s still possible to average above 100 FPS, with the 1% low still above 60 FPS, and much higher frame rates are possible with lower settings.
CS:GO was tested with the Ulletical benchmark, and as expected at 1080p we’re getting very nice results in this test, with around 400 FPS possible at minimum settings, and still over 300 at max settings.
At 1440p the results don’t really change too much, as this game seems to rely more on the CPU than the graphics in my experience. Rainbow Six Siege was tested with the built in benchmark, and like CS:GO this test always scores very high frame rates, with well above 200 possible at 1080p even with ultra settings.
Even at 1440p with max settings the results are still very good, with the 1% low sitting at 120 FPS, so even the dips in performance aren’t going to be that bad. Shadow of the Tomb Raider was also tested with the built in benchmark, and 1080p high settings allowed us to reach 100 FPS in this test.
The frame rates drop back quite a bit at 1440p, by over half, and now only low settings is capable of getting above 60 FPS. Watch Dogs 2 was tested as a resource intensive game, and considering I can play the game fine with a solid 30 FPS it’s running very well even at ultra settings.
There’s not really any changes at the lower settings once we go to 1440p, but still at higher settings it was easily playable, with 60 FPS possible at ultra settings. PUBG was tested using the replay feature, and at 1080p the results were very good, though not too different between the different setting levels.
Again at 1440p with lower settings there isn’t too much difference, but at higher settings it’s still going quite well, even at ultra settings we’re seeing 90 FPS averages with 60 for the 1% low.
I’ve also gone back and retested the witcher 3 as a couple of people requested it in the comments of my 1080 vs 2080 video. At high settings even 4K was able to average above 60 FPS and was easily playable, with fairly high frame rates at lower resolutions.
As expected these crazy specs are giving us very nice results in games, really high frame rates at both 1080p and 1440p as well. Check the card in the top right if you want to see how the 2080 in this laptop compares with the GTX 1080.
Now for the benchmarking tools, I’ve tested Heaven, Valley, and Superposition from Unigine, as well as Firestrike, Timespy, and VRMark from 3DMark, just pause the video if you want a detailed look at these results.
As we saw earlier we’ve got the option of overclocking the graphics and CPU and boosting the power limit and undervolting the CPU to increase performance, so let’s see how this actually helps in gaming.
Far Cry 5 was retested using the built in benchmark at ultra settings. At 1080p there’s basically no difference to average frame rate, although the 1% low was lower with these changes. At 1440p there’s a 3% increase to average frame rate, and a larger 13% improvement to the 1% low.
At 4K there’s a slightly larger 4% improvement to average frame rate and a 9% boost to 1% low. Storage benchmarks will vary depending on the drive that gets selected when ordering, and my 1TB NVMe SSD was performing quite well.
The SD slot was performing very well too, basically getting full speed of my card. For updated pricing check the links in the description, as prices will change over time. At the time of recording, here in Australia the base model starts at $3600 AUD.
If you spec it out with the 9900K, RTX 2080, 144Hz screen and extra memory like I’ve got here that price rises to around $5400 AUD, not too surprising given the specs in this form factor though. For my international viewers that’s around $3500 USD taking away our taxes.
So what do you guys think about the P775TM-G laptop? Obviously this isn’t something for everyone, I’m not even sure it’s for people that want a serious gaming machine. To me it seems to be a desktop replacement, it’s more of a portable workstation I think, although of course with that 9900K and RTX 2080 you’ll be able to play any game no problem.
As expected it does get very hot when the CPU is under load, which should be expected in a laptop, the 9900K is a crazy hot chip even in a desktop PC with good cooling. That said though it still does perform very well, and if you need plenty of raw CPU power in a laptop I think these Clevo units are unmatched at the moment.
Despite the fairly big battery though it still doesn’t last too long given these high end power sucking specs, and the other issue I had was the bad backlight bleed, though that will vary. Basically if you’re after extreme power in a still fairly portable size this is the best in that regard I’ve tested so far.
Let me know what you guys thought about the Metabox P775TM-G laptop down in the comments, and leave a like to let me know if you found the review useful. Thanks for watching, and don’t forget to subscribe for future tech videos like this one.