I've been monitoring how many developers are watching all of the segments of the video series we've posted on the Take Five site.   It is natural that not everyone will make it to the end, but I'm wondering if there is anything we could do better.

First, I'd like to know if any of these are the causes:
  A) Didn't know series were there
  B) Watched the video from a different page w/out the series navigation
  C) Didn't have time to watch them all
  D) The video, well, just wasn't that interesting
  E) My boss walked by, and I had to quickly switch screens from videos
  F)  Other:  ______________ 

Take a look at two of the series we've posted from the Game Developers Conference:  Optimizing DirectX for Mulicore, or the SkullTrail series with GRIN Software* and UbiSoft*.  Or the very long Confronting ManyCore series.   Once a video is playing in the player at the top of the site, there is a link that says "see Next in series" and/or "see Previous in series" to easily (at least I thought) move from one to the next. 

While you in the player, you can rate the videos, get the embed code and direct links, leave comments, and/or link back to the blogs.

To see more series, they are in "series boxes" on the tabs below the Take Five player.  If you mouse over the numer, you will get a video description. Be sure to click on the blue arrow to see the Virtualization and Community, and coming soon, Open Source tabs. 

I'd love to hear your comments, and answers to the mulitple choice above.  I'm hoping I don't get too many D's.

Happy Friday!

I cannot envision a more exciting time to be working in the technology industry, and especially, at Intel. Tick-tock, multi-core, low power, atom, larrabee, wimax, mids, 45nm are just a few of the potentially game changing strategies and technologies that Intel is driving today. However, none of these breakthroughs will make a big difference in the marketplace unless ISVs can effectively incorporate them into products that deliver real value and excitement to customers. This is the ‘simple’ premise behind ISPP: Make it easy for ISVs to align their products with the Intel roadmap and enjoy increased business success as a result.

In my past life as an engineer, I got a real kick in the pants from creating new SW and hearing feedback from customers about how cool our app was…or sometimes, wasn't! These days my ‘kick’ comes from driving our talented engineering, ISN developer community, and marketing teams to deliver enabling programs that help you and your company to capitalize on Intel's technology leadership. I’m looking forward to sharing my thoughts and ideas about SW enabling in future posts, but am even more eager to hear your feedback about how we can make ISPP work better. Simple enough?

First, ...2014?  I'd love to know what is going to be happening by the end of this year, let alone six years out.   That's an Internet Eternity, especially in this product area.  How old is  your cellphone, smartphone, or favorite tech toy? Would you have really envisioned how you use it two years ago?  Four?

Second, there are too many wild cards in the cast of characters.  Apple, Google, Asus, Microsoft, and even... what's their name again... oh yeah, Intel, might have a plan or two up their sleeve in this time frame, and any one of them could send things scrambling in a new direction.  Apple wrenched the phone world sideways with the iPhone, and Asus launched a whole new market with the Eee PC.  Google has Android cooking and never fails to keep developers amused with new forays, and good ol' Microsoft may be sometimes slow to the party, but they always show up with an entourage.

Third, the article claims the "proven advantages of the ARM ecosystem in mobile devices will eventually outweigh those of the Intel platform."  I've talked to a lot of people inside and outside Intel who develop for these devices, and the fragmentation of the existing ecosystems drives many of them nuts.   One of the great things about moblin is that is provides an compliance base across devices so you can reduce testing and reach more target systems, even running on multiple form factors like MIDs and netbooks.

ARM has an impressive part of the market right now in MID and ultra mobile devices without a doubt.  The iPhone alone tips the balance quite nicely.  I just don't see how anyone watching this industry, and who has looked at Atom and Intel's plans for it, could do anything but chuckle at any six-year prediction around MIDs other than it is going to be one crazy ride.

Mobility technology provided by Intel & its partners such as Lenovo makes the rest of the world a step closer to the Everest. At the 5200M base camp, the world's highest Internet Café is serving & hosting the journalists from around the world, including CNN & Reuters. In this tough environment, there are 10+ Thinkpad T61 notebook powered by Intel, connected through wireless and reach out to the world. The latest Intel Atom based Ideapad U8 Olympics edition which is the first MID product from Lenovo provides convenient connectivity to the digital camera and help the press folks send the picture home with a few clicks. From base camp & up, you can spot a lot of notebooks powered by Intel mobility technology, providing the access to the internet for personal blog or heavy duty work such as editing the HDV video footage using software Non Linear Editing system and photo retouching & send information back home.

How mobile can you be? It is up to you. But guess what, there is a Lenovo X300 notebook that is up at the top of the Everest, which is a place thinner than air. Now it has been safely back to the 6500M camp with his owner and working happily there. How high can your mobile go?

Photo Courtesy of Xinhua News Agency.

Now I think PowerPoint is a fine tool, we just use and abuse it beyond its intended function. For example, sometimes a discussion is just better served with visual cues, whether charts, pictures, or even (I suppose) those stick figure clip-art people. Likewise a discussion sometimes needs to have a constrained, focused agenda, which is what I get with a series of PPT slides. PowerPoint definitely has some valuable uses.

In our case, we wanted an interactive discussion. The scenario was a presentation on the status of the mobile developer community on ISN to people two levels up the food chain. The recipients were my manager, his peers, and his boss. The outcome we wanted was to level set on the status of our community, get input into our strategy and plans, and actively discuss some of the challenges we're facing. Interactive discussion = Good. Silence while everyone does email = Bad.

Our preparation was to put all of the background material and presentation content on a page into an outline in our internal wiki. The wiki allowed for collaborative editing, and brought a lot of transparency to the discussion. We didn't want to surprise the viewers with the no-PowerPoint approach, so we sent out the wiki address two days in advance with an invitation for comments. We cautioned them that the content wasn't finalized, but welcomed input on the planned topics and outline. We incorporated several comments into the final presentation, so it worked well.

During the presentation itself, I held the page open for editing while others viewed it either projected in a conference room or on their own laptops. We explained the agenda of the presentation (Current Status, Future Plans, then Roadblocks & Challenges) and encouraged people to ask questions about anything they read - not just what I was talking about. We knew we couldn't cover every point in the wiki, but by letting people read on their own we could jump to the hottest areas of discussion for more detail.

Keeping the wiki open for editing allowed me to make changes as we talked. At one point I was able to update some metrics that were IM'd to me by someone else on the call. We added some of the follow up tasks and questions as we went, then fleshed it out in more detail afterwards. The wiki page is now the full agenda, content, notes, and meeting minutes for the entire presentation.

Specific thoughts on using a wiki for presentations:

• More interactive discussion - People were not bounded by the topic of the current PPT slide, so we got exactly the sort of engaged debate we were after.
• Trickier to guide the flow - The flip side of the above point is that without a slide declaring "This the current topic", things jumped around. We had to be willing quick on our feet.
• Have to know the material - I do not recommend the wiki approach if you don't know your topic very well. Again, this is interactive so there isn't any room to hide if people start firing hardball questions.
• Easier collaboration - Ability to create the content for the presentation in a centralized area rather than passing around files saved tons of time (and had revision tracking for a bonus).
• Standard format - Everyone didn't have to have a Windows machine to follow along. Also easier to print for later reference, or to grab sections out to drop into other presentations or collateral.
• Greater detail - Only so much detail can go onto a PPT slide before it becomes an eye chart. In the wiki, people could read in greater depth, read it quickly, then ask questions.
• Not as visual - If we had a more visual oriented presentation rather than discussion based, this may not have worked as well. To me, one of PowerPoint's strengths.
• Harder to time the discussion - We use a "10 minutes per slide" rule of thumb when presenting PPT, but I had no idea what that translated to for a wiki. We included more material than we could cover, which we need to watch closely. It's not entirely bad because people were still able to read it and get the info even if it wasn't discussed.

Overall, it was a definite success and we'll be trying it again in more meetings. It helps that our team is willing to try odd things in the name of both Social Media Experimentation and Saving The World From PowerPoint Poisoning, but I can't imagine others won't be willing to give it a shot.

I hope this was of interest, and I'd love to hear any thoughts on how to improve this technique for the next time around!

Fundamentally, it's all about how open your project really is. I thought the days of saying your product was "open" when you published the APIs developers could use was over but based on the panel discussion, I was clearly mistaken.

To me, openness is about doing your development in a manner that allows people who are interested and capable to contribute. Clearly, you don't want to give world write access to your source code repositories but you do want people who have proven their ability to contribute positively to do so during the course of the development process. And this is one of the things that differentiates the projects under Moblin.org from some of the others that are trying to create a mobile Linux stack.

There is more that we plan to do in the Moblin project to make it easier for developers to contribute, but for now I think the approach we are taking is being well received by the community.

If you are a developer interested in creating software for MIDs and other devices but you haven't looked at the Moblin.org project yet, I strongly recommend you do so. Go to http://www.moblin.org and check it out. There is a lot of good stuff there with more coming all the time. We appreciate your input and feedback - please sign up for the Moblin developer mailing list at http://moblin.org/community/community_mail-list.php

Thanks,
Derek

There's got to be a catch

There are 5 CC-states and, depending upon how you count, 6 PC-states in the Penryn line of Intel processors. And, in Microsoft XP, there are 4 OS C-states. So are there 5 C-states, 6 C-states, 4 C-states, or 15 C-states? Choose the number that you are least uncomfortable with. Personally, I first imagine a 3 set Venn diagram with overlapping elements and added transition annotations. Then I get confused and give up. 

Given what we've talked about above, it seems as if we should always drop a core into the lowest permissible CC-state, right? 

There are a few reasons for not doing this. First off, the OS's Power Management (PM) policy, and not the hardware, determines when a core enters a CC-state. From our standpoint as a hardware manufacturer, we have little to do with this. I'll talk about why this is important later. Secondly, there is always a cost for dropping into a lower C-state. That cost is the amount of time required for the core to transition from an idle state, e.g. CC5, to C0. As you start using deeper CC-states, latency becomes significant. For example, the latency to go from CC3 to CC0 is around 20 us, literally ages when we're talking about 3 GHz processors. 

This latency penalty is even worse once you realize that the phrase, "in a given C-state," is misleading. As I've mentioned above, it's easy to think of a core as descending as a waterfall from C0 into C1 into C2 into C3. (See Figure A.) If this were the case, you'd suffer only one 20 usec penalty. No, it's oscillating between C0 and C3 hundreds, if not thousands, of times a second until the OS's PM code decides that the percentage residency merits ascending / descending to the next C-state (e.g. CC3 to CC2). In Windows, the C-state that a core transitions to is based on the % idle over a certain interval. This means that each transition exacts that 20 usec penalty, and there are hundreds of transitions. Doing the math, experiencing a 20 usec delay 100 times a second is a whopping 2 msec of added latency per second. (See Figure B.) 

Even if it is possible to drop a core into a deeper CC-state, the OS has to ask itself various questions, such as what is the likelihood that processes are going to be doing more work very soon, so that dropping into a deeper CC-state might actually cost an unacceptable penalty? Similarly, the processor has to ask whether dropping a core into a lower CC-state is going to cause incorrect operation, say whether the delay in the processing of an interrupt will cause an event to be lost.

Figure A. The waterfall misconception.

Figure B. What actually happens "in a given C-state".

I decided to upgrade the version I have running in a VirtualBox guest under Windows XP on my Sony VAIO laptop with a Core2Duo processor. (VirtualBox and OpenSolaris is of course open source software and totally free).

Of course, my job seems like a constant stream of interruptions (just the nature of my job) so I started the upgrade late in the day, and had to go home while the upgrade was still running.

When I came in this morning, I noticed that the laptop appeared to be powered off. Oh yeah, I forgot to change the defaults, so Windows had put me in what they call "Stand By" mode. This is the state referred to as "Suspend to RAM" because the processor is powered down, but the memory remains powered.

Uh oh, did the upgrade finish before the suspend? Well, let's find out.

I hit the power button, and Windows came back up. The OpenSolaris guest was running, and I could see that the Upgrade progress was only at 35%. Since I was doing the Upgrade from a DVD, I figured that it was a totally lost cause.

But hang on, the OpenSolaris guest is trying to access the CD, and to my surprise, the upgrade starts up again and continues to progress normally!

I don't know why I should be surprised by this. The exasperated engineer will say, "Of course! It worked just like it was supposed to!" Yes, but in the area of suspend/resume power management, I am normally accustomed to things either not working or working very slowly.

Anyway, after this experience, I will definitely feel more comfortable using suspend-resume on the host OS.

I'm happy because I think this could be a great resource.  The good feedback I received from my Atom 101 post made it clear there are lots of questions out there around MID development.  Even the definition of a MID has a lot of variations, then throw out UMPCs, UMDs, and Nettops on top of it   How does moblin compare to Android?  What is Atom?   I'm still sorting it all out, and a fleshed out wiki would certainly be a huge help.

I'm cautious because this is a different approach than we've tried before, and because the wiki isn't already full of content.  I'm hoping to start moving some of Intel's static, PDF, white-paper content into dynamic, evolving, community driven material.  This is a new idea to a lot of internal people, and my regular question of "Could we add this to our new wiki?" is usually met with silence or some hesitation.  Fortunately a few hearty souls are willing to give it a try so I will keep at it.

I'm also cautious because I'm deliberately opening the wiki up without content so it doesn't start as a giant Intel document repository.  I want to get lots of Intel content in there as part, not all, of the collection.  As long as it relates to mobile device development, this wiki is wide open to non-Intel content.

I am going to keep adding material, and urge people to at least subscribe to the new articles or edit feeds.   We'll do more advertising of the site once it reaches a greater critical mass of content.  In the meantime, I'd love to see people hop in there and add their own material, or leave a comment here and tell me what would make the MID wiki a valuable tool for you.

For those of you still wondering what I'm talking about, I'll let you in on a little secret. We're showing the old design to about half of you, and the new design to the other half. Yes, this is intentional. We're collecting metrics to guage the effectiveness of the new design.

Would you like to do me a favor? Let me know what you think about it.  Does it work better than the old design?  How would you improve it?