Jem Davies is ARM Fellow, VP of technology in the Media Processing Division of ARM, he speaks at the AMD Fusion Developer Summit conference. He talks about Power, Heterogeneous computing, Moore’s Law in the next decade, open standards for massive parallel GPU computing and more.
What do you think about ARM speaking at the AMD conference? First step before AMD announcing that they’ll make an ARM processor? Why wouldn’t AMD also make an ARM Processor? Post your opinions about this keynote in the comments. Jem Davies posted this blog post on blogs.arm.com a few days ago in preparation for his keynote at the AMD conference.
Check out this awesome video interview with ARM President Tudor Brown conducted by Rob Symes of Campbell Black, hear Tudor Brown talk about how ARM was founded 20 years ago, a sometimes hilarious interview about ARM’s Cambridge British/European technology company attitude:
We assumed it wouldn’t work. That was the nature of our British Cambridge attitude. Here we go again, this would never work. But never mind, it’s fun, let’s make a go of it. And that was kind of the attitude that had been bred into us at Acorn. Many of us had been working at Acorn for many years, we created some great technology, but commercialy it had been a flop. And indeed the first products that were based on ARM, when ARM set out as a separate company, were all flops commercialy. It was a set of sales flops, I can list them:
– the Acorn computers which were really not very successful,
– the Apple Newton which was the first PDA which was the whole raison d’etre for the company that was clearly a flop,
– the 3DO multiplayer first CD based games machine in the world, based on ARM, commercial flop,
– the Sega Dreamcast, games machine, flop
– and so it went on.. we were kind of used to that.
It was a very important point that we were running the company as a bunch of pessimists, very lean and mean, on the basis that we weren’t going to build the company up on the hope of some future growth, we were going to run the company on the basis that what we had today had to last.
Both of the new amazingly priced Nook Touch ($139) and Kobo Touch ($129) seem to use the new Freescale i.MX508 (I first filmed i.MX508 a year ago) to facilitate smooth touch interfaces using Infrared Touch technology (first used in the Sony PRS-650) but also to easily enable those to be the first Android powered e-ink e-readers! Basically i.MX508 and Android makes first real useful use of WiFi and Touch on e-ink e-readers! Using Android on e-ink is awesome to provide for great files support, web browser, easily add apps like news readers, rss readers, Chrome-To-Phone functionality (just beam any article or any text from your Chrome browser on your laptop to your e-ink e-reader in one click!), synchronizing of bookmarks, wireless beaming of web based ebooks, all in one click. Android on e-ink just increases the e-readers potential greatly. Using infrared touchscreen technology is a great way to provide the best possible readability, not adding any layers on top of the screen.
Look forward to much more fascinating devices to come with: i.MX508 + Android + infrared touch + WiFi/3G + e-ink = Low-cost e-ink e-readers!
Also see my video of the alternative Acoustic Pulse Recognition (APR) touch technology.
An interesting thing to follow though will be to see how open Android i.MX508 e-ink e-reader makers will be towards allowing for alternative e-book stores, and allowing for full open Android apps installation and full unrestricted access to the web browser. As it may be that part of the profit margins on these new low-cost e-ink e-readers may be subsidized by the hardware makers expecting to make later huge profits on selling e-books, like Amazon does with their Kindle e-book store, it will be fun to see if all the e-ink e-reader makers will seize the opportunity to be open with Android on e-ink and at the same time find new interesting ways to monetize those. I for example would prefer paying less than $10 per month for unlimited legal access to all e-books, all newspapers, all blogs, all websites, everything, perhaps even certain 3G bandwidth included, but will content owners and hardware makers be open minded enough to let that happen quickly?
When will Google officially announce open source Android optimizations for e-ink devices and with deep integration of Google Books, Google Reader, Google Fast Flip and more?
Here’s how I think those e-ink e-readers could be sold in the best way: $99, $10/month optional pre-paid plan gives 100mb 3G data (upgradable bandwidth for $1/100mb or less) and full unlimited access to all e-books, all e-newspapers, e-magazines, blogs through RSS/search and more. Full unrestricted web browsing on e-ink. And nice complete cloud syncing with your Smartphone, Tablet and Laptop. While e-ink e-reader market is doing great perhaps reaching 20 million units to be sold this year, if they price them right and open them fully, I think they’ll do even better.
Archos hacker Bubu just released the multi-boot menu support for the Archos Gen8 Tablets, it reportedly works great on Archos 70 Internet Tablet and Archos 101 Internet Tablet, find the release here: http://www.openaos.org/archives/739
Then you can install alternative firmwares and have them right there available in your multi-boot menu on your Archos, so you could for example today choose between Archos official Android firmware, Debian Linux and Angstrom Linux.
If more software hackers contribute, Archos tablets may also run Ubuntu 11.4, Chromium OS, Gingerbread and any other embedded Linux OS through this multi-boot system.
Bubu and the team at OpenAOS.org previously released this solution for the Archos Gen6 and Gen7 tablets as well.
Check out how ShiZhu’s new Pixel Qi tablet series perform outdoors and indoors against other regular LCD based new Android tablets such as the HTC Flyer (with an anti-glare coating) and the Archos 70 Internet Tablet, you can see how the Pixel Qi screen is much more readable outdoors. This tablet is based on the Samsung Hummingbird Cortex-A8 processor. They are releasing also a 7″ and they can provide them in several different designs.
A quite small modern solar panel can produce 1W of power, enough to power both the Pixel Qi and the whole ARM Powered tablet motherboard behind it. Think for a second how amazing this is. They can put solar panels on the bezel and the Tablet could basically be fully solar powered, have a bigger solar panel on the back of the tablet if you want to just charge it and not use it. This compact 1W Solar panel is $3. OLPC could be using this for the upcoming OLPC XO-1.75 Laptop and the XO-3 tablet.
While this is not yet with the anti-glare anti-reflective coating and the reflective mode for touch is not yet implemented in this ZTE sample, this is how the ZTE Light with Pixel Qi might look like, the fully optimized version should be shipping in the third quarter of this year.
Ian Drew is the Executive VP of Marketing at ARM, here is my recording of most of his keynote at Computex 2011 in Taipei today. Sorry about the missing first 3 minutes, sorry for the slight shaking in the first few minutes (as I was trying to get to a better chair) and sorry that I don’t have direct sound recording from the microphone. I couldn’t see if Computex organizers were actually recording these speeches on official video to publish on their website, if you find another better recording of this keynote please post the video link in the comments.
Here’s a new comparison between Pixel Qi and the iPad followed by Mary Lou Jepsen’s status report on the latest Pixel Qi news, their first showing of the new 1280×800 thinner 10.1″ wide view screen.
Tudor Brown, President of ARM, delivers a keynote speech on the status of the ARM industry and where it’s headed. Here at the Computex 2011 trade show in Taipei Taiwan, look forward to a whole bunch of new ARM Powered devices. As 25 Billion ARM processors have been shipped thus far, look forward to even more uses for these ARM Processors, with the Internet of Things and infinite other uses to turn everything smart.
Google wants to replace your wallet, your passport, your ID, to be used for ticketing, for local offers, coupons, deals and more. But is it secure yet?
We need this pin code mode and it needs to be fully 100% secure. But is it yet secure in this first implementation with NFC on the Nexus S 4G? Does Google yet use some type of deep hardware level security like the ARM TrustZone Mobile Payments platform?
We need this pin code screen to show up full screen, and there needs to be some kind of light diode indicator confirming that you are in 100% secure mode. That kind of pin code screen needs to come up to confirm every login, every payment, every money transfer. If they can do that in the way ARM is suggesting with TrustZone, this should make of this system a fully secured way to replace wallets, ID, Passports, tickets, coupons and more.
I want to login to my Google Account using my phone’s pin code security system. I want this system to replace all login username/passwords on the web. This system needs to become the new interface for a new type of OpenID system. Google released in February an SMS based secure login service that they offer to all Google Account holders today. But SMS is not seamless, it’s not really usable, the pin code screen needs to popup on your smartphone right there as you are trying to login, authenticate your access or to pay for something. That pin code authentication mechanism could perhaps be replaced by some kind of bio-metric authentication, or a kind of screen lock mechanism. Think of it like that calculator that you use for your security for your net banking, it needs to be the same integrated right into your phone.
Here’s the 1-hour video of this Google Wallet announcement, embedded to start at time-code 22 minutes (you can rewind and watch the whole thing if you want) where Rob von Behren talks about the NXP PN65 based Secure Element solution, which sounds like this is true hardware based security!
He has said it before, but perhaps not as explicitly when talking about tablets, and it’s always fun to see Eric Schmidt, Chairman of Google, explain Android’s strategy vs Apple.
Eric Schmid:
It’s a classic contest in high-tech. In that contest, you have a very well run, very focused, closed competitor who builds a great product that does something that is very usefull. That would be Apple. You have another competitor who makes all the technology available to everybody else, and using various creativity and various partnerships gets the benefit over everyone else’s creativity. Because there are more people involved in the open side of that, that side will eventually get more volume, have more investment, therefore have more creativity, more innovation, and ultimately the end user will choose the open one over the closed one.
Fareed Zakaria:
Except right now, the open one, all these tablets that are Android based, let’s be honest, they are not as good as the iPad and they are more expensive, which strikes me as unusual.
Eric Schmidt:
Which product will produce a lower cost product quicker? One manufacturer for a product or many manufacturers competing? The matter of fact is that we are just at the beginning of this fight. And the fight between two very well run, very large, very significant ecosystem companies, will ultimately produce great value for consumers because the fight between them will keep prices low, keep these systems honest and open and encourage the kinds of investments that people want to see. One of the greatest things about this contest is that people who win in this are the consumer.
(…)
There is pride in both approaches but they are completely different. In Apple’s case, they can continue to build beautiful and excellent products. The ecosystem that Google represents will continue and already has more volume, more users and will have more investment in the platform. Ultimately that will produce cheaper, better and faster products for everybody.
for the price issue, Chandra said that Google has now qualified ARM chips to be used to run the Google TV software, instead of just the Intel Atom chips that currently power the Revue. Moore’s Law–the inevitable increase in chip performance driven by increasing transistor density–will push the performance of the cost-optimized ARM chips up high enough to compete with Atom, while helping drive down the overall platform price, Chandra said.
I have thus far video-blogged over 60 ARM Powered Set-top-boxes from all the consumer electronics trade shows over the past 2 years, most are running Android, all of which could in theory run the Google TV software.
Of course, it is up to Google to decide what kind of hardware requirements they want to enfore for Google TV on ARM, if they want those to only feature the full suite of HDMI pass-through features, meaning HDMI input and output, Infrared blasters (to change the channel on your cable/satellite set-top-box), USB hosts, Bluetooth and more, then that would disqualify just about all of the ARM Powered set-top-boxes that I have filmed thus far. I wouldn’t know how much more those hardware features require, and perhaps that requires an ARM Cortex-A9 at the minimum to run all the overlay user interface features and preferably 1080p at 60fps support at the minimum.
I think it is more likely and more logical that Google will decide to be as open as possible about Google TV on ARM, and thus support all the SoC that are currently being used and that will most likely be used. I think that means Google TV on ARM could work in “AppleTV/Roku mode”, meaning no HDMI inputs, just the Google TV experience of bringing the Web and WebTV on the TV on this separate HDMI port to your HDTV. That is why I expect there to be two kinds of Google TV on ARM:
1. Basic Google TV on ARM, this is HDMI output only, Bluetooth or RF/USB keyboards with mouse pad accessory can be used. This solution could work on 100% of the ARM Powered Set-top-boxes that I have filmed. And I believe this will be included turning every Android Smartphone/Tablet with Ice Cream Sandwich and every Tablet with Honeycomb 3.1 into a Google TV “for free”. See the Google Android Team’s response to my question submitted on the possibility of turning all Android devices into free Google TV devices when HDMI is used:
2. The Full ARM Powered Google TV experience, including HDMI pass-through, IR blaster, USB hosts, Ethernet, etc.. Since Chris DiBona answers to my question above “There’s all this other stuff that goes into a Google TV that isn’t in a phone”, well then, the Full ARM Powered Google TV will be that type that does it all. But that should not prevent an Android device with a basic HDMI output and not much else to still display many if not most of the Google TV UI features right there on the HDTV.
3. There is also a third scenario that I am envisioning, Google might use their Android Hardware division to plan out a new type of Multimedia TV Docking system for Android, using nothing more than HDMI, USB slave/host and evt MHL (that combines both into one Micro-USB connector). Basically the idea here is a cheap TV Dock that should work with most if not all Android Smartphones that have HDMI, USB (or MHL) to turn those into full Google TV, thus using the USB slave/host to transmit the right infos back and forward and feature in that Dock the right HDMI input and output, IR blaster, USB host duplicators, Ethernet connector, charging and more. The idea is a new Google Open Accessory design that could sell for $49 to dock any Android Smartphone with HDMI/USB or MHL and thus turn those into full Google TV. A solution which could evt also turn any ARM Powered Set-top-box into a full Google TV box also with adding the HDMI in/out, IR and more to those. Maybe it could be called the Google TV adapter, converter or extender.
Here’s the 56 minute session on some of the Google TV Honeycomb 3.1 upgrades and development tools at Google I/O:
My take on it is that the Chromebook is the first serious challenger to Windows/Mac in terms of being installed in a mass market retail product. It’s the first ever mass market Linux laptop (after the One Laptop Per Child non-profit reaching 2.5 million children with Linux Laptops in the developping world since 2007). It’s the first ever mass market ARM Powered laptop. It can be configured to be the cheapest laptop to make, the safest, the fastest, the thinnest, the lightest and the easiest to use. Chromebook may be the first successful carrier subscription based laptop.
For Chromebook to sell more than Windows, here’s what I think Chromebook needs to be:
– $199 or less in an ARM Powered configuration
– Use Pixel Qi with ARM and you’ve got 30 hours battery runtime in a sub-1kg 11.6″ or 12.1″ super slim form factor
– They should subsidize these in partnership with the carriers to do a subscription model for normal consumers like this:
1. Sell it for $99 or less on a 2-year contract with $10/month/100mb or $20/month/1GB 3G/LTE data plan
2. Bandwidth upgrades should be max $10/GB, $20/5GB, $30/10GB on-demand one-click
3. They can use carrier billing (thus carrier revenue share) for bandwidth upgrades, for cloud media subscriptions, on-demand, Chrome Web Store web apps and for all Google Checkout based online shopping
4. Provide an optional hardware upgrade once a year with contract extension. Used devices can be resold refurbished.
5. Provide 100GB or more cloud storage and full Google Apps for consumers with the subscription, offer guarantee of available of advanced web apps such as HD video editing (with many or most of the features of Avid/Finalcut), photo editor. And all these web apps must feel near instant to load and work offline, a web app should only need to get reloaded if it detects that there is a new version available. Gmail should load instantly for example.
6. Obviously, Google Voice and Google Music needs to be worldwide. They should also expand with a Google Video cloud storage. Basically they can allow people to upload 20’000 songs and 1’000 movies for free, the reason being, Google only needs to store one copy of each song or movie, and if the upload client (also on Chrome OS) detects that the file you want to upload already exists on Google’s servers in equal or better quality, it should instantly beam it to your account without actually requiring you to upload anything. Google should not care to try to filter out any “illegal” Mp3, Flac, DivX, MKV files. Eventually they can introduce unlimited music/movies subscription plans like Spotify/Netflix but they should aim at being able to include access to everything in those unlimited subscriptions, this might only be achievable through Government regulation of online content subscriptions.
If Google can deliver on those things and quickly, which is what I expect them to be able to do, then I think it’s obvious Chromebook could become the number 1 PC/Laptop OS as quickly as they became number 1 OS in smartphones since the Nexus One was released.
Can we expect to see some ARM Powered Chromebooks (or Chromiumbooks) at Computex in Taiwan at the end of May from all the Taiwanese notebook designers (Inventec, Pegatron, Wistron, Foxconn, Shuttle, Gigabyte etc..) who design upwards 90% of worldwide notebooks?
The biggest announcement at Google I/O was the launch of the Android@Home Open Accessory Development platform. This is the platform for a whole new world of accessories and connecting everything through Android to the Internet. Suddenly, we are smart about all things and all things become smart.
Now we are not only talking about about connecting 7 Billion people to the Internet with Android Smartphones, now it becomes about connecting 100 Billion things through Android to the Internet.
Why the Internet of things? Why using Android?
The cost to add an ARM processor such as one of the ARM Cortex-M series, with sensors, switches and wireless connectivity in every appliance in your home may cost as little as few cents or a few dollars per device. It’s so cheap that as soon as an open standard is established and as soon as applications are planned out, all devices will get connected with this technology. Watch my video with Nuvotron NuMicro at Embedded World 2011 about the cost ($0.50-$2) and use of ARM Cortex-M0 32bit microcontrollers in all types of devices.
Here are some of the infinite amounts of uses for putting ARM processors in everything:
– Put a smart control in every lamp and the lights follow you, if you move to another room the lights automatically turn off, you save power. They automatically dim if they detect you’re relaxing or watching TV.
– Put a smart control in all your doors, in all your windows, in all your power outlets, integrated with your heating systems, water systems.
– Add sensors, ARM Processors in your pillow, blanket and in your bed, to monitor your sleep and wake you up at the right time between the right sleeping cycles. You’ll feel better the whole day and you’ll optimize your sleeping times. It’s more healthy, makes you more productive and saves you time.
The trick is that even as the Internet of things has been possible for a while, and even as prices to add smart controllers and sensors in each thing costs $3, people haven’t been doing much of it yet just because the control, management, interactivity systems around this have not been standardized and open yet. If you want to build the Internet of things it has to be built around you and for you and not among each thing and only for each thing. That is why Android is your interface into that world of things, and Google supports the open Adruino platform to enable these developments in an open industry. Android is the UI for the Internet of Things. Android is how you guide it, how you see it, it’s how you control your things.
Expect the next consumer electronics trade shows to showcase more and more ARM Powered things to connect with the Android ecosystem, look forward to an industry about to get really creative in how to use and feature that Internet of Things most efficiently.
ARM President Tudor Brown talks about the Internet of Things at the ARM Technology Conference:
Andy Rubin’s former co-founders on Danger (Sidekick, see these videos from 2004) Matt Hershenson and Joe Britt demonstrate and launch the new Android Open Accessory API and Android@Home platform at the Google I/O 2011 Day 1 Keynote:
As a reply to my ARM Powered Chromebook Chrome OS notebook question posted online, the Google Chrome Team confirms that while they are focusing right now on getting the Intel Powered Samsung and Acer devices out the door on June 15th, the Chrome OS team is also working hard on ARM support and the ARM Powered Chromebooks should be released soon after. Their wording is the ARM Chromebooks are “hot on the trails” of the Intel Powered ones. This could be the first mass market ARM Powered laptop ever released.
The advantage of an ARM Powered Chrome OS device is that it is thinner, lighter, runs longer on a battery and especialy could be sold for a lot cheaper. I’m expecting the ARM Powered Chromebooks to sell below $199 for them to become the new best selling notebook platform.
Chrome OS is I think the first real challenger to Microsoft Windows and Mac OSX to actually be able to reach a mass market through retail stores and carrier subscription plans. Google has announced the global release of Chromebooks starting June 15th, but for now, those are only Intel Atom based. I asked them about the ARM Powered Chrome OS on their Google Moderator, let’s see if we get a good answer about the status of Chrome OS on ARM, the live Fireside Chat with the Chrome Team starts now at 3PM PDT (Midnight Central European Time):
Please vote now for the following two questions for the Google I/O Fireside Chat with the Chrome team (to be live streamed on http://google.com/io at 3PM PST today, about 2 and a half hours from now), please use the following links to the Google Moderator to up-vote my questions so that we are more likely to get these answers:
When are you showing the ARM Powered Chromebooks? Are ARM Cortex-A9 processors with lots of RAM fast enough today? What is the status of Chrome OS on ARM?
What hardware requirements do you enforce for manufacturers to get access to the Chrome OS source code? Can anyone use Chromium OS today with the full Chrome Web Store on any hardware without asking Google for permission?
Also feel free to add your own good questions to the Chrome team in that Google Moderator and post your question links in the comments so we can also vote for your questions.
Who in the industry has access to Honeycomb 3.0 and 3.1 source code today? We were hoping for Google to announce 3.1 being open sourced around Google I/O but now it seems Google might not provide any Honeycomb open source code before Ice Cream Sandwich in Q4 this year?
I understand the gigantic work involved for Google to write all the code, implement the programming APIs and everything else involved around Ice Cream Sandwich. Regardless of how quickly Google put together Honeycomb for tablets to have something ready at Mobile World Congress in February, I think it would just look wrong if for some reason we only have $500+ Android tablets based on Tegra2 made by a handful of priviledged companies somehow having any type of Honeycomb software on them for another 6 months. Given the amount of companies (nearly 375 of them filmed here), small to medium sized, who are investing their futures in making Android tablets, Tegra2-Honeycomb-exclusivity-until-Q4 would probably be quite scandalous. This is what I am expecting must be happening right now secretly with the Honeycomb 3.0 and 3.1 source code behind the scenes:
1. All “serious” tablet companies using all the major ARM Processors do have access to Honeycomb now, or will get access very soon. By “serious” company, I could mean the companies in which Google can trust not to leak the source code. That could mean that these “serious” Android tablet companies need some kind of a track record of being serious with this market.
2. Google should be transparent about which chip provider does have access to Honeycomb source code today, and which chip provider will get access soon. I believe all chip makers from at least ARM Cortex-A8 performance and upwards should be allowed to work on optimizing any current Honeycomb source code to work and timely be shipped with all the tablets that do get released with those specific chips in them. I do not believe that a Tegra2-only club for Honeycomb would be taken with a smile from the rest of the industry. All chip providers that have tablet makers showing products on the market and showcasing them at all the “serious” tradeshows today, including AmLogic, Freescale, Marvell, NEC/Renesas, Qualcomm, Rockchip, ST-Ericsson, Samsung, Telechips, Texas Instruments, VIA, all those should get that access and be able to ship Android tablets with Honeycomb in Q3 this year.
I’ve sent some of my Google contacts some questions regarding the actual status and plans for Honeycomb’s source code and support on the variety of ARM chip providers, while I am waiting for their reply, I wouldn’t know for sure what the actual happenings are behind closed doors before, during and after Google I/O in terms of officially supporting Honeycomb 3.1 on other platforms than just Tegra2.
Google needs to officially confirm that they are working with these ARM processors to support Honeycomb 3.1 in Q3 this year and I think that most Android tablet fans would be totally happy and satisfied:
If Google has been and is cooperating with at least each of these SoC platforms, then I think we have no problem.
But if on the other hand, it somehow turns out that most of these alternative SoC vendors are somehow locked out of the Honeycomb party until finally getting source code access in Q4 and maybe not being able to release actual tablets with that code before Q1 2012, well then I think there will be some very angry people around the worldwide Android tablet industry.
Given the relatively big level of secrecy from all SoC vendors involved, I would like to interpret that as a clue that they must all be silently and cooperatively working closely with Google ever since before even the Motorola Xoom was released last February. And if they didn’t all have access already before last February, then hopefully they have all quietly gotten access or are getting access by now.
Google announced that over 100 Million Android Smartphones have been sold thus far. If 80% of those are sold on 2-year contracts generating revenues at an average of $1200 per phone over the 2-year contracts (in the US for example that number is most often higher than $2000), then that could mean Google’s Android has a huge influence on a global revenue of potentially more than $120 Billion, possibly over $80 Billion of which have been generated just in 2010 alone, with 2011 global Android industry revenues possibly clinching upwards twice as much as Smartphone growth is more than doubling every year. The Android Smartphone may be a $160 Billion industry in 2011 alone. Over $250 Billion in 2012 maybe. We are not talking peanuts. And with all analysts saying Tablets are the post-PC interface, Google may feel some type of pressure from the big guys of tech, not only the manufacturers but also the carriers (who are touching most of those huge Android related revenues), so it can be understandable that Google does things very carefully around Android, and to bring Android’s market share from 40% to over 80% in the next few months, Google may want to focus on top level secrecy in all of their cutting edge Android developments.
While I can understand that, and as a huge Google fanboy I want them to dominate over everything, but let’s see if we can get more informations on Honeycomb openness and the industry’s access to Honeycomb now under this Google I/O conference. I haven’t yet watched the Google Executives Q&A with Andy Rubin where some of the Android openness questions are answered, if anyone has the link to that video or any other related sources of informations please post those links in the comments.
If you are an industry insider and if you would like to tell me any secret information about the status of Honeycomb in the industry related to 3.1’s likelyhood to work on any or all of these SoCs during Q3 this year, you are welcome to contact me at charbax@gmail.com and if you want I can keep your name secret if you allow me to report here on your info.
