We tested and compared the Apple MacBook Pro versus Razer Blade 15 in terms of (Gaming) Performance, Price, Display Quality, Portability, Battery & more.
Above you can see the Ranking with the results and below you will find the in-depth test reports of the Apple & Razer Laptop.
Ranking First: Razer Blade 15
- Best GPU (Graphics card) performance
- Quieter than most other gaming laptops in gaming mode / under load
- High quality 144 Hz display
- More expensive than MacBook Pro
Sharp blade. Although the Blade-15 series is only one year old, it already covers two Nvidia GeForce graphics card generations and offers display options that include everything from 144-Hz FHD to OLED. In this review, we put the RTX-2070-Max-Q-SKU to the test and look at how it compares to the GTX-1070-Max-Q variants.
Razer unveiled the new Blade 15 with the GeForce RTX-Turing graphics card at CES 2019 along with the 240Hz and OLED display configurations. While the latter will have to wait until Q2 2019, the Blade 15 RTX with GeForce RTX 2060, RTX 2070 Max-Q or RTX 2080 Max-Q is already available.
This is also the first time that a class 80 Nvidia GeForce graphics card can be found in the Blade 15 series, which was otherwise reserved for the larger 17.3-inch Blade Pro series.
Our current test device is the RTX-2070-Max-Q-SKU, which is equipped with a 512 GB Samsung NVMe SSD and a matte 144 Hz display. A 256 GB NVMe SSD and a reflective 4K UHD touchscreen are also available. The Blade 15 with RTX 2070 Max-Q graphics card costs between 2,400 and 2,900 US dollars depending on the configuration.
The competition in this area is constantly growing. The 15.6 inch gigabyte Aero 15 Y9, Dell Alienware m15, Aorus 15, MSI GS65 and Asus ROG GL504 are just a few examples of other devices with GeForce RTX graphics. In this review we also compare the Blade 15 RTX 2070 Max-Q with the Blade 15 GTX 1070 Max-Q and look at how much additional performance you can expect from the new model.
We refer to the existing Razer laptop reviews below, as many of our comments apply here as well. The design of the Blade-15 series has hardly changed since its inception.
Should Apple ever offer a black MacBook Pro 15, it will probably look very similar to the Blade 15. Razer laptops are often praised for their seamless, minimalist design, and the Blade 15 series will continue to maintain this.
All of our existing comments about older Blade 15 Pro SKUs apply here as well, as the case material, hinges and quality have not changed. The system continues to feel more stable and less flexible than the Alienware m15, MSI GS65 and Gigabyte Aero 15, which together with the case that doesn’t creak under pressure, makes for a captivating first impression.
From future iterations, we would like to see a more stable lid and a narrower lower screen edge, which would allow for an even more compact layout, as for example in the ZenBook 15 UX533.
Even though it is by no means weak in any way, the lid is more flexible than on the XPS 15, but these are just small details that contrast with an otherwise appealing and stable case.
The RTX Blade 15 is a bit thicker than the GTX Blade 15 from last year. Razer had to slightly increase the Z-height in order to accommodate the physically larger Nvidia Turing chip. Length and width are otherwise identical to all previous Blade 15 SKUs, making it one of the smallest 15.6-inch gaming laptops available.
The dimensions are very similar to the XPS 15, but a much stronger graphics card is built in. Please note that the Alienware m15 is significantly larger than the Razer, although it is advertised as a thin gaming laptop.
In terms of weight, the Blade 15 is heavier than the MSI GS65 and at the same time it is more compact, which makes it look denser and more powerful.
Nevertheless, it is significantly lighter and more portable than thicker 15.6-inch gaming laptops like the MSI GT63 and the Alienware 15. It should also be noted that the 4K UHD SKU is slightly heavier due to the additional reflective touchscreen.
The connection equipment and position remain identical to that of last year’s GTX Blade 15 professional model despite a slight increase in thickness. It should be noted that RJ-45 is only available on the GTX Blade 15 base model for the time being. Users hoping for an integrated SD card reader will have to cross their fingers for the next case version.
Razer has replaced the Intel 9260 in last year’s GTX Blade 15 model with the newer Intel 9560. The WiFi modules’ performance is otherwise identical, as both support Bluetooth 5 and WiFi transfer rates of up to 1.73 GBps.
The Intel 9560’s advantages over the Intel 9260 are instead limited to LTE coexistence and don’t affect the Razer system. Competing models like the Alienware m15 are equipped with the Killer 1550, which offers more gaming oriented functions than our Razer.
During our tests, no connection problems occurred in our test device. The actual performance can exceed the results recorded below because our server configuration is limited by a 1GBps line.
The cover on the bottom can be easily removed after loosening the T5 Torx screws. The internal layout and upgrade options are almost identical to those of the Blade 15 professional model, as you can see below.
Purchases made in the United States are covered by a one-year limited warranty. The optional RazerCare program extends the warranty by two years, for a total of three years.
Keyboard & Touchpad
The chroma-capable keyboard is found here unchanged. We hope that dedicated macro keys or volume hotkeys, which simplify multimedia control, will be added in future models.
The key stroke is very small and the pressure point soft. We like the clearer key feedback of competing models, like the Alienware m15 and GS65, better.
The customisable RGB key illumination uses pulse width modulation for brightness control. This means that the LEDs switch between “on” and “off” very quickly, which could potentially lead to eye problems for a few users. This phenomenon is also associated with the screen flickering of some laptops.
The roomy touchpad (13 x 8 cm) is larger than the touchpads on the MSI GS65 (10.5 x 7 cm) and the Alienware m15 (10.5 x 6.5), making pointer control and multi-touch input easier at any speed. Pointer movements feel especially fluid thanks to the panel’s native 144 Hz refresh rate.
The touchpad’s biggest drawback is still the soft key feedback. Instead of the quiet acoustic feedback, we would have liked a clearer and more satisfying click here.
Razer has made no changes to the display compared to last year’s professional model. In this case, the LG Philips LGD05C0 1080p IPS panel with 144 Hz is the same as the Blade 15 GTX 1070 Max-Q and the Gigabyte Aero 15x v8. Key data such as black to white response times, color space coverage and contrast are therefore excellent and comparable for these laptops.
This is a thoroughly high-quality panel that is free of image grain effects and other characteristics that often plague matt panels. The 60Hz panel in the Blade 15 base model has significantly slower black-to-white response times, resulting in more noticeable ghosting.
As mentioned earlier, future Blade 15 SKUs will offer 240Hz and OLED display options that deliver near-perfect black levels and smoother motion.
Razer states 100% coverage of the sRGB color space for the FHD SKU and 100% coverage of the AdobeRGB color space for the 4K UHD touchscreen SKU. In our independent measurements, the former achieved 95 percent sRGB coverage, making it similar to the MSI GS65 and Gigabyte Aero 15X v8.
Users requiring an even wider color gamut for professional applications will have to forego the native 144 Hz refresh rate until such a panel is available for laptops.
Further measurements with an X-Rite spectrophotometer show that the panel already displays colours well ex works. The impact of Razer’s promise to calibrate each unit before shipping from the factory is clearly seen in our CalMAN results below.
Despite this, our calibration tests have further improved greyscale and colour accuracy, resulting in average DeltaE values of 1.4 and 1.71 respectively. This even comes close to the MacBook Pro 15.
Outdoor display readability is neither better nor worse than most other gaming laptops because of the standard 300 cd/m² backlighting on the matte panel. It’s a pity, after all, the Blade 15 is thinner and more mobile than many of its competitors, which probably tempts owners to use the device more often on the go.
In direct sunlight, colours may appear washed out, but dazzling reflections can be minimised by taking advantage of the wide viewing angles of the screen. As long as you don’t look at the display from extremely steep angles, neither the colors nor the contrast will be distorted.
The Coffee Lake-H Core i7-8750H CPU returns and this time it has the new GeForce RTX 2060, RTX 2070 Max-Q or RTX 2080 Max-Q graphics card in tow.
In contrast to many competing laptops, there is no Core i5 configuration and no variant with less than 16 GB of memory that could have lowered the base price. Razer instead concentrates fully on offering powerful gaming laptops, as the data sheet shows.
While the memory configuration is identical to the Blade 15 (DDR4-2666), each SODIMM slot now supports up to 32 GB, making a total of up to 64 GB possible. Earlier models were limited to a total of 32 GB. Since all SKUs are configured for Optimus technology by default, G-Sync is not supported.
At the time of this article, the same Core i7-8750H is used in all Blade 15 SKUs. Unfortunately, the multi-thread performance with our Razer laptop is 10 percent lower than the average i7-8750H in our database, which includes 60 other laptops with this processor.
The Alienware m15 scores 22 percent higher than our Razer with the exact same CPU in the same CineBench benchmark. In fact, the multi-core performance overall rather resembles a Core i5-8300H. This result is due to the poor preservation of Turbo Boost, as the stress test section will show.
By running the CineBench R15 multi-thread test in a loop and recording the results (see below), we can determine the effects of CPU throttling. In the second run, the score drops from 999 to 825, and comparatively speaking, the Alienware m15 can maintain higher scores, averaging 1,164.
You can find more technical details and benchmark comparisons on our Core i7-8750H page.
In the PCMark benchmarks, the deviations between our system and the Blade 15 GTX 1070 Max-Q-SKU are in the single-digit percentage range. Thus, the upgrade from Pascal to RTX doesn’t have a significant impact on the performance in everyday use.
Even the result of the Digital Content Creation Benchmark, which is more strongly influenced by the graphics card performance, is only 11 percent higher in the new Blade 15.
In contrast to the Blade Stealth, which has to struggle with some problems, we didn’t encounter any hardware or software problems with this Blade 15 tester.
A 512 GB Samsung PM981 NVMe-SSD is used in our test device, whereas it was still a 512 GB PM961 in last year’s Blade 15. Users who need a second storage bay have to fall back on the older Blade 15 base model or opt for the MSI GS65 or the Alienware m15.
The sequential write rate of the PM981 is higher than the PM961 and most SSDs from competing manufacturers. For example, the Toshiba XG5 in the Alienware m15 has an almost 40 percent lower sequential write rate. We appreciate the decision to limit the entire Blade 15 and Blade Pro series to Samsung SSDs.
So how much graphics performance can users expect in comparison to the original Blade 15 GTX 1070 Max-Q? According to 3DMark Fire Strike and Time Spy, the new RTX 2070 Max-Q is about 15 percent faster, making it almost as fast as the GTX 1080 Max-Q from the last graphics card generation.
Furthermore, the ray tracing performance in the new Port-Royal benchmark is about 20 percent higher than the RTX 2060, which is used as standard in laptops. The leap from one generation to the next is definitely smaller with Pascal to Turing than with Kepler to Pascal when considering pure gaming performance.
If you decide on an RTX 2070 Max-Q, you should use ray tracing technology to get the most out of the new GPU architecture.
Since the Blade 15 is one of the very first RTX gaming laptops, there aren’t many benchmarks we can compare it to, so we don’t have a feel for how it compares to other RTX laptops. Of course, our database is growing over time, which will allow us to better assess how the Blade 15 is doing overall.
Like the rest of the RTX graphics cards, the RTX 2070 supports Max-Q for increased gaming performance DLSS. However, developers must first implement this new Nvidia-exclusive technology, as it is currently only supported by a few titles.
In practice, the gaming performance is roughly on par with a GTX 1080 Max-Q, or about 30 to 35 percent below a desktop RTX-2080. To fully exploit the potential of the 1080p display and its 144 Hz refresh rate, a preferably constant refresh rate of 144 Fps is desirable, but difficult to achieve in demanding titles.
However, the system is perfectly suited for typical eSports titles such as DOTA 2, Rocket League, CS:GO and Overwatch.
In 4K gaming, you can expect an average of 42 fps in Battlefield 5 on the “Ultra” graphics setting, which is comparable to a GTX 1080 Max-Q. If you decide on the more expensive Blade 15 RTX 2080 Max-Q, you can expect an increase in performance of about 20 to 30 percent compared to the RTX 2070 Max-Q.
There are no frame drops during idle in The Witcher 3, which shows that in gaming mode no interference from background activity is to be expected.
Noise & temperature levels
Apart from the slightly increased thickness, the cooling solution is visually identical to that of last year’s Profi model. In the energy-saving profile, both fans remain switched off at low loads such as surfing or video streaming, which makes the device practically silent during operation.
If you choose the high performance profile, you can expect a fan volume of 29.4 db(A) with a background volume of 28.8 dB(A).
It gets more interesting at higher loads because the fan noise seems to be lower than on the GTX-1070-Max-Q-SKU. In the first scene of 3DMark 06, for example, the maximum fan volume was 35.7 dB(A) instead of 46 dB(A).
The model from the last generation and also The Witcher 3 shows a similar difference. The Razer is quieter than the competition in gaming mode, including the Alienware m15, MSI GS65 and especially the Gigabyte Aero 15X. The highest recorded volume, which temporarily occurred when Prime95 and FurMark were running simultaneously, was 45 dB(A).
Electronic noise and coil beeping didn’t occur in our test device.
The surface temperatures are the same on both sides of the laptop when idle. The temperature of the palm rests is always slightly above the ambient temperature, which is why these surfaces always feel warm.
In opposition to the keyboard’s center, which heats up to 48°C during gaming, the WASD keys stay relatively cool. However, the wrist-rest then reaches a temperature of a good 30 °C, which we find quite warm.
The Alienware m15’s wrist-rests stay a bit cooler than those of the Razer laptop under about the same gaming load.
Fresh air is sucked in from the side of the keyboard and the warm exhaust air is expelled from the back of the laptop. This approach helps to keep the heat away from the fingertips and the external mouse.
To detect potential throttling and possible stability problems, we expose the notebook to synthetic loads. Under Prime95 load, the processor reaches a clock frequency of 2.8 GHz in the first few seconds and then settles down to a constant 2.3 to 2.4 GHz.
As the Core i7-8750H’s base clock is 2.2 GHz, the Blade 15 can thus maintain a Turbo Boost of +100 to +200 MHz. These results coincide with the CineBench R15 loop test above, where the scores are also below the expected values for this CPU.
However, the below average clock frequency has one advantage: lower core temperatures. The processor in the Blade 15 stabilizes at only 60 °C compared to 80 °C in the Alienware m15, which is equipped with the exact same i7-8750H processor. However, the Alienware’s clock frequency also settles down at 3.6 GHz, whereby it can unfold a measurably higher processor performance.
If you let Prime95 and FurMark run at the same time, CPU and GPU settle down at 80°C and 75°C respectively. Interestingly, the CPU ran a bit faster at 2.6 GHz under these extreme conditions. This behavior is extremely unusual because the CPU usually clocks lower when Prime95 and FurMark are running simultaneously.
Running The Witcher 3 however represents practical gaming loads better than the previously mentioned benchmarks. Here the CPU and GPU stabilize at 80 and 74 °C respectively, the GPU gets a constant boost clock rate of 1.410 MHz. The lower CPU performance in Prime95 and CineBench fortunately has no negative impact on the gaming performance.
The performance is reduced in battery mode. It should be noted that in battery mode, an Fps limit of 60 is enforced even in “prefer maximum performance” mode.
Razer laptops have relatively large battery capacities despite their small thickness. The Blade Pro GTX 1080, for example, is equipped with a 99 Wh battery, just like the thicker Alienware 17.
The battery life is above average for a gaming laptop with about 6.5 hours of practical WiFi use. At best, you can expect up to 11.5 hours of battery life if you stay completely idle on the lowest brightness level on the desktop.
Thus, the runtimes are even higher than in our original GTX-Blade-15 configuration, which is probably due to hardware optimizations or reduced background activity in our RTX model.
Fully charging the empty battery takes 1.5 hours, which is a normal value for many laptops. Unlike the MX150-based Blade Stealth, it is possible to use and charge the device for gaming at the same time.
Just like the Blade 15 , this early 2019 model is in our opinion one of the best thin & light gaming laptops on the market. It has the same advantages (thin and stable case, narrow display edges, pre-calibrated display) and the same disadvantages (low keystroke, limited CPU turbo boost, no SD card reader, RJ-45 and no second internal memory slot) as the last version of the Blade 15 professional model.
Apart from the expected GPU change, Razer has also upgraded the WiFi module and memory from Intel 9260 to Intel 9560 and from Samsung PM961 to PM981, respectively.
In terms of graphics card, the RTX 2070 Max-Q is faster than the GTX 1070 Max-Q without being louder or warmer. The gaming performance is, if you leave out the ray tracing technology, roughly on par with a GTX 1080 Max-Q laptop.
So if you don’t necessarily need ray tracing and DLSS (these are algorithms that 99% of all PC games don’t support anyway), you only get about 15% more performance here than on the much cheaper Blade 15 GTX 1070 Max-Q. Therefore, the older Blade 15 professional model is a much better offer for many users looking for a thin & light gaming laptop.
A lot hasn’t changed and that’s a good thing. Everything that we like about the Blade 15 is here again together with Nvidia RTX, which is why the Razer Blade 15 is ranking first versus Macbook Pro.
Ranking Second: Apple MacBook Pro
- In everyday life most very quiet and cool
- Better Price than Razer Blade 15
- Good Battery
- Bad Webcam
Now with four cores. Apple is refreshing the MacBook Pro 13 with a new quad-core processor and True Tone Display, but only on the more expensive model with the Touch Bar.
In this review, we take a close look at the performance of the new MacBook Pro, among other things. What can be expected for the advertised price of $1500 (i5)?
Secretly, quietly and silently, Apple has updated its MacBook Pro range. Externally nothing has changed, but the manufacturer from Cupertino finally relies on Intel processors of the 8th generation with four cores (Coffee Lake U) for the processors. Apart from that, the displays now also support True Tone technology, which is already known from the mobile iPhone and iPad.
This review is about the smaller 13 inch model. It’s noticeable that only the more expensive model with the touch bar gets the updates. The cheaper version with the function keys is only sold with the 2017 specification. Apple may be orienting itself on the MacBook Air, which hasn’t received an update for a long time and has therefore become cheaper and cheaper over time.
However, we don’t like this development, because as we could see in the reviews of the 2017 versions, the more expensive touch bar model wasn’t the better choice for every customer.
There are no changes in terms of price, as it still starts at $1500. In return, the customer gets a Core i5-8259U including Iris Plus 655 graphics, 8 GB LPDDR3-2133 RAM and 256 GB SSD memory. The RAM can be expanded to a maximum of 16 GB and the mass memory to up to 2 TB, whereby the surcharges are, as usual, extremely high.
We’ve known the case design of the MacBook Pro 13 since 2016, and the current model is also available in both silver and space gray. There is still no criticism from us in terms of quality, which also applies to the feel of the metal surfaces.
From our own experience, we can say that the cases present themselves in excellent condition even after a longer period of use, whereas plastic cases are more likely to show optical signs of wear (greasy surfaces, etc.).
That the case is now entering its third year can be seen in its weight and dimensions. The MacBook Pro 13 is by no means a heavyweight with a little more than 2.86 lbs and also not overly large, but the competition has noticeably caught up and partly overtaken in the last few years. There are more and more devices with thinner screen edges, which makes larger panels possible, and the weight is also going down further.
The basic connectivity hasn’t changed at all, as there are still 4x Thunderbolt 3 with USB-C plugs and a 3.5mm audio plug. Due to the change to the Coffee Lake U CPUs, however, more PCIe lanes are available this year, so that all connections are operated at full speed (the two right-hand connectors were previously only at 2x speed).
There are more and more peripherals with USB-C, but in everyday life you won’t be able to avoid a dongle or adapter.
In last year’s Apple MacBook Pro models, we could find WiFi connection problems under certain circumstances when several USB-C devices were connected. With the version, there were also occasional packet losses and higher ping times, but the connection was maintained.
The communications module in the new MacBook Pro has also received an update. The biggest improvement is the support of Bluetooth 5.0, but the WiFi module’s transfer rates (especially when sending data) turn out worse than last year’s model.
Apple customers still have to do without a WWAN module, and Apple is fully committed to combining this with smartphones.
Apple equips the MacBook Pro with the specially developed T2 chip, which combines several functions, including encrypted memory and secure boot functions.
But the user doesn’t notice anything of this. To protect access to the laptop, the touch-ID fingerprint scanner is available, just like on last year’s model. In practice, this works very quickly and reliably.
Keyboard & Touchpad
The current keyboard of the Apple MacBook Pro has been the subject of lively discussion since its release. Due to the thin base unit, Apple has built a keyboard with an extremely limited stroke.
Thanks to Apple’s own butterfly keys, the (quite loud) strokes are extremely precise, but the typing feel simply doesn’t please every user. A big advantage is, however, the integrated backlight, whose intensity can be adjusted in 15 levels.
It is also automatically adjusted to the surrounding light. Actually nothing special, but very practical and unfortunately not available in most Windows laptops.
Recently there have also been more and more reports of problems with defective keyboards in Apple notebooks, whereupon Apple has also set up a special repair program. Even small things (such as dirt, crumbs, etc.) can cause problems (keys no longer work).
Apple has not addressed this problem further when announcing the new models, but the manufacturer is advertising a quieter keyboard. Our colleagues from iFixit have already taken a model apart and found a new silicone buffer under the keys.
Apart from a quieter typing noise, this probably also prevents foreign objects from entering the keyboard and should therefore be a reaction to the previous problems. However, we can’t of course check whether this solution will really work in the long run in this short time.
Subjectively, you can clearly see the difference between the old keyboard (for example in the 2017 MacBook Pro 15) and the new one. The stroke is clearly dampened and also a bit more sluggish.
The difference in volume actually only becomes apparent when the keyboard is used with distance and relatively swinging. Overall, it is difficult to say whether the touch has improved. In the beginning, however, a short period of getting used to the keyboard is necessary.
Update: Apple’s internal service documents have now confirmed that the new silicone membrane is designed to prevent the penetration of dirt particles.
Nothing has changed on the large touchpad this year. It is still the reference for touchpads and can be operated perfectly. Thanks to the large surface, gestures with several fingers are also no problem.
Under Windows (Bootcamp) the basic operation also works, but the optimization is not quite as good as with Apple’s own operating system. Especially moving files or windows is more complicated.
Nothing has changed on the touch bar above the keyboard either. Here, different symbols can be displayed depending on the context.
In normal use, these are for example the usual functions for screen brightness or volume, but you can also optionally display the normal function keys (F1 – F12). Whether the Touch Bar offers a real productivity advantage depends greatly on the software support.
The basic characteristics of the 13.3-inch IPS panel have not changed this year: 2,560 x 1,600 pixels (227 DPI) aka retina display, 500 cd/m² brightness and complete coverage of the P3 colour space. What’s new, however, is the support of TrueTone technology, which we already know from the iPhone and iPad.
This automatically adjusts the white balance to the colour temperature of ambient light, which is supposed to make for more natural vision. A manual night mode (Night Shift), which reduces the blue color components of the image, is still available.
This can also be adjusted in the system settings (automatic activation depending on the time of day, warmer or cooler image). Subjectively, TrueTone is very pleasant and you quickly get used to it.
The display is called APPA03E, which is different from last year’s two models (APPA033 and APPA034 respectively). However, it’s not clear whether this is simply a new panel revision or whether you need a new panel due to the TrueTone function.
We determine an average brightness of a bit more than 500 cd/m² for our MacBook Pro Retina in this case. Thus, the panel does fulfill the factory specification, but our two models from last year (especially the model without touch bar) were even brighter.
In return, the illumination is better and the black value also turns out lower at 0.33 cd/m². Thus, black image contents appear more saturated and the contrast ratio could also be increased once again (>1,573:1).
The display’s calibration ex factory also turns out a bit worse than last year, at least in comparison to the touch bar model. However, we criticize here on a very high level, because already ex factory all deviations (in comparison to the P3 reference) are below the important limit of 3.
The P3 color space is covered to 99% according to CalMAN. The smaller sRGB color space is also completely covered, the more extensive AdobeRGB reference is covered to around 77 percent. These values are on the level of the predecessor.
Apple’s compact MacBook Pro 13 Touch Bar already relied on Intel processors with a consumption rating of 28 watts in previous years, so it had a noticeable performance advantage over almost all other sub-notebooks equipped with the usual 15-watt models. Another advantage was the integrated eDRAM cache, which helped both the processor and especially the integrated GPU.
This year there is a significant performance boost when it comes to using multiple cores, thanks to the new quad-core processor. In addition, the two available CPUs now offer an eDRAM cache twice as large (128 vs. 64 MB), which should bring an additional punch.
So far Apple has been able to use the performance of the CPUs very well in the long run, so we are already looking forward to the results of the new quad core.
Due to the implementation of a real quad-core processor, the new Apple MacBook Pro 13 is sufficiently equipped for most requirements. Only the GPU performance can become a bottleneck, depending on the requirements, although there are also expansion possibilities with external GPUs. We will go into this in more detail in the chapter “Graphics Card”.
While Apple is now switching to DDR4 in the larger 15-inch MacBook Pro, the smaller 13-inch system will still be delivered with a maximum of 16 GB LPDDR3 RAM. LPDDR4 is not supported by Coffee-Lake-U, and Apple was probably a thorn in the side of the increased power consumption of DDR4-2400.
Practically all new subnotebooks have been using Kaby-Lake refresh CPUs since the end of last year, which now offer four instead of two processor cores and thus significantly improve multi-core performance. Apple is now catching up or overtaking the competition thanks to the use of Intel 28-watt CPUs in its MacBooks, which are based on the Coffee Lake architecture (Coffee Lake-U).
These are therefore more “castrated” Coffee Lake chips than highly bred Kaby Lake R models, which also differ in the manufacturing process (14nm++ vs. 14 nm+). The new manufacturing process also includes an updated microcode against Spectre v2.
Apple offers two different models from the Coffee Lake U series, the Intel Core i5-8259U and the Core i7-8559U. The only difference, however, is the clock (CPU and iGPU), as both models are quad-cores including hyper-threading, which don’t support vPro.
In our entry model, the Core i5 works with a base clock of 2.3 GHz, but thanks to Turbo Boost, frequencies of up to 3.8 GHz are possible when using a maximum of 2 cores; it’s still 3.6 GHz with all 4 cores. An interesting detail: According to Intel, the price difference between the two models is “only” 111 US dollars.
Of course, there are different purchase prices depending on the manufacturer, but the surcharge of 360 Dollars for the Core i7 is extremely high.
In the past, Apple has almost always given Intel processors free rein, making temperature the only limiting factor. This is also the case here, and we see consumption values of almost 50 watts. In the last few months, we have noticed that the Kaby-Lake refresh quad-core ULVs have become real hotheads from clocks of 3 GHz, and we were naturally curious to see if this would change with Coffee-Lake-U and the improved 14nm++ process.
The result in the Cinebench R15 single test is still within the expected range, but when we saw the result of the multi-test, we were surprised. Based on our previous experience with the Kaby-Lake-R quad-cores, we had expected a value of more than 700 points, but it is just 608 points (tested under macOS).
Thus, the test device falls behind systems like Dell’s XPS 13 or the ThinkPad X1 Carbon. Because many manufacturers also dynamically control the TDP, even many models are faster with the quite normal Core i5-8250U.
But what is the reason for this? Under macOS we monitored the processor’s clocking and power consumption with the Intel Power Gadget tool. In the following screenshot you can see the behavior of the processor during a single multi-run.
At the beginning of the test, the processor works at the full 3.6 GHz for about 5 seconds, which corresponds to a consumption of just under 48-49 watts. But the cooling doesn’t keep up with this for long and the CPU temperature rises to 100 °C, whereupon the performance has to be reduced.
As a result, the consumption fluctuates extremely strongly (between 20-40 watts), which naturally also affects the clock rate. And this already with only one run of the multi-test.
Due to the high temperatures, however, even this performance level cannot be maintained. In the further course of our Cinebench multi loop, the MacBook settles down at around 580 points, which means that it loses about another 5% in performance.
Update: Apple has meanwhile released a software update, which noticeably increases CPU performance.
What particularly surprised us is the apparently poor efficiency of the Coffee Lake U processor. We’ve been intensively occupied with the Kaby-Lake-R quad-core processors in the past few months, as they are found in practically every current subnotebook.
The normal Core i5-8250U is actually the best example here, because in the case of the ThinkPad T480s it is also only limited by the temperature, just like in the MacBook. With a consumption of around 30 watts (thus even a bit less than the new MBP 13), it works stably at 3.4 GHz and reaches more than 700 points.
We will of course investigate this further, but for now Coffee-Lake-U seems to have disadvantages compared to the Kaby-Lake-R chips when it comes to efficiency. However, this is not only a problem of Apple, but also of Intel, of course.
This naturally raises the question why Kaby-Lake-R was not chosen for the new 28-watt models. The new MacBook Pro is thus faster than the old 2017 model, but some Windows computers do better with the supposedly slower processors.
Thus, the performance advantage of the optional Core i7 processor should only turn out marginal, maybe it even works slower than the i5 due to higher temperatures (especially under constant load).
As we were a bit perplexed by the comparatively low Cinebench results, we wanted to get to the bottom of it a bit more precisely. One possibility is the massive fluctuations of the clock rate. As the processor immediately shoots up again with some leeway, it has to cool down again for some time at a low clock rate.
So we wanted to see how the processor or cooling behaves when we set the consumption manually. Under Windows you can use the Intel XTU (freeware) tool for this, which offers a lot of setting options. Under macOS things look more difficult.
Here we use the Volta tool for testing purposes, which doesn’t offer as many settings and is also subject to a charge (7-day trial). In addition, we can only set the consumption up to the specified 28 watts, more is not possible. We’ve also summarized more information, including the larger MacBook Pro 15, in this article.
First we limited the consumption to 31 watts, since this value is also reached with the default settings. The result is surprisingly good, as the four processor cores work at around 300 MHz more and the results of the Cinebench loop are stable.
Under macOS (blue curve) the performance is slightly lower due to the 28 Watt limitation, but stable and still better than with the old default settings. After the software update the CPU performance has increased and remains stable, which can be seen on the green curve.
Thanks to the fast processor, sufficient RAM and a very fast PCIe SSD, the system reacts quickly to input and there are no long waiting times. The installation of Windows via Bootcamp no longer requires a USB stick; an image file from Windows is sufficient.
Operation under Windows also works without any problems, but the results in the synthetic PCMarks are not particularly good and the test device can only outperform its predecessor in the PCMark 8 work test; in the other two scenarios, both last year’s models are faster.
A possible explanation is the missing optimization of the BootCamp driver for the MBP 13, which was released only a few days ago.
The MacBook Pro 13 is only available with fast PCIe memory. A drive with 256 GB memory is used in our test device; a maximum of 2 TB is optionally available ex factory.
This is fast memory, but it’s not easy to determine the exact performance data, as our usual tools only work to a limited extent and also depend a little on the optimization of the BootCamp drivers. At most, we determine a maximum of just under 3 GB/s when reading, so there are no speed problems here.
The graphics card integrated in the Apple laptop is an old acquaintance, because basically the Intel Iris Plus Graphics 655 is just a regular Iris Plus Graphics 650 with a larger L4 cache (128 MB vs. 64 MB eDRAM).
We are still dealing with the GT3e expansion stage of the Kaby-Lake GPU, which achieves a maximum clock rate of 1,050 MHz in combination with the Core i5-8259U. If you decide on the faster Core i7 in the MacBook, the maximum is 1.200 MHz. More technical information about the Iris Plus Graphics 655 is available here.
If you need more graphics performance, you should fall back on an external Thunderbolt solution, which is now natively supported by macOS. Thanks to the new quad-core processor, there should be no major limitations by the CPU.
Apple itself, for example, offers a solution from BlackMagic that uses a Radeon Pro 580. We will soon receive a similar solution from Gigabyte (RX580 Gaming Box) for testing and of course we will also try it with the new MacBooks.
The integrated GPU can hold its own well in the synthetic 3DMarks. In the older 3DMark 11, the new MacBook Pro 13 can position itself about 10% ahead of the predecessor as well as the dedicated GeForce MX130, but the MX150 is still a bit faster.
The picture shifts a bit in the demanding 3DMark 13 test, like Fire Strike or Time Spy. The predecessor is overtaken by 10-20%, depending on the test, while the dedicated models are also ahead of the test device. The comparison with the two integrated AMD GPUs, Vega 8 and Vega 10, is also interesting, but they beat the Intel iGPU in all tests.
The integrated Iris Graphics 655 is only sufficient for very undemanding games, but the performance remains constant over a longer period of time.
However, there is no performance advantage over the predecessor. We didn’t have any driver problems under Windows, by the way.
Noise & temperature levels
The cooling system for MacBook Pro 13 has not changed this year. Furthermore, two fans take care of the processor’s waste heat, with the speed being continuously adjusted to the load.
In idle and under low load, the two fans rotate (1,294/1,285 rpm), but are only audible at 30 dB(A) when you hold your ear directly against the case. This changes under load, whereby the speed is adjusted quite quickly and the speed level therefore often fluctuates within a certain range.
By the way, the two fans’ maximum speed is 6.336/6.864 rpm, which corresponds to a sound pressure level of 46.7 dB(A).
We didn’t reach this value during testing, though, but a maximum of 46.2 dB(A) at the beginning of the stress test under Windows (Prime95 + FurMark). After a certain time, the fans also slowed down again (35.4 dB(A)), in order to later stabilize at around 38 dB(A).
However, the speed increases over time (32.7 -> 34.5 -> 36.1 dB(A)) under a somewhat lower load, such as the first scene of 3DMark 06. It gets noticeably louder during gaming, which we tried out with the title Fortnite.
During gaming, the fans clearly turn up and fluctuate all the time between 5,400 to a maximum of 6,800 rpm and are also clearly audible with 43 to 46 dB(A). At least the fan characteristic isn’t unnecessarily high-frequency.
The entire case of the MacBook Pro 13 stays pleasantly tempered in idle and under light load. As expected, the rear area of the base unit warms up under load because the components and heat pipes are also located here.
It doesn’t become really critical with a maximum of 43°C and even on the thighs (not necessarily directly on the skin), use in this case is possible. The power supply stays significantly cooler in comparison to the larger MacBook Pro 15.
Looking at the components, we can see that the processor temperature turns out a bit lower in the stress test in comparison to the pure CPU load. The integrated graphics card consumes part of the TDP reservoir and doesn’t develop as much heat as the processor.
At the same time, the processor mostly works under its base clock and the graphic card also only works with about 600-700 MHz (maximum is 1,050 MHz).
Apple has increased the battery capacity of the MacBook Pro 13 from 49.2 to 58 Wh. We first conducted our WiFi test with an adjusted brightness of 150 cd/m². We also compared under macOS how the result differs when the Safari browser is displayed in window or full screen mode.
But this difference is not worth mentioning with 9 minutes. Due to the higher capacity, the runtime turns out one hour better than the 2017 MBP 13 Touch Bar and is now more than 9.5 hours. This beats the Windows competition, but the MacBook Pro 13 2017 with the function keys is still well ahead of the more expensive model with more than 11 hours.
The runtime has also improved once again with the software update, as we can now determine about 40 minutes more in the WiFi test.
The video runtimes could also be increased in comparison to last year’s model, whereby software optimizations probably also play a role here. The version doesn’t quite come close to the result of the old model without the touch bar, but just under 11 hours isn’t a bad result. You can expect about 1.5 hours under load.
The battery’s charging time is pleasantly fixed with a total of 105 minutes, but there is no special quick charge function. After 30 minutes, 38% is available, after 45 minutes 55% and after 60 minutes 71%.
The release of the new MacBook Pro models hit the headlines. In particular, the performance of the new processors was in the spotlight, which was not at the expected level. A week later, a software update was released to improve performance.
However, with ~1.5 GB it was very extensive, and it rather gave the impression that the software just wasn’t ready for the launch. We know this from other manufacturers, but with the MacBooks this was not yet the case.
However, the performance of the new quad-core processor is now stable and therefore much higher than its predecessors, which were still delivered with dual-core CPUs. Nevertheless, two things stand out: The performance could be a bit higher with better cooling, but this has been adopted unchanged and must now cope with more cores.
Furthermore, the Coffee Lake processor doesn’t seem to be quite as efficient as the Kaby Lake R chips, where we have seen better results with comparable consumption values. Under full load, the cooling, which normally works pleasantly quietly, can also get quite loud and the power supply is also borderline with its 61 watts.
Besides the higher performance, the new Coffee Lake processor also offers more PCIe lanes, which is why all Thunderbolt interfaces are now connected with the full bandwidth. The WiFi module has also been updated and now supports Bluetooth 5.0, but the WiFi transfer rates are somewhat lower than the predecessor and the WiFi problem with connected USB-C devices has at least been mitigated.
The keyboard has also been updated. Apart from the actual purpose of the new silicone buffers, namely to protect against dirt particles, the stroke is a bit more muffled and quieter. Still, the limited hub is a matter of taste.
More performance and longer battery life, but otherwise a lot remains the same. The MacBook Pro 13 is also a very good Ultrabook, but the design including the cooling is slowly reaching its limits.
Apart from that, Apple continues to offer a very good Ultrabook and implements a few things that are often not available from the competition.
Besides the very high-quality case, the very good 16:10 display and the first-class clickpad, there are also smaller things like the automatically regulated keyboard illumination or the display’s TrueTone function.
Whether or not the TouchBar represents a real added value still depends on the software support and can’t be answered generally. However, it’s a pity that Apple only updates the more expensive TouchBar model and doesn’t consider the cheaper version with the function keys, which is why the Apple Macbook Pro is ranking behind vs Razer Blade 15.