Pascal Goes Mobile; GTX 1070 Notebook Review

Editor-in-Chief

Author: SKYMTL
Date: August 15, 2016
Product Name: ASUS ROG G752VS OC Edition
Part Number: G752VS-XB78K
Warranty: 1 Year

Another week and there’s another launch from NVIDIA. While the TITAN X, GTX 1080, GTX 1070 and GTX 1060 have made headlines for their ability to move the gaming goalposts forward in every segment. However, despite all of those achievements today’s announcement is arguably the one I’m most excited about since it represents the culmination of everything NVIDIA’s Pascal architecture has set out to achieve. Today, not one but three distinct graphics cards are being introduced into the notebook market, a true triple salvo into a space where there’s an already dominating GeForce GPU presence.

While it may have taken just over two months for NVIDIA to roll out their first four desktop GPUs, on the notebook side they’re ripping off the band-aid in one shot. In an effort to hit as many price points as possible right before the key back to school shopping season, mobile versions of the GTX 1080, GTX 1070 and GTX 1060 will all be available in notebooks at various price points in the coming weeks.

The quick launch of these notebook parts so soon after their desktop counterparts is interesting but not surprising in any way; the Pascal architecture is literally a perfect fit for both markets. From its low heat output to high overall efficiency, the duality of this core design is incredible.

One thing you will notice right off the bat is a lack of the distinctive –M moniker at the end of these cards’ names and there’s good reason for that. Unlike previous generations where bits and pieces had to be locked away to meet stringent efficiency guidelines absolutely none of these cards is being cut down from their desktop counterparts. For example, the GTX 980M may have been near the pinnacle of last year’s lineup but it actually used the GM206 core from the mainstream GTX 960 graphics card. While there was a late-bloomer in the form of a mobile version of NVIDIA’s full GTX 980, it proved to be quite late, supremely expensive and quite power hungry even for high end gaming laptops.

As you can see in the chart above, while every one of these notebook-bound cards’ baseline specs align with the non-mobile versions, there are some very minor sacrifices in clock speeds. This was likely done to reduce the amount of core voltage which also minimizes heat output and power consumption, two key factors when determining notebook compatibility. The sole stand out however is the GTX 1070; in its notebook form it has an additional SM unlocked, pushing it to 2048 cores and and 128 texture units. This is an interesting development considering the desktop GTX 1070 could now fall behind its mobile sibling.

Another interesting addition to the lineup is a 3GB version of the GTX 1060. While this was initially rumored to be a desktop SKU, it looks like NVIDIA is first announcing it here. Don’t take it to be a notebook-exclusive product though since we may just see an iteration of it as an add-in-card sometime down the line.

Throughout every single review of NVIDIA’s Pascal architecture, I’ve always talked about the relative performance of these new cards in comparison to their predecessors. The GTX 1080 was between 60% and 90% faster than the GTX 980, the gap between the GTX 1070 and GTX 970 was about the same while the GTX 1060 also featured massive intergenerational framerate uplifts. The notebook GPUs are looking to continue that trend and also extend their performance envelope into 4K territory, which used to be a domain solely reserved for extremely high end desktop cards.

It also goes without saying that performance like this coupled with Pascal’s inherent efficiency will open up a whole new range of options for gaming laptops. NVIDIA’s marketing slides call it a “quantum leap” and I tend to agree. 120Hz GSYNC displays, significantly longer battery life, ultrathin options that can offer awesome in-game framerates and more are all now possible and will likely see the light of day in very short order. Indeed, some companies like Gigabyte and Razer are announcing today that their sub-20mm notebooks weighing as little as 4lbs will house some of NVIDIA’s newest mobile GPUs like the GTX 1060.

Overclocking is also playing a surprisingly large role in this launch for both system integrators and end-users and that’s something which runs contrary to history in notebook segments. While mobile Maxwell GPUs did offer a range of limited options for boosting their clock speeds (typically a few presets within a manufacturer’s control panel), this generation will feature direct compatibility with some of today’s most popular overclocking tools like EVGA’s Precision and MSI’s AfterBurner. Provided the included thermal solution has been deemed sufficient, it’s very possible we could see clock speed boosts of 300MHz or more on some products.

From the manufacturer’s side of this coin, NVIDIA has supposedly been quite conservative with their reference clocks on the GTX 1080, GTX 1070 and GTX 1060 for notebooks. As a result and like we’ve continually seen from add-in board partners, some gaming laptops could come pre-overclocked out of the box. The ASUS G752 OC Edition I’m reviewing in this article overclocks the GPU’s core and memory by 50MHz and 100MHz respectively when in its present Extreme mode. Now this may not offer much in the way of real-world performance improvements but as they say; something is better than nothing.

Despite the inherent efficiency of these new Pascal-based graphics cores, NVIDIA has still looked towards improving some of their unique software-based technologies in an effort to increase battery longevity without killing performance. One key weapon in the battery life versus performance war is BatteryBoost which essentially allows a user to set in-game framerate targets and then modules the GPU’s output to hit those targets. As a result efficiency is improved since, in theory, the GPU shouldn’t have to utilize its full amount of horsepower to hit the highest framerates possible.

Now BatteryBoost is receiving a bit of a makeover. Instead of fluctuating performance it should achieve much smoother framerate output while also attaining better battery life. How much better? Well, when combined with the advancements made within the Pascal architecture, according to NVIDIA we can expect 30% longer runtimes with these new GPUs.

With the GTX 1080, GTX 1070 and GTX 1060 for notebooks, it looks like NVIDIA is simply attempting to bury their competition under an avalanche of high level launches. Truth be told, it is becoming increasingly obvious that NVIDIA doesn’t even consider AMD much of a competitor anymore and as a result they’re marching to the beat of their own drummer. This launch isn’t a reaction to anything from the Radeon lineup since notebook-destined AMD cards at the GTX 1070 and GTX 1080’s performance level could be more than a year away. Unfortunately this leaves us with something of a monopoly and I don’t doubt for a second pricing of these new notebooks will very much reflect that.

The only potential competition comes in the form of AMD’s RX 480 against the notebook GTX 1060 but judging from Polaris 10’s high TDP values, actually integrating it into some lower-cost, slim and light chassis may prove to be challenging.

Where this leaves us is in an unfortunate situation where NVIDIA has a long running stranglehold on a segment that’s virtually begging for some competition. I also have to wonder how, given the very real challenges of maintaining stock of these cards’ desktop alternatives on retailers’ shelves, there’s any hope of seeing more than a smattering of notebook GTX 1080’s, GTX 1070’s and GTX 1060’s. If suddenly we start seeing widespread availability of Pascal-based notebooks before the inventory situation on the desktop side resolves itself, NVIDIA will need to answer some very challenging questions.

There is a ton to cover in this article but even though every one of these new notebooks GPUs looks enticing, only so much could be done before the veil lifted on this particular launch. As such, the first review unit I got my hands on is featured here: ASUS’ Republic of Gamers G752VS OC Edition. I’m not complaining though since this thing is absolutely awesome and if its performance is any indication, we have a whole lot to look forward to from NVIDIA’s notebook Pascal lineup.

A Closer Look at the ASUS’ RoG G752VS OC Edition

As with every single notebook range on the market, the products which house NVIDIA’s newest Pascal GPUs will come in every shape and size and should fit into a vast number of price ranges. For this review, ASUS sent us one of their new G572VS OC Edition which includes a GTX 1070 and retails for an astounding $2999USD or $3999CAD. Hey, no one said these new graphics cards would come cheap, did they?

For this epic outlay of money you get an unlocked Intel i7-6850HK processor, 64GB of memory, a 512GB Toshiba NVMe SSD, 1TB hard drive a G-SYNC compatible 17.3” IPS screen and countless other features. The G572VS OC Edition is obviously a high-end notebook that includes everything but the kitchen sink but three grand will be hard to swallow for most PC gamers. Luckily there are some other options detailed above that will still get you into the portable GTX 1070 game but won’t require the sale of your right testicle.

The OC Edition’s main claim to fame is its out-of-box overclock. With it, ASUS has automatically boosted the CPU speeds to a maximum of 3.8GHz, memory rings in at 2400MHz and the GTX 1070’s core / GDDR5 work 50MHz and 150MHz faster respectively. This has been achieved through two distinct modes within ASUS’ ROG Gaming Center: Standard and Extreme, the latter of which is the default mode for this notebook. The Standard setting is simply there for folks who want slightly lower acoustics and longer battery life. Something else to note is that regardless of the Gaming Center setting, the GPU maintains its higher-than-reference frequencies.

The Gaming Center can also be considered as the jump-off point for a large number of other features. For examples you can overclock the system even further (both CPU and GPU), monitor system readounts, set up macros and even enable smartphone control over the notebook. It’s actually quite brilliant and intuitive at the same time.

NOTE: For the purposes of this review, there will be two ASUS G752’s in the charts. One is labeled as “OC”, denoting the notebook running in its Extreme mode while the other just indicates Standard mode.

The G752VS may be one of the best looking, most ruggedly built dedicated gaming notebooks I’ve seen in a long time. Its faux aluminum housing includes some tasteful red accents and a glowing ROG logo that’s flanked by a pair of slim illuminated strips. Meanwhile, the hinge mechanism is perfectly integrated into the back-facing chassis.

Around back there’s a bright red grille that hides ASUS’ new cooling assembly which consists of an advanced 3D vapor chamber and a special funnel design that nearly eliminates dust buildup in key areas. That vapor chamber may not be seen but it plays a massive role in properly regulating the G752VS’s component temperatures so neither the GPU nor CPU throttles under load.

Remember, throttling is typically a huge problem with gaming notebooks in particular since their internal heat dissipation systems aren’t large enough to properly cool off some higher end components. If ASUS has figured out a way around those challenges with clever engineering, the GTX 1070 could extend its lead even more over past notebook GPUs.

Under the lid lies that spectacular looking 17.3” IPS display with a matte coating and ASUS has given particular attention to its longevity. Everything you see here may be fabricated out of plastic but it has been thoroughly reinforced, giving the G752 an extremely solid feel.

The full size backlit keyboard ASUS has designed for this notebook may not have a brand name on it like the SteelSeries units MSI’s alternatives comes with but it is nonetheless extremely well laid out, exhibits good key travel and provides just the right amount of feedback. Its WASD keys have a subtle red highlight that helps with locating them in the daytime and an additional halo for nighttime viewing. ASUS has also included 30-key rollover and anti-ghosting technologies.

The typing experience on the G752VS is good but not perfect. On one hand the keys are well separated, their tactile feedback never missed the mark and the soft-touch material around the keyboard provided an adequate non-slip surface. However, I did find the notebook’s lower edge to be slightly too sharp for my liking so actually finding a comfortable typing or even gaming position was quite difficult.

Slightly above and to the left of the function keys are five dedicated Macro buttons which can have functions assigned from within ASUS’ ROG Gaming Center. Actually reaching them during an intense firefight is a bit tricky but they get the job done. There’s also a button that enables game streaming through the included XSplit Gamecaster software.

The trackpad provides a Dr. Jekyll / Mr Hyde experience that runs from awesome to absolutely infuriating. On one hand its resistance and physical left / right clicks are second to none. However, the palm detection is missing in action even with Windows’ own algorithms working for all they’re worth. Luckily this is something that can be fixed with a simple update.

As you could probably tell from the chart above, the ASUS ROG G752 has an enviable set of I/O connectors integrated into its somewhat portly chassis. The left hand side houses a Kensington lock, a pair of USB 3.0 ports and the Blu Ray drive.

Meanwhile, the right side is where most of the action happens with the power input, a LAN jack and display outputs in the form of a full sized HDMI 2.0 connector and a single mini DisplayPort 1.4 output. There’s also a pair of additional USB 3.0 connectors, a combo Lightning / USB-C and the usual headphone and mic jacks.

Unlike some other notebooks, there’s actually quite a bit happening below the G752. Not only is there an easily accessible upgrade bay that can be opened by detaching a single screw but there’s also a small window which shows the internal CPU and GPU heatsinks and a small grille for the included subwoofer.

Quite a few of the components can actually be swapped out of the G752 and are easily reachable. The 64GB of DDR3 memory is right there for the taking while the NVMe SSD lies buried under an aluminum heatspreader in the upper left hand corner of the image above. Meanwhile, the storage drive (in this case a 1TB Hitachi HDD) sits inside a caddy that can be removed by loosening a quartet of screws and gently pulling upwards and sideways. ASUS needs to be commended since they seem to be actively encouraging user upgrades unlike some of their competitors who lock these items behind Warranty Void if Removed stickers.

Test Systems & Setups / Competitive Landscape

True apples to apples comparisons in the notebook market are almost impossible since there’s such a wide array of components that vary from one prebuilt product to the next. With that being said, we’ve already tested out quite a few other high end notebooks over the last few months, some of which included desktop-class hardware. Below is a quick chart of the notebooks in this particular review and their respective specifications.

One thing to note is that all results we have in the gaming charts are done through a notebook’s DisplayPort connector so resolution consistency can be maintained.

DX12 Benchmarking

For DX12 many of these same metrics can be utilized through a simple program called PresentMon. Not only does this program have the capability to log frame times at various stages throughout the rendering pipeline but it also grants a slightly more detailed look into how certain API and external elements can slow down rendering times.

Since PresentMon throws out massive amounts of frametime data, we have decided to distill the information down into slightly more easy-to-understand graphs. Within them, we have taken several thousand datapoints (in some cases tens of thousands), converted the frametime milliseconds over the course of each benchmark run to frames per second and then graphed the results. This gives us a straightforward framerate over time graph. Meanwhile the typical bar graph averages out every data point as its presented.

CineBench R15 64-bit

The latest benchmark from MAXON, Cinebench R15 makes use of all your system’s processing power to render a photorealistic 3D scene using various different algorithms to stress all available processor cores. The test scene contains approximately 2,000 objects containing more than 300,000 total polygons and uses sharp and blurred reflections, area lights and shadows, procedural shaders, antialiasing, and much more. This particular benchmarking can measure systems with up to 64 processor threads. The result is given in points (pts). The higher the number, the faster your processor.

PCMark 8

PCMark 8 is the latest iteration of Futuremark’s system benchmark franchise. It generates an overall score based upon system performance with all components being stressed in one way or another. The result is posted as a generalized score. In this case, we didn’t use the Accelerated benchmark but rather just used the standard Computational results which cut out OpenCL from the equation.

RealBench

RealBench is an integrated benchmark distributed by ASUS which incorporates several real-world tests which are run, after which a score is calculated. In the standard combined test we are using, there are four consecutive tests being run: GIMP photo manipulation, Handbrake h264 video compression, OpenCL rendering using LuxMark and a Heavy Multitasking scenario using all three other tests at the same time. This may look like a simple benchmark result but it combines multiple tasks with single and multi threading in an effort to replicate an accurate depiction of system performance.

WPrime

wPrime is a leading multithreaded benchmark for x86 processors that tests your processor performance by calculating square roots with a recursive call of Newton’s method for estimating functions, with f(x)=x2-k, where k is the number we’re squaring, until Sgn(f(x)/f'(x)) does not equal that of the previous iteration, starting with an estimation of k/2. It then uses an iterative calling of the estimation method a set amount of times to increase the accuracy of the results. It then confirms that n(k)2=k to ensure the calculation was correct. It repeats this for all numbers from 1 to the requested maximum. This is a highly multi-threaded workload but we have also included single thread results. Below are the scores for the 1024M benchmark for multi thread and 32M for single thread.


AIDA64 Memory Read / Write

Memory bandwidth benchmarks (Memory Read, Memory Write) measure the maximum achiveable memory data transfer bandwidth. The code behind these benchmark methods are written in Assembly and they are extremely optimized for every processor variant by utilizing the appropriate x86/x64, x87, MMX, MMX+, 3DNow!, SSE, SSE2, SSE4.1, AVX, and AVX2 instruction set extension.

CPU PhotoWorxx Benchmark

This benchmark performs different common tasks used during digital photo processing. It performs a number of modification tasks on a very large RGB image:

This benchmark stresses the SIMD integer arithmetic execution units of the CPU and also the memory subsystem. CPU PhotoWorxx test uses the appropriate x87, MMX, MMX+, 3DNow!, 3DNow!+, SSE, SSE2, SSSE3, SSE4.1, SSE4A, AVX, AVX2, and XOP instruction set extension and it is NUMA, HyperThreading, multi-processor (SMP) and multi-core (CMP) aware.

Crystal Diskmark

Crystal Diskmark is a simple storage subsystem benchmark which measures sequential bandwidth across a large number of file sizes. For the purposes of this test, we are using a 1GB file size and measuring both read and write performance.

7-Zip

At face value, 7-Zip is a simple compression/decompresion tool like popular applications like WinZip and WinRAR but it also has numerous additional functions that can allow encryption, decryption and other options. For this test, we use the standard built-in benchmark which focuses on raw multi-threaded throughput.

Blender

Blender is a free-to-use 3D content creation program that also features an extremely robust rendering back-end. It boasts extremely good multi core scaling and even incorporates a good amount of GPU acceleration for various higher level tasks. In this benchmark we take a custom 1440P 3D image and render it out using the built-in tool. The results you see below list how long it took each processor to complete the test.

Handbrake

Video conversion from one format to another is a stressful task for any processor and speed is paramount. Handbrake is one of the more popular transcoders on the market since it is free, has a long feature list, supports GPU acceleration and has an easy-to-understand interface. In this test we take a 6GB 4K MP4 and convert it to a 1080P MKV file with a H.264 container format. GPU acceleration has been disabled. The results posted indicate how long it took for the conversion to complete.

GIMP Photo Manipulation

GIMP is a free, open source photo manipulation program which features several GPU-accelerated algorithms. In this benchmark we use a 30MB, 5K image and load a custom script to apply the following filters: gaussian blur, distort rotation, flare, stained glass and extrude. The score indicate is the time it takes the notebook to complete all tasks when run sequentially.

WinRAR

WinRAR is one of those free tools that everyone seems to use. Its compression and decompression algorithms are fully multi-core aware which allows for a significant speedup when processing files. In this test we compress a 3GB folder of various files and add a 256-bit encryption key. Once again the number listed is the time to completion.

Wireless Network Performance

Testing wireless network performance isn’t an easy thing since everything from environmental conditions to latent secondary network load can influence the results. In order to eliminate as many variables as possible, we used ASUS RT-AC5300 which is considered to be one of the most powerful 5GHz Wireless AC routers currently available and set up a dedicated access stream for just the notebooks being tests. We gave this access point absolute priority so its spatial steam wouldn’t be interrupted to constrained by any other elements.

The 5GHz Wireless AC spectrum has some amazing potential throughput numbers but the actual range in which those theoretical performance figures are attainable is limited at best. In this situation, the better the onboard router, the better the achievable performance but the situation is far from cut and dried. As such, we are testing through three separate zones. Zone 1 is located within 10 feet of the router and has a clear line of sight to it. Zone 2 is about 20 feet away but also forces the signal to pass between 3 walls. Finally Zone 3 is not only about 30 feet away but it also requires the signal to pass through walls, floors and a single concrete slab.

In order to achieve the benchmark numbers, we timed how long it takes to transfer files from the notebook in question, through the router and onto an Ethernet-connected external SSD. The first test focuses upon transferring a huge 2.62GB folder containing everything from images to word documents to program executables. Meanwhile the second test transfers a simple 4K video file.


3DMark

Synthetic testing is never an accurate way to depict potential in-game results but Futuremark’s 3DMark has the capability to run some straightforward tests and spit out a reasonably good representation of where a given system will land once we get into the actual games. In this test we run the Fire Strike, Fire Strike Extreme and DX12 Time Spy benchmarks, the former of which puts additional stress on the CPU while the latter two typically causes a GPU-centric bottleneck.



Doom (OpenGL)

Not many people saw a new Doom as a possible Game of the Year contender but that’s exactly what it has become. Not only is it one of the most intense games currently around but it looks great and is highly optimized. In this run-through we use Mission 6: Into the Fire since it features relatively predictable enemy spawn points and a combination of open air and interior gameplay.

Fallout 4

The latest iteration of the Fallout franchise is a great looking game with all of its detailed turned to their highest levels but it also requires a huge amount of graphics horsepower to properly run. For this benchmark we complete a run-through from within a town, shoot up a vehicle to test performance when in combat and finally end atop a hill overlooking the town. Note that VSync has been forced off within the game’s .ini file.

Far Cry 4

This game Ubisoft’s Far Cry series takes up where the others left off by boasting some of the most impressive visuals we’ve seen. In order to emulate typical gameplay we run through the game’s main village, head out through an open area and then transition to the lower areas via a zipline.

Grand Theft Auto V

In GTA V we take a simple approach to benchmarking: the in-game benchmark tool is used. However, due to the randomness within the game itself, only the last sequence is actually used since it best represents gameplay mechanics.

Hitman (2016)

The Hitman franchise has been around in one way or another for the better part of a decade and this latest version is arguably the best looking. Adjustable to both DX11 and DX12 APIs, it has a ton of graphics options, some of which are only available under DX12.

For our benchmark we avoid using the in-game benchmark since it doesn’t represent actual in-game situations. Instead the second mission in Paris is used. Here we walk into the mansion, mingle with the crowds and eventually end up within the fashion show area.

Overwatch

Overwatch happens to be one of the most popular games around right now and while it isn’t particularly stressful upon a system’s resources, its Epic setting can provide a decent workout for all but the highest end NOTEBOOKs. In order to eliminate as much variability as possible, for this benchmark we use a simple “offline” Bot Match so performance isn’t affected by outside factors like ping times and network latency.

Rise of the Tomb Raider

Another year and another Tomb Raider game. This time Lara’s journey continues through various beautifully rendered locales. Like Hitman, Rise of the Tomb Raider has both DX11 and DX12 API paths and incorporates a completely pointless built-in benchmark sequence.

The benchmark run we use is within the Soviet Installation level where we start in at about the midpoint, run through a warehouse with some burning its and then finish inside a fenced-in area during a snowstorm.

The Division

The Division has some of the best visuals of any game available right now even though its graphics were supposedly downgraded right before launch. Unfortunately, actually benchmarking it is a challenge in and of itself. Due to the game’s dynamic day / night and weather cycle it is almost impossible to achieve a repeatable run within the game itself. With that taken into account we decided to use the in-game benchmark tool.

Total War: Warhammer (DX11)

The newest Total War game not only looks great and in this benchmark we use a custom game scenario which is recorded as a timedemo with the in-game tools. The playback takes into account all elements of the game and runs for about 2 minutes.

Witcher 3

Other than being one of 2015’s most highly regarded games, The Witcher 3 also happens to be one of the most visually stunning as well. This benchmark sequence has us riding through a town and running through the woods; two elements that will likely take up the vast majority of in-game time.

1080P DX12 Results



Doom (OpenGL)

Not many people saw a new Doom as a possible Game of the Year contender but that’s exactly what it has become. Not only is it one of the most intense games currently around but it looks great and is highly optimized. In this run-through we use Mission 6: Into the Fire since it features relatively predictable enemy spawn points and a combination of open air and interior gameplay.

Fallout 4

The latest iteration of the Fallout franchise is a great looking game with all of its detailed turned to their highest levels but it also requires a huge amount of graphics horsepower to properly run. For this benchmark we complete a run-through from within a town, shoot up a vehicle to test performance when in combat and finally end atop a hill overlooking the town. Note that VSync has been forced off within the game’s .ini file.

Far Cry 4

This game Ubisoft’s Far Cry series takes up where the others left off by boasting some of the most impressive visuals we’ve seen. In order to emulate typical gameplay we run through the game’s main village, head out through an open area and then transition to the lower areas via a zipline.

Grand Theft Auto V

In GTA V we take a simple approach to benchmarking: the in-game benchmark tool is used. However, due to the randomness within the game itself, only the last sequence is actually used since it best represents gameplay mechanics.

Hitman (2016)

The Hitman franchise has been around in one way or another for the better part of a decade and this latest version is arguably the best looking. Adjustable to both DX11 and DX12 APIs, it has a ton of graphics options, some of which are only available under DX12.

For our benchmark we avoid using the in-game benchmark since it doesn’t represent actual in-game situations. Instead the second mission in Paris is used. Here we walk into the mansion, mingle with the crowds and eventually end up within the fashion show area.

Overwatch

Overwatch happens to be one of the most popular games around right now and while it isn’t particularly stressful upon a system’s resources, its Epic setting can provide a decent workout for all but the highest end NOTEBOOKs. In order to eliminate as much variability as possible, for this benchmark we use a simple “offline” Bot Match so performance isn’t affected by outside factors like ping times and network latency.

Rise of the Tomb Raider

Another year and another Tomb Raider game. This time Lara’s journey continues through various beautifully rendered locales. Like Hitman, Rise of the Tomb Raider has both DX11 and DX12 API paths and incorporates a completely pointless built-in benchmark sequence.

The benchmark run we use is within the Soviet Installation level where we start in at about the midpoint, run through a warehouse with some burning its and then finish inside a fenced-in area during a snowstorm.

The Division

The Division has some of the best visuals of any game available right now even though its graphics were supposedly downgraded right before launch. Unfortunately, actually benchmarking it is a challenge in and of itself. Due to the game’s dynamic day / night and weather cycle it is almost impossible to achieve a repeatable run within the game itself. With that taken into account we decided to use the in-game benchmark tool.

Total War: Warhammer (DX11)

The newest Total War game not only looks great and in this benchmark we use a custom game scenario which is recorded as a timedemo with the in-game tools. The playback takes into account all elements of the game and runs for about 2 minutes.

Witcher 3

Other than being one of 2015’s most highly regarded games, The Witcher 3 also happens to be one of the most visually stunning as well. This benchmark sequence has us riding through a town and running through the woods; two elements that will likely take up the vast majority of in-game time.

1440P DX12 Results



Battery Life Testing

Battery life is a key component of notebook testing since without adequate unplugged time, one of these systems could hardly be called “mobile”. In order to accurately measure how long a notebook can last away from the mains, we devised a quartet of tests. First is a light usage Browsing scenario which simply refreshes a webpage every 30 seconds while the second test runs an infinite loop of PCMark’s Work preset which simulates some light photo manipulation, word processing and Excel document creation.

The last two tests are more extreme in nature since they consume the majority of a notebook’s onboard resources. The first of these loops PCMark’s Creative benchmark, a testing suite that combines multi-threaded photo editing, GPU-accelerated video transcoding and web conference streaming. Finally the Gaming test uses Rise of the Tomb Raider output at 1080P to the notebook’s screen.

All of these tests are done in the notebook’s default power profile, with screen brightness set at 75% and set the notebook to go into Sleep mode with 5% battery remaining. To log results we use PassMark’s handy BatterMon tool which polls the battery levels every minute and logs them to a text file.


From an efficiency perspective the ASUS notebook reigns supreme even in its overclocked form. While the differences between it and the GTX 980-equipped MSI, we can’t forget that notebook tended to throttle due to temperatures and excess GPU power consumption.

The GTX 1070 is obviously an efficient graphics card but exactly how efficient is still unknown since all of the laptops we tested for the chart above have quite different specifications. Nonetheless, longer battery life alongside some impressive performance results is very much a possibility with Pascal.

Temperature Testing

In order to log thermal results we used AIDA64’s logging modules and loaded up the program’s CPU Burn test for any CPU-centric results while the GPU results are taken from an actual 30 minute gameplay sequence. Both are done while the notebook is plug in and displaying images through its screen. The recorded temperatures below are absolute maximums.

CPU Temperatures

GPU Temperatures

The amount of heat put out by 16nm Pascal GPUs is minimal at best when compared against previous generation designs and these charts affirm that. However, there may be something to say about ASUS’ cooling solution too but regardless of the source, the GTX 1070 remains cool enough that it doesn’t even come close to thermal throttling. That is likely why it was so far ahead in some of the gameplay tests.

Acoustic Results

What you see below are the baseline idle dB(A) results attained within a completely quiet sound-isolated environment in both in idle and load scenarios. The meter we use has been calibrated and is placed at seated ear-level exactly 12” away from the notebook. For the load scenarios, Tom Raider is used in order to generate a constant load over the course of 30 minutes.

Quiet is the name of the game here which is likely due to the combination of an efficient GPU alongside a very good internal cooling solution on ASUS’ notebook.

Conclusion; Notebooks to the Next Level?

So here I am at the tail end of another rapid-fire launch and for the first time in a while, I find myself a bit torn between giddy excitement and apprehension. It feels like a privilege to experience the power of notebook that can offer high end desktop level performance metrics without any major sacrifices and yet there are a few existential questions surrounding this launch.

There’s absolutely no denying that NVIDIA’s Pascal architecture will usher in a revolution within the gaming notebook space. The GTX 1080, GTX 1070 and GTX 1060 will be a dream come true for system integrators since they’re now able to shoehorn some impressive graphics horsepower into thin and light chassis. While I can only speak to my experience with a lone GTX 1070-based product thus far (more about the incredible ASUS G752 below), all indicators point towards these new GPUs being the quantum leap NVIDIA promised.

Perhaps the most incredible thing about this launch is how little has been sacrificed on the performance front. Previous generations of mobile GPUs cast aside cores, memory capacity and frequency to achieve TDPs necessary for their integration into even the largest of gaming notebooks. Indeed, while past GPU architectures claimed to offer desktop performance for the notebook market, the difference between theory and reality could be measured in light-years. Pascal graphics cores on the other hand have been ported as completely holistic entities with only some minor cut in base clocks. Not only will this provide true high end performance but it’s been achieved without revising those already-stringent notebook design specs and it’s those guidelines that have effectively cut AMD out of this segment. As a matter of face these cards are so damn capable that users could face the very real possibility of CPU bottlenecking at 1080P, particularly with low voltage mobile processors.

This brings me to the ASUS G752VS OC Edition and the GTX 1070 8GB beating at its heart. This is an incredible, albeit expensive solution that is priced well above most gamers’ affordability levels but it highlights both the strengths of NVIDIA’s architecture and how well ASUS’ engineers have integrated that technology into their notebook design. This gaming laptop is well built, feature-rich and, despite being pre-overclocked it is surprisingly well behaved in the acoustics and efficiency departments.

Unfortunately actually putting the notebook version of NVIDIA’s GTX 1070 8GB through any sort of direct competitive analysis is extremely difficult. Apples to apples comparisons for specific components in the notebook market is nearly impossible since everything from a product’s cooling efficiency to secondary hardware can drastically impact benchmark results.

With that being said, what I can tell you about the gaming results is this: I put the G752 up against some of this year’s fastest notebooks and it consistently came out on top, sometimes by a country mile. Despite the ASUS notebook utilizing a mobile-centric CPU versus the desktop i7 6700-series some others have, the only solution that was able to come remotely close was a loud, power hungry $4000 GTX 980M SLI setup from Eurocom. Even then, the SLI system’s dual card nature proved to be its downfall since scaling varied wildly from one title to the next. Regardless of whether you choose to use ASUS’ integrated 1080P G-SYNC display or output to a higher resolution monitor, the GTX 1070 won’t disappoint.

The in-game framerates provided by the lone GTX 1070 8GB are noteworthy but the secondary experience may be the bigger news here. That record-setting performance is in part due to the GPU core remaining well under its thermal throttle limit, allowing NVIDIA’s Boost algorithms to project frequencies into the stratosphere. This is accompanied by some notable battery life improvements which may not look groundbreaking in our charts but really do represent a big step forward.

The worries I mentioned earlier largely stem from the complete lack of competition in this segment. This isn’t something anyone can hold against NVIDIA since they’ve been killing it as of late but it is nonetheless something that needs to be discussed. Due to their inefficient core architectures, AMD has been completely shut out the high end notebook market for the last few years and knowing Polaris’ performance per watt challenges, that situation may not change anytime soon. This has left NVIDIA with an unassailable position and as we all know, near-monopolies can lead to increased prices; just look at the current CPU market for evidence of that. How that translates to future notebook prices is anyone’s guess but I’m hoping some semblance of affordability remains since these Pascal GPUs are really game changers for mobile gamers.

So more than a thousand words into this conclusion and I think it’s time to wrap things up. The GTX 1070 that’s destined for notebooks brings to the table extremely high levels of performance without any of the sacrifices normally associated with gaming laptops. How it is integrated going forward will be determined by NVIDIA’s partners but for the time being there’s very little reason for concern, you just have to pick the right combination of components and pricing to fit your needs.

Posted in

Latest Reviews