A quick news post for anyone following the HTC Vive announcements: as we reported back at MWC, the official pricing and timing for pre-orders was announced.  For your $799 on February 29th at 10am EST, you get the following:
  • HTC Vive head mounted display (HMD)
  • Two Base Stations (for location)
  • One pair of controllers
  • A Link Box
  • Vive Ear buds
As part of the package, two titles were also announced as being bundled: Job Simulator (the demo of which I still remember vividly as trying managed to microwave steak and wine) and Fantastic Contraption. As of yesterday, HTC announced that Google’s Tilt Brush will also be a part of the pre-order bundle for a limited time. Tilt Brush is the VR equivalent of MS Paint, that initial drawing platform allowing users to create, but this time in a full 3D environment.
The pre-order time is set to go green in about an hour:
  • 10am ET
  • 7am PT
  • 3pm GMT (UK)
  • 4pm CET
The listed $799 price last week was the US pricing, without tax. HTC has announced regional pricing as well:
RegionMSRP Pricing
* indicates without tax pricing
USAUSD $799*
CanadaCAD $1149*
AustraliaUSD $899
ChinaCNY 6888
TaiwanNTD $28,288
JapanJPY ¥111,999
UKGBP £689
New ZealandUSD $949
EurozoneEUR €899
Credit card purchases will use a hold on the funds, whereas Paypal transactions will be processed immediately. Shipping beings April 5th, indicating that pre-orders will be first come, first serve, limited to one per customer. The link to preorder will be www.htcvive.com – we have not heard if there is limited preorder units, or if it will be ‘as we build them, they will be shipped’.

Last time I wrote about smartphones beyond the odd news post and live blogs at events, I had purchased a $160 smartphone for my own use that ran four days on a single battery charge. Not too long after, Huawei invited me over to their UK office to show me the Mate S, and their new Force Touch edition. Andrei’s review of the Mate S had been uploaded only the previous day, so part of the discussion was around some of the findings in that review, but the purpose of bringing me into the London office was to actually try the new features. I was taken through the motions of Force Touch on Huawei’s hardware. I wasn’t expecting it at the time, but at the end of the demonstration they handed me the unit and said ‘try it for a while, tell us what you think’.
So I’ll admit from the start – it isn’t every day that I use a $650 smartphone. This is Huawei’s top model of the Mate S, featuring 128 GB of storage, Force Touch, and in rose gold, which sets it apart from other Mate S implementations. On the color, which is typically not one I would have considered, I had more positive than negative comments on the rose gold in the past couple of months, with most questions about the color being directed to the comparison to Apple’s equivalent. A number of people also suggested that the Mate S was copying the iPhone, which was not an uncommon sentiment.
In these ‘a month with’ pieces, I use the smartphone as a daily driver, trying to see how easy the device fits into my slightly uncommon use case. I also try to use the new features on offer when they become relevant to what I do. If you want the full review of the Mate S, please read Andrei’s very detailed breakdown of the Mate S and the Kirin 935 SoC – this piece is my expression my personal opinion as a technical user but not a smartphone reviewer.

This review is going to be a list of pros and cons that I’ve come across in the last couple of months that are purely related to me. Your mileage may vary. When I gave Huawei this feedback, they agreed with most of what I had to say. I also took a small video explaining some of the interesting features that were difficult to explain in text:

Pro: Fingerprint Sensor Speed

When I road-tested the HTC One max, I struggled using the fingerprint sensor they had on it. It was an early generation swipe sensor on the rear of the phone, and no matter what I did it would only unlock 30% of the time or even less. With the swipe sensor, it only worked within a small range of angles, and invariably with a large phone you do not always hold it at the same angle to unlock. Having never used any other fingerprint sensor other than that at the time, the sensor on the Mate S was superb. I logged five different fingers in the system (first two on each hand and a third finger on my right), and it beats my previous experience in two ways. Firstly, it’s a touch sensor, which means no swiping. Secondly, the response time is super quick, so if it fails to unlock another attempt is less than a third of a second away. It’s not only quick by comparison to the One max, but I’ve also played with the 6P a little bit and the Mate S easily beats the 6P in time to calculate the unlock and then retry. Because I’ve gotten so used to the Mate S, I end up removing my finger from the sensor too early on the 6P, and it asks to try again. It’s going to be a tough habit to break out of.

Con: EMUI, Software Glitches and Response to Voice

Here’s a triple whammy of annoyance. EMUI, for all its best wishes, attempts to be very iOS like. Despite being on an Android base, there is no true list of apps or app drawer, limiting each app icon to only one screen. My typical Android experience up to this point involves some app on multiple screens, just to save time, but that isn’t possible here. The background seems only to be changeable to a set of Huawei defaults, as even after two months I haven’t found a way to put my own picture there (all the normal Android ways don’t work). EMUI also interferes with a number of apps I like to use, such as a network speed monitor so I can tell when I’m on a limited connection if some app in the background is hogging all the data. My app of choice, while it starts up, seems to be ejected from memory automatically very quickly when the screen is turned off.
The voice glitching is another aspect to be wary, and this comes in two forms. Firstly, at the end of my last smartphone test, my use of an Android phone had evolved to the point where I was starting to use Google Now to do simple tasks like set reminders and alarms, or ask basic questions very easily. On the Cubot H1, a smartphone with four A53 ARM cores running at 1.0 GHz, Google Now responded immediately without hesitation. However, the Mate S has issues. Sometimes saying ‘Google Now’ doesn’t activate the feature, and when it does, the system ends up trying to load something for three seconds and you get the ‘I didn’t hear that’ because the processor was busy. That means every time you ask it something, it’s an additional delay that I could have spent typing the question into Google for.
The other voice glitch is completely random and happens about twice a month. I will have the phone on the desk or in my pocket when talking to friends when out of the blue it will start talking about how it can’t understand a command, reach a contact, or let you know where it is. It has even done this while sitting on my bed side table while speaking with my wife. There is an always on voice feature to help you find where the phone is (which never seems to activate when you are actually looking for the phone), but the fact it activates randomly and usually when it is at high volume makes the user look like a random idiot. It’s not a good look.

Pro: Knuckles

One of Huawei’s interesting features on the Mate S is the ability to do certain things with your knuckles. It sounds odd, but essentially this means striking the screen with one or two knuckles and drawing out a box or a letter to perform certain functions. Writing a C for example, no matter where you are in Android (or even when locked), will force the camera application to load. This makes it easy to take pictures rather than searching for the camera icon (which you can’t put on multiple screens due to EMUI). The software recognizes C, M and two other letters, all of which can be customized to run the application you want. Please watch the video above for a demonstration.
The other knuckle-based features aren’t that useful so much, but are still there for the odd time you show them off to friends. A double tap will take a screenshot, while forming a box on any screen will cause what is in the box to be cropped into an image, which can then have filters or shapes applied. A double knuckle double tap will let the smartphone record both audio and on-screen activity, so if you want to show how to do something to someone else, you can make a direct video recording and send it to them.

Con: Gmail App on EMUI, Split Windows

I’ll round the EMUI issues off with these two. I use the Gmail app like a number of Android users, however the EMUI implementation is extremely annoying, especially with the backgrounds in play. When you have email and a number of new messages, then scroll down the notifications bar, EMUI displays the name of the sender and the email topic in black. This is black on a gray notifications bar which is semi-transparent, but most of the Huawei backgrounds for EMUI are quite dark, meaning you can’t read anything. This just adds another couple of seconds to looking at email and completely negates a quick notification look at the subject to see if it’s something you should respond to.
 
Huawei does implement a variant of multitasking in EMUI called Split Windows, and is extremely similar to other multitasking arrangements in other Android and iOS devices. It’s a good feature that I can see myself using a lot of, so why list it as a con? Well I’m not sure whether it is new, or it was a last decision thing, but the split windows require each app to be able to dynamically handle the adjustment of two screens side by side – the split windows feature can’t manage it on the fly. As a result, there are only eight apps that use the feature, and they are all from Huawei, and I don’t tend to use any of them. My best use for this, because of the limited application limitation, is typically when I want to listen to music on YouTube while searching in another browser window. But both browser windows have to be using Huawei’s browser, and not the one I normally use with all my bookmarks and cookies.

Windows splt between browser and notepad

Meh: Force Touch

So Huawei technically came out with a Force/3D Touch implementation before Apple, however this implementation seems limited in scope. The two best things you can do with it is zoom in on pictures in the gallery using one finger (it comes up like a magnifying glass with a zoomed view) and weigh round fruit between 100 and 400 g (0.15-0.90 lbs). There are other non-obvious uses for Force Touch, but these are the only two I’ve used since having the device, and even then I don’t use the latter except to show the feature off. The picture zooming thing gets a bit more use, but even then it’s fairly minimal. In order to make Force Touch more of a thing, it has to be obvious that it does more in more apps.
  

Pro: Screen Size

Ever since I used the HTC One Max as a daily driver, I am sold in large phones. I get why they exist, and I somehow adapted to them rather quickly. The Mate S is 5.7-inches, below the 6-inch of the One max but above the 5.5-inch of the Cubot H1. The screen has a reasonable brightness range, although doesn’t go as low as some other devices like the Nexus 6P. What I would say though is you won’t be using the Mate S in high brightness mode that often. This brings me to an issue I’ve had with the Mate S: battery life.

Con: Battery Life

We covered this in the original Mate S review, but due to an issue in the SoC silicon, the memory controller consumes a lot of power when at high frequency, even when idle. For most smartphone users that means nothing, but it relies on the OS to de-clock the memory when speed is not required. I’m not sure how often this happens, but I can easily burn through 50-60% of the battery in three hours, watching video/flipping between Skype and Twitter even at low brightness while at an airport. True story.
Putting it in Flight Mode and not using it during the 10-hour flight by comparison and the device was barely sipping at the power. But it does mean the Mate S is certainly a device that requires a daily charge, sometimes before the end of the day. This was a very quick slam to the face coming from a four-day charge smartphone in my last road test.

Pro: Software Eviction on Battery

Because I’m not so liberal with spending money on smartphones, I’ve often ended up with low power, low memory devices, or I worry about what background services are eating power. I get the typical case of memory and battery anxiety, wondering if I’m still going at full speed when I need it or if the device is going to last the full day. This is especially true when using the device at an event. Luckily there are two things on the Mate S that help here: an optimization widget and a software power detector. The first one is software which I’m not sure is part of Android, but does an optimization of currently running apps. (I know that removing apps from memory means it costs power when you need them next, but somehow memory anxiety still exists, despite having 2GB to play with.) The second is the eviction on battery use – if the OS detects that a certain piece of software is causing the SoC to draw more current, it will tell you and offer to close the software. I have seen people complain that on the Mate S this feature is fairly aggressive, and it does come across that way, but at least it gives me the option to do so.

Con: Camera

As mentioned in Andrei’s review, the camera on the Mate S isn’t one of the plus points in the device. It certainly takes a picture, but I’m an unsteady photographer who wants the photo taken as quickly as possible, and sometimes that doesn’t happen. Don’t get me wrong, I’ve used the Mate S to take pictures for AnandTech Live Blogs, but when an image gets compressed to the size of ravioli it doesn’t matter too much anyway. But on the important stuff, the family photos, there are smartphones with better quality implementations at this price point. As with my other road tests, here are some example pictures.

Pro: Camera and Document Adjustment

The camera has one nice feature worth mentioning, especially for anyone having to take notes on lectures/meetings or someone who photographs a lot of documents. The Document Readjustment option in the camera attempts to detect the edges of a piece of paper, a monitor or a projection, and will tilt, stretch, rotate, expand and adjust the image as necessary to give a flat representation. This means that you can take the image at a wide angle (I’m not sure the limit, but it certainly works up to 45 degrees) and it will come out, around 80% of the time, like the photograph was taken head on. I have used this on a number of presentations over the last couple of months, and in the review of a server motherboard recently I used it when the BIOS did not have a print screen option. In this case, I’d normally set up my DSLR on a tripod and take pictures that way, but with the Mate S I was able to simply raise the camera each time on a different screen, wait a second, and then take the image. So rather than having to crop then color balance, all I had to do for 95% of the photographs is a bulk color balance. I would note however that it does seem to strip the EXIF data from the image.

Con: Call Quality

I don't have any tools that will directly measure call quality, but I found the Mate S to be difficult to use in any environment with mild background noise. When in a quiet environment though, a call was clear to hear and easy to speak to. I'm not sure if it was the position of the microphone on the bottom, but more often than not I would have to fish out my headphones and plug them in, so I could arrange the phone and speak directly into the bottom of it. 
 

Con: Too Many Cons

I’ve been using the Huawei Mate S 128 GB with Force Touch in rose gold for over two months now. It has many good features, and some of them feel top notch in my limited testing sphere, such as the fingerprint sensor and the ability to knuckle into the camera at any time. But it has too many issues to make me fully content with the device. Most are derived from software, most likely the nature of the latest version of EMUI supported on the Mate S, but things like not being able to read email notifications at a glance or the random nature of how it likes to talk to you in the most random situations mean that I’d probably end up buying something without EMUI or flashing the device (assuming I could get all the feature I like some other way).  I feel if that it was running a base version of Android, it might have given a better user experience. All that being said, the one waking day battery life (or under one day of battery on active days) is something not easy to overcome, and means carrying around a battery pack. At least anecdotally it felt like it charged rather quickly.
Along with the Mate S, part of the way through I decided to also try out the Huawei TalkBand B2 smartwatch for some time. The TalkBand B2 was a one of the cheap press event freebies, but it tracks a few smartwatch type features in a non-Android way at a low price and comes with a monochrome display. For someone who has resisted the smartwatch market up until this point, as well as being apprehensive about the utility of such devices, I decided to bite the bullet and at least have a go. I’ll write up my experiences is a separate piece and upload it shortly.
G.Skill has introduced new additions to its Trident Z family of DDR4 memory modules, which are designed to simplify the lives of anyone who wants to color-coordinate their PC. The new Trident Z lineup includes memory sticks with five new color schemes which are designed to match the aesthetics of overclocking and gaming motherboards.
Just 10 to 15 years ago, PC modding used to be reserved for hardcore enthusiasts, who were willing to spend time and money to build beautifully looking PCs using custom-made components. PC cases with built-in LEDs or even transparent windows were in the minority, with only fully custom liquid cooling systems as a deviation, and the vast majority of motherboards were either slowly implementing color schemes or remained green/brown. Memory modules with heat spreaders were considered stylish. Fast forward to 2016, we have plenty of mass-produced components with PC modding features and almost everyone with moderate knowledge of PC hardware can build a PC with matching color scheme as well as lighting. The new Trident Z memory modules further simplify building of computers with unique designs, something that gamers and enthusiasts want to do.
The G.Skill Trident Z series comes with two vectors: the main body and the top bar. The main body is either black or silver, and the top bar can come in orange, yellow, white or black. The original Trident Z modules featuring a silver body and black brushed aluminum heat spreader with a red top-bar highlight will remain on the market. Meanwhile, new color schemes (such as those with orange and yellow bars) will come in handy for those building new PCs based on the latest ASUS ROG, GIGABYTE Super Overclock or MSI XPower/MPower motherboards.
The new G.Skill Trident Z memory modules have 8 GB or 16 GB capacities and are based on Samsung’s 8 Gb DDR4 ICs. The Trident Z will be available in dual-channel and quad-channel kits with 16, 32, 64 and 128 GB capacities, targeting everything from gaming desktops to higher-end workstations. Initially, G.Skill will offer colorful Trident Z kits with DDR4-3200, DDR4-3300, DDR4-3333, DDR4-3400 and DDR4-3466 speed-bins, CL14, CL16 or CL16 latencies as well as the higher DDR4 standard voltage (which is normal for high speed DDR4). The Trident Z Color are aimed at a good price/performance ratio for the majority of PC enthusiasts. Offering too many DDR4-4000+ SKUs with different color schemes could make lives of retailers uneasy since such modules are not too popular because of their high price. Therefore, if you want to go DDR4-4000 and higher, you will have to stick to “classic” Trident Z color scheme: black and silver heat spreader with red top-bar highlight.
I know Ian has taken a delivery of some of these modules for future testing, in the silver body and white bar design. As shown below, these are 16GB modules at DDR4-3200 and 14-14-14-34 sub-timings, installed in the GIGABYTE X170-Extreme ECC.
G.Skill intends to make the new Trident Z modules available this month. Since such modules only feature new heat spreaders, but continue to use the company’s own design PCBs as well as Samsung’s mass-produced 8 Gb DDR4 chips, they should not cost significantly more than existing Trident Z solutions
For a few years now, NVIDIA has been flirting with the server business as a means of driving the growth of datacenter sales of their products. A combination of proof-of-concept hardware configurations and going into spaces not necessarily served right away by the OEMs, NVIDIA has over the years put together boxes like their Quadro Visual Computing Appliance and the DIGITS devbox. Though a side business for NVIDIA, it’s one that has taken on some importance, and nowhere is this more applicable than with the Pascal-based Tesla P100 announced this week.
One of the more interesting aspects we’ve known about P100 for some time now – and well before the product was formally announced this week – was that NVIDIA would be pursuing a new form factor and new interconnect technology. Though the GP100 GPU at the heart of the P100 supports traditional PCI Express, NVIDIA has also invested heavily in NVLink, their higher-speed interconnect to enable fast memory access between GPUs, and unified memory between the GPU and CPU. For NVIDIA this is a logical progression: as GPUs become increasingly CPU-like in their flexibility and necessity, they move from being a peripheral to a core processor, and so too must their I/O and form factor evolve to match.
This brings us back to servers. Because NVIDIA is launching P100 with a new form factor and connector, P100 requires completely new infrastructure to run. NVIDIA’s OEM partners will in successive quarters be offering their own systems supporting P100 and its mezzanine connector, but right now at day 1 there is nothing from the OEMs ready to support P100. So in order to avoid the long gap for OEM servers to catch up, to serve as a pathfinder for their own development of P100 and its software, and to further their own goals for the server market, NVIDIA is producing their own P100 server – the first P100 server – the full scale DGX-1.
I call DGX-1 a full-scale server because it is a maximum implementation of Tesla P100 right off the bat. 8 P100s are installed in a hybrid mesh cube configuration, making full use of the NVLink interconnect to offer a tremendous amount of memory bandwidth between the GPUs. Each NVLink offers a bidirectional 20GB/sec up 20GB/sec down, with 4 links per GP100 GPU, for an aggregate bandwidth of 80GB/sec up and another 80GB/sec down.
The mesh itself is a Non-Uniform Memory Access (NUMA) style design; with 4 links, not every GPU can be linked to every other GPU. But in the case where two GPUs aren’t linked, the next GPU is no more than 1 hop away. Meanwhile, the GPUs connect back to their x86 CPU host over standard PCI Express.
NVLink in turn enables each GPU to directly execute read, write, and atomic memory operations out of the memory of the other GPUs, functionality that has not been available to NVIDIA GPUs before. For this reason, NVIDIA has not shied away from pointing out that while NVLink may technically be an interconnect, it’s really more of a memory technology, which would make data transference one such application of the technology. Fundamentally this ends up being a very different take on an interconnect than PCI Express, which is a data bus with certain memory-like features layered on top of it. In this case it’s probably safe to say that NVLink is closer to the likes of Intel’s QPI and AMD’s HyperTransport.
Getting back to the server itself, backing up the P100s is a hefty dual Intel Xeon based system. The DGX-1’s 3U chassis holds a dual 16-core Xeon E5-2698 v3 arrangement, 512 GB of DDR4-2133 LRDIMMs, four Samsung PM863 1.92 TB storage drives, dual 10 gigabit Ethernet (10GBase-T) as well as four EDR Infiniband connections. This system serves not only to feed the Teslas, but to further drive home NVIDIA’s scalability goals as well, with the Infiniband connections in particular put in place to allow for high-performance DGX-1 clusters. Of course with so much hardware on hand you’ll need a lot of power to drive it as well – 3200W, to be precise – as the 8 P100s alone can draw up to 2400W.
NVIDIA DGX-1 Specifications
CPUs2x Intel Xeon E5-2698 v3 (16 core, Haswell-EP)
GPUs8x NVIDIA Tesla P100 (3584 CUDA Cores)
System Memory512GB DDR4-2133 (LRDIMM)
GPU Memory128GB HBM2 (8x 16GB)
Storage4x Samsung PM863 1.92TB SSDs
Networking4x Infiniband EDR
2x 10GigE
Power3200W
Size3U Rackmount
GPU ThroughputFP16: 170 TFLOPs
FP32: 85 TFLOPs
FP64: 42.5 TFLOPs
In terms of construction, as hinted by in NVIDIA’s sole diagram of a component-level view of the DGX-1, the Tesla cards sit on their own carrier board, with the Xeon CPUs, DRAM, and most other parts occupying their own board. The carrier board in turn serves two functions: it allows for a dedicated board for routing the NVLink connections – each P100 requires 800 pins, 400 for PCIe + power, and another 400 for the NVLinks, adding up to nearly 1600 board traces for NVLinks alone – and it also allows easier integration into OEM designs, as OEMs need only supply an interface for the carrier board. It’s a bit of an amusing juxtaposition then, as the carrier board essentially makes for one massive 8 GPU compute card, being fed with PCIe lanes from the CPU and power from the PSU.
On a quick aside, I also asked NVIDIA about cooling for the P100s, given that there’s 2400W of them to cool. While they aren’t going into great detail about the heatsinks used (e.g. whether there are any vapor chambers involved), they did confirm that they are 2U tall heatsinks, and as NVIDIA is wont to do, the tops are painted green. Apparently these are the standard heatsinks for P100, and when OEMs start assembling their own systems, these will be the heatsinks that come with the boards.
Meanwhile, besides the practical necessity of constructing the DGX-1 as the pathfinder system and reference implementation for the P100, at the end of the day the DGX-1 is meant to be a workhorse for NVIDIA’s customers. Given its 8 P100s we’re looking at 28,672 CUDA cores and 128GB of shared VRAM, the DGX-1 is rated to be able to hit 170 FP16 TFLOPs of performance (or 85 FP32 TFLOPs) inside of 3Us. And though the server should be good at just about any HPC task given the versatility offered by the P100 – it offers fast FP64 and ECC as well – NVIDIA is initially pitching the box at the deep learning/neural network market, where they’re intending to fully exploit the FP16 performance improvements of the Pascal architecture. NVIDIA has of course been building towards this market for some time now, and it is their hope that DGX-1 combined with their ever-increasing collection of tools and SDKs (cuDNN, DIGITS, etc) that this will serve as a catalyst for that market.