HP Enterprise demonstrates next-generation computing prototype as The Machine comes together

• By Joel Hruska on November 29, 2016 at 7:30 am
• Comment

• Facebook
• Twitter
• Google Plus
• Reddit
• Hacker News

Two and a half years ago, HP (now HP Enterprise after the company split) revealed a new, revolutionary computer architecture it dubbed “The Machine.” This new computing platform would combine cutting-edge and still unproven technologies like memristors, silicon photonics, and truly massive amounts of addressable memory. HPE was forced to dial back some of its ambition when it proved too difficult to bring the entire project to market all at the same time, but it refused to give up on the idea of what it calls “Memory-Driven Computing.”

Today, HPE is announcing that it has demonstrated the major components of this new type of system, albeit in prototype form. The Machine as currently constituted consists of:

• Compute nodes accessing a shared pool of Fabric-Attached Memory
• An optimized Linux-based operating system (OS) running on a customized System on a Chip (SOC)
• Photonics/Optical communication links, including the new X1 photonics module, are online and operational
• New software programming tools designed to take advantage of abundant persistent memory.

The Machine (well, one “blade” of it, anyway). The DIMM slots are filled with NAND Flash; HPE wants to transition to lower-latency memory that competes directly with DRAM in the next few years.

HPE has previously shown off some of these components, like its X1 silicon photonics module. The X1 module is capable of transferring data at up to 1.2Tbps (150GB/s of bandwidth) over a 30-50 meter distance. HPE has also demonstrated silicon photonics technology that can move data up to 50 kilometers (30 miles) at 200Gbps. HPE’s major goal with The Machine is to create a system in which non-volatile memory (NVM)serves as a true DRAM replacement, offering at least equivalent latency with drastically reduced power consumption and low-latency optical interconnects.

Customers will still have the option to deploy The Machine as a conventional system but HPE’s plan is to offer huge pools of NVM that can be shared across many SoCs. While the diagrams below only refer to CPUs, there’s no reason this model couldn’t be extended to other types of accelerators — vector processors like Intel’s Xeon Phi or GPUs from AMD and Nvidia could at least theoretically be paired with HPE’s new architecture. The following slideshow steps through some of HPE’s design elements, and the benefits it expects to offer with The Machine compared to traditional systems. Images can be clicked to enlarge them in a new window.

HPE's The Machine is built around certain key concepts. A shared pool of non-volatile memory and high-speed optical interconnects are both vital components of the system.

<>

<>

When HPE announced that it would re-purpose The Machine’s design around conventional technology in mid-2015, it seemed to imply that the project’s groundbreaking potential had been largely buried beneath financial realities and the slow pace of technological innovation that characterizes modern semiconductor development. Today, I have to acknowledge that this dismissal was premature. HPE may not be planning to commercialize memristor technology in the near term, but The Machine is more than a conventional server with a huge amount of RAM, and the company’s work on low-latency non-volatile memory and optical interconnects could have significant implications for HPC and Big Data problems for years to come. The Machine won’t debut as a single system with all-new technologies but should transition to new memory standards as they become available. This might make it a touch less exciting on launch, but should lay the groundwork for a long-term virtuous cycle of improved performance and reduced power consumption, up to and including (eventually) exascale-class deployments.

Read More

No Man’s Sky ‘Foundation’ update is live, but don’t expect a whole new game

It would be fair to say No Man’s Sky has been one of the biggest disappointments in modern gaming history. After two years of extraordinary hype, the game launched earlier this year to mostly negative reviews. Many of the features that were promised by developer Hello Games were missing, and the final game was scarcely a match for the awe-inspiring trailers, and the company’s silence after months of frantic communication went over poorly with much of the community, to the point that hackers took over the firm’s social media accounts (briefly) to apologize for the game. Hello Games is trying to redeem itself after months of silence with the “Foundation” patch, which has just launched on PC and PS4.

I’ve been playing around with the Foundation patch (AKA v1.1) to see if it fundamentally changes the experience. The headlining change in this patch is the option to build bases. Yes, in a game of exploration, you can choose a home planet. The resources you gather from around the galaxy can be used to build new sections of the base, eventually leading to complex, labyrinthine corridors. Your bases can be used to grow resources, research new technologies, and house your alien helpers.
If setting up shop on a planet doesn’t appeal to you, there are expensive freighter ships available that are essentially flying bases. These ships can move between star systems, taking all your items, grow chambers, and alien helpers along for the ride. In addition to having more space to keep things in your base and freighter, there’s a quick-access inventory system that makes it easier to find items. This is much appreciated as the clunky inventory system at launch was one of the most tedious gameplay features.

The most serious issue with No Man’s Sky at launch was just now uniform everything was. If you’d seen one planet, you’d seen them all. The random nature of the game was so random that planets lacked large scale distinctive features like mountains, oceans, and deserts. Hello Games says that changes to the landscape algorithm allows for more aesthetically pleasing planets and biome-specific resources. I don’t know that I’m seeing much difference yet, but the graphical tweaks like motion blur and antialiasing do look nicer. One big drawback here; Hello Games had to regenerate all planets in the game. Continuing a previous game might mean you character could be in a different environment than before the update.

No Man’s Sky has also been split into three game modes with the Foundation update. There’s the original game mode, creative mode, and survival mode. Creative mode grants you unlimited health and resources to build whatever you want—it’s basically sandbox mode. Survival mode is the opposite; planets are more dangerous and resources are more rare. You will die a lot in survival mode.
Admittedly, the Foundation update brings a lot of new features, but this is still No Man’s Sky. If you didn’t like it before, you probably still won’t. There’s very little story to speak of and the animals still look fairly ridiculous much of the time. If you were on the fence, some of this might entice you to give it another shot. Players at least seem to be giving No Man’s Sky 1.1 a chance. After dropping dramatically after launch, No Man’s Sky is back in the top 50 most played games on steam with a peak of about 7,700 players on Monday.

Read More

Cold, dead fingers: Will motorists give up in-car cellphone access without a fight?

Highway deaths are up in the United States, as is the use of smartphone apps. The National Highway Traffic Safety Administration sees a link and last week issued guidelines – not mandates – that would have the major phone-OS developers figure out a way to severely limit the functionality of smartphones while a car is moving. But only the driver’s phone would be crippled, not the passengers’ phone, in NHTSA’s proposal.

Calling features would still be enabled, via handsfree link. Incoming texts would be read aloud and a speech-to-text response could be generated. Music playback would also be available. That’s about it.

Spike in fatalities unseen since the 1960s, NHTSA says

NHTSA says the 7% increase in fatalities in 2015 is unprecedented in the past half century and preliminary data on 2016 suggests the death toll rose a further 10% this year. The last time it was that high was the 1960s, when fatalities year over year increased by 7% to 9% in four of the 10 years. Since then, the death rate (total fatalities) has been flat or gone down in 26 of 46 years.

Traffic deaths each year per 100,000 population and per 100 million vehicle miles traveled, left axis, and total motor vehicle deaths per year (car, truck, bus, motorcycle, bicyclist, pedestrian). Source: NHTSA FARS database.

How big is the problem? Depends on what stats you use

The concern by safety officials comes from a spike in traffic fatalities. How big of a spike depends on which data gets publicized. The chart above is from NHTSA’s Fatality Analysis Reporting System (FARS). Here are three ways to look at the data.

• Raw fatalities. This is how many people died in a year. Since the US population grows each year, using this data set often provides the most dramatic increases. The death toll in the US in 1970 (the start of the safety era with seat belts mandatory since 1968) was 52,627 when the population was 205 million; in 2015 the US population was 321 million, 57% more, but the fatality rate was a third lower, 35,092 in 2015. So, historically the death rate is down. But in 2015 the death toll was 2,417 more than in 2014, or 7% more. The five-year increase, 2015 vs. 2010, was 6%.
•  
• Deaths per 100,000 per year. This accounts, somewhat, for our growing population over time. It’s a measure that’s understandable. If you live in a town of 100,000, on average 11 people died last year. In 1970, it was 26 per 100,000. From 2014 to 2015, deaths per 100,000 went up 7%, too, same as the raw fatalities rate. The five-year increase was 2% (2015 vs. 2010). Deaths-per-100K peaked in the 1930s at almost 30 per 100,000 people. It drops in bad economic times, during World War II (driving restrictions), and during gasoline shortages (mid-1970s, early 1980s), but it’s generally heading down.
•  
• Deaths per 100 million vehicle miles traveled (VMT). This statistic does the best job of factoring in fluctuations in the economy and fuel shortages, as well as safety improvements in cars and roads. The line slopes steadily downward, from almost 25 deaths per 100 million VMT in 1921 (the first year of more detailed auto fatalities record-keeping)  to just over 1 last year. In 2014 it was the lowest in history, 1.08 deaths per 100 million VMT. In 2015 it was 1.11, an increase of 3%, the only increase in the past decade. That’s noticeable and worrisome, but 3% isn’t 7%.

Even as auto fatalities have gone up 3% or 7% in the past year, you’re still pretty safe in a car. Multiply 11 fatalities per 100,000 people times the 80 years a person is in a car as a passenger or driver (or out walking or biking), you get 880 fatalities per 100,000, or slightly less than a 1% chance of dying in a motor vehicle accident over your lifetime. That’s low but still high enough that you’ll likely know of a friend or family member killed in an accident. It’s a leading cause, (for years the leading cause), of deaths in children and adults under 25. For better or worse, suicides, homicides and drug deaths are now challenging auto accidents among those 25 and under.

What are the odds a cellphone kill switch is adopted?

By asking cellphone-makers and OS providers to voluntarily restrict what services the phone provides in a moving car, the feds look a bit less like spoilsports. There is likely to be pushback. Whenever drivers are asked to self-rate their driving, the majority say they’re above average, which isn’t possible statistically, and they might bristle at seeing their quote freedoms curtailed.  (See below.)

There are reports that Apple in 2008 applied for a patent on technology that would lock out phones that were in motion; the technology supposedly could determine which phone was the driver’s. The patent was cited in a 2014 lawsuit over an alleged distracted-driving fatality.

CellControl, a Baton Rouge, LA, company, offers DriveID, a dash-mounted $129 hardware-software product that can limit the driver’s use of his or her phone. DriveID is managed by parents in a family situation or by a fleet manager in a commercial setting. Pokemon Go has been cited as a factor in some car crashes, so much so that the current version has a lockout that disables the app when it’s being played at greater than walking speeds. The change was hastened by a video showing a Pokemon Go motorist sideswiping a police car.

It’s possible phone-makers will use a limit-your-phone-access mandate as a reason to keep from adding competing apps into the handful now enabled under Apple CarPlay and Android Auto. Specifically: You wouldn’t see Apple Maps on an Android phone and you wouldn’t see Google Maps or Waze on an Apple phone running CarPlay.

Americans didn’t like enforced ignition interlocks

History shows motorists reacted badly to heavy-handed behavior modification. In the mid-1970s, cars were equipped with ignition interlocks that wouldn’t allow the car to be started until the driver and front seat passenger had buckled their seat belts. This was supported by the auto industry in part to stave off mandatory airbags, which at the time were seen as a substitute for seat belts. The buzzers were loud, sometimes electrical gremlins kept the cars from starting, and on many cars it was easy to defeat the interlock. Automakers pulled back within a couple years.

Circa 1980 some automakers, again trying to keep airbags out of the picture, installed automatic seat belts with complex (read: not always reliable) mechanical arms that draped the seatbelt across the driver’s and passenger’s torsos. If driver or passenger opened the door while backing up, the belt retracted and wrapped itself around the occupant’s neck. This, too, died an early death.

Is there a sensible compromise?

Given how little Americans like to be told what to do when driving, they’re not going to like limits on their freedom to use their phones in ways ranging from distracting or potentially hazardous to downright stupid.

Driver-assist technology can help make the occupants of a car safer should the driver be distracted while creating or reading a long text. A car equipped with lane departure warning, adaptive cruise control, pedestrian detection, and forward emergency braking will alert the driver if the car ahead suddenly slows, or if the car is drifting out of lane. For someone who’s texting, that is either a safety feature … or it’s an enabler for someone to write longer texts before they get in trouble.

While texting is the main concern, better voice recognition and one-shot destination entry will make it easier to enter an address without stopping. Right now, every phone has that feature, but only some embedded navigation systems. Already, most automakers lock out the LCD display’s ability to tap in the address. Automakers so far have resisted implementing technologies that could recognize when a passenger, not the driver, is entering address information. Mercedes-Benz offers a unique technology, SplitView, that shows the driver and passenger different views of the same center stack screen. Alternating pixels point left and right.

If NHTSA moves to reign in the use of smartphones in cars, there’s likely to be pushback. Even if distracted driving is the cause of the current 3%-7% increase in fatalities, drivers still see the overall risk to themselves as minor. And there’s conflicting research as to what constitutes distraction. Studies show that just talking on a hands-free phone takes up some of the driver’s attention. Tuning the car radio has always been distracting. Some research says talking to a passenger can be distracting. In the past, two of the biggest distractions were the driver dropping a lit cigarette in his or her lap, or a bee flying in the car. Those are small distractions now with the decrease in smoking and air conditioning that allows for windows-closed driving.

Read More

Homeopathic solutions now have to be labeled to disclose that there’s no science behind them

The FTC is playing whack-a-mole with pseudoscience again, and this time it’s targeting homeopathy. Their latest comments contend (PDF) that the standard disclaimer isn’t enough to dissuade consumers from buying this crap, so now not only do homeopathic products have to carry the standard disclaimer, they also have to say there’s no science behind them.

I support the idea that it’s your body, so do with it what you want, and if taking homeopathic sugar pills makes you feel better, sure, keep taking them. But the only reason I can’t make a blanket statement that homeopathy provides no benefit whatsoever is the placebo effect. To paraphrase Tim Minchin’s memorable, NSFW rant on alternative medicine, Storm, without fail or exception every kind of alternative medicine has either not been proved to work, or been proved not to work. Do you know what alternative medicine is called that’s been proven to work? Medicine. Homeopathy must not be confused for medicine.

The FTC agrees, and they’re really emphatic about it. Homeopathic products don’t work: because further dilution is supposed to increase a homeopathic remedy’s strength, many homeopathic products contain no detectable traces of their title ingredients. This hilarity is a good thing for consumers, because homeopathy’s other fundamental principle of “like cures like” means the solutions could be deadly if they actually contained any of their title ingredients: among a great many other products, homeopaths prescribe nux vomica (strychnine), arsenica album (arsenic) and even teething products containing extract of belladonna, which it turns out is not great for kids. In reality, homeopathic products tend to be recalled when it’s discovered that they contain their title ingredient. Three separate times in their recent press release, the FTC maintains that the standard “has not been proven to treat or cure any condition” disclaimer is unlikely to be enough to adequately convey how completely valueless homeopathic solutions are. Homeopathy has been rejected by science, medicine and now by the FTC.

Debunking homeopathy has already been done so beautifully that nobody needs me to do it here. If you haven’t yet had a headache today, go search for “succussion” and that’ll fix that right up for ya. I present to you Munroe’s razor:

 

Image: Randall Munroe, xkcd.com/808/

This is not government overreach. This is the government intervening to put the kibosh on outright fraud. Products which definitionally cannot do what they claim to do and contain none of what they claim to contain are fraudulent. Nobody benefits from their sale but the fraudster. This is the FTC attempting to intervene in the one way it can, because the next step up would be banning homeopathic products outright and we don’t have a reason to go there yet. Consumers can buy what they want, even if it’s terribly advised. The principle of caveat emptor puts the responsibility on the buyer to not go carting off pallets of triple-distilled homeopathic 40C oleus serpentius.

Read More

New teardown reveals Surface Studio packs ARM CPU core, upgradable storage

Microsoft’s new Surface Studio is an interesting foray into high-end premium hardware, but it’s clearly not intended as a user-expandable system. It’s always interesting to see how companies balance slim, cutting-edge hardware against basic repairability, and Microsoft’s track record in this area is pretty mixed. iFixit’s recent teardown of the Surface Studio found some expected downsides, but a few intriguing details and user-friendly options as well.

It’s not too difficult to open the bottom of the system, but there are some wire leads to watch and you’ll have to remove the midframe to access the guts of the machine. The cooling solution appears quite robust — Microsoft is using a multi-heatpipe system with dual fans, one for the CPU and one for the GPU. The larger fans are used for the GPU, while the smaller serves the CPU (the GPU also exhausts directly, while the CPU fan pushes air into a plastic channel).

The dual heatpipe. Image by iFixit

Microsoft is using a hybrid cache solution, but earlier reports that implied the system used a Seagate SSHD were apparently invalid. There are two general ways to create a hybrid SSD platform: You can use a small amount of SSD storage to accelerate both reads and writes (we’ve covered this type of solution in previous reviews), or you can deploy an SSHD. SSHD’s contain both NAND flash and conventional magnetic media in the same 2.5-inch enclosure, and they typically have different caching strategies and less total NAND than hybrid solutions that pair separate NAND and magnetic media storage pools. iFixit’s Surface Studio has 64GB of storage total and could theoretically be expanded to 128GB if you have an appropriate workbench and the necessary expertise.

Unfortunately, anyone who dreamt of upgrading the onboard RAM or other components will be disappointed — all of the onboard DRAM is soldered to the motherboard, as is the CPU. There’s no upgrading the GPU, either; the GTX 965M / 980M isn’t mounted to an MXM-compatible PCB. It might one day be possible to perform a full mainboard swap by buying a Studio on eBay and tearing it down for parts, but that’s obviously not a cost-effective solution for most people and won’t be a realistic option for several years, if at all.

The Surface Studio’s RAM, GPU, and CPU are all mounted on the same piece of silicon. Image by iFixit.

The display can be removed and replaced, but only by cutting through a glue layer and removing the rear hinge. iFixit rates this as a fairly straightforward repair if all you need to do is replace a cracked glass screen. The display includes a second motherboard with some Microsoft-branded chips, some NOR flash, and a 32-bit ARM Cortex-M7 processor. We don’t talk much about ARM’s embedded controllers, but the M7 is supposed to be a significant improvement over ARM’s previous embedded chips, as shown below:


The Cortex-M7 is used as part of Microsoft’s PixelSense display, and it’s not the first time we’ve seen ARM controllers popping up in PC hardware. For years, conventional wisdom predicted that x86 and ARM were headed for an inevitable confrontation across the smartphone, tablet, and even the laptop market. That seems much less likely to happen now, but we’re seeing some interesting trends from companies that integrate ARM and x86 silicon side-by-side, whether that’s as an on-chip security solution (AMD’s TrustZone) or using an embedded ARM controller to handle specific tasks, as Microsoft has done.

Overall, the Surface Studio wins a 5/10 score for its repairability. The 8GB limit on the lower-end Surface Studio (to the extent that any $3000+ system can be called “lower end”) could give some buyers pause, given that there’s no way to upgrade the platform. Conventional desktop users can easily get by with this, but the content creation professionals that Microsoft is targeting really could’ve used a 16GB baseline configuration.

Read More