Petafloppers, Quantum Supercomputers, ISC 2013 and More…

HPC geeks ponder 100 petafloppers and quantum supercomputers

ISC 2013 The next big barrier for supercomputing is punching through 100 petaflops peak performance, which frankly could be done in a heartbeat if someone had a few hundred millions dollars lying around. And now that Google and NASA are monkeying around with a quantum computer, thoughts are turning to how a QC might be deployed to replace some of the work done by traditional supercomputer clusters.

Euro students cluster fest: Configurations LAID BARE

The configurations of the systems to be used by the young HPC warriors in the 2013 International Supercomputing Conference’s Student Cluster Challenge were released last week

Edinburgh students’ heaving racks: UK’s only hope for cluster-wrestling glory

Eight universities have traveled to the 2013 International Supercomputing Conference in Leipzig, Germany, to participate in the 2013 ISC Student Cluster Challenge. They’ve deployed their clusters and are busily working to turn in the best results on a series of HPC benchmarks and scientific apps.

SDSC GeoComputing Lab named winner of HPC Innovation Excellence award by IDC

The High Performance GeoComputing Laboratory (HPGeoC) at the San Diego Supercomputer Center (SDSC), an organized research unit at the University of California, San Diego, was named a winner of the HPC Innovation Excellence Award by the International Data Corporation (IDC) for developing a highly-scalable computer code that promises to dramatically cut both research times and energy costs in simulating seismic hazards throughout California and elsewhere

 
nCore launched BrownDwarf, an actual ARM- and DSP-based supercomputer. What started in cell phones has moved upwards into smartphones, tablets, servers and now even supercomputers as well.
 
When I look at the technology ecosystem, I look far and wide. I take off the covers and peek inside. I marvel not just at Instagram, but I marvel at innovation that Apple has packed inside the iPhone and its laptops. Did you know that Apple has employed some of the top material scientists who work with other Apple engineers and help craft together the MacBook Pro or the iPhone 5? The material science isn’t all that exciting to some but to us it isn’t incremental work. I look at those videos Apple puts out about how its computers are made, and I get excited — because I see old fashioned engineering & tinkering at work to make my computing experience even more beautiful.

How Virtualization is Key to Unlocking Cloud HPC

Josh Simons from VMware discusses why the high performance computing community is starting to leverage virtualization technologies for Cloud HPC.

BookLog: D is for Digital

From D is for Digital by Brian Kernighan

Some computing is highly visible: every student has a computer, each one of which is far more powerful than the single computer that cost several million dollars, occupied a very large room, and served the whole Princeton campus when I was a graduate student there in 1964. Every student has high speed Internet access, as does at least half the population of the United States in their homes. We search and shop online, and we use email, texting and social networks to keep in touch with friends and families.

But this is only part of a computing iceberg, much of which lies hidden below the surface. We don’t see and usually don’t think about the computers that lurk within appliances, cars, airplanes and the pervasive electronic gadgets that we take for granted — cameras, video recorders, DVD players, tablets, GPS navigators, games. Sometimes their pluses and minuses come to the surface, almost accidentally, as in a newspaper article that quoted an executive of Hewlett-Packard as saying “In essence, a digital camera is a computer with a lens.” The same article also quoted an unhappy consumer: “This isn’t a camera, it’s a computer,”comment on how hard it sometimes is to use computers.

Nor do we think much about the degree to which infrastructure depends on computing: the telephone network, television and cable, air traffic control, the power grid, and banking and financial services. The pervasive nature of computing affects us in unexpected ways. Although we are from time to time reminded of the growth of surveillance systems, incursions into our privacy, and the perils of electronic voting, we perhaps do not realize the extent to which they are enabled by computing and communications.

Meta:

I was reading this interview with Brian Kernighan and came across the link to the book. I looked at the table of contents and reviews and felt that it may be a great book for the first year Engineering students. In Chennai Engineering colleges, they have a subject titled Fundamentals of Computing. While the CS students know what it is about, many others from different branches of Engineering, are not able to relate to this topic. I was hoping that a book like this would make students understand the “computing that is all around us” .

 

 

Kernighan, Brian (2012-02-04). D is for Digital (Kindle Locations 73-83). Brian W. Kernighan. Kindle Edition.

Some Examples of Machine Learning Problems – BookLog

Statistical learning plays a key role in many areas of science, finance andindustry. Here are some examples of learning problems:

  • Predict whether a patient, hospitalized due to a heart attack, will have a second heart attack. The prediction is to be based on demographic,diet and clinical measurements for that patient.
  • Predict the price of a stock in 6 months from now, on the basis of company performance measures and economic data.
  • Identify the numbers in a handwritten ZIP code, from a digitized image.
  • Estimate the amount of glucose in the blood of a diabetic person,from the infrared absorption spectrum of that person’s blood.
  • Identify the risk factors for prostate cancer, based on clinical and demographic variables.

The science of learning plays a key role in the fields of statistics, datamining and artificial intelligence, intersecting with areas of engineering and other disciplines.

Meta

Not a single word above is mine. One of the tweets I received talks about using R for machine learning. I am curious about this subject and may even use it for some of our products. R comes up a lot in many things I read about data. So I decided to investigate. Here is the journey of articles I found and ended downloading the free book on Elements of Statistical Learning.

  1. I lost the link to the tweet and decided to get help from Google and searched for “R for machine learning”
  2. Found this link interesting and made it my starting point – Guide to getting started in Machine Learning.
  3. Found a fascinating link on How Google and Facebook are using R. Who can resist that?
  4. Got started with Elements of Statistical Learning

So here I am. I have not even started the book and not sure how long it will take me. I found this fascinating list of examples of machine learning and thought that it may be a great motivator for others like me who may wish to embark on this learning journey.

Does the Full-featured DBMS Scale to Web Scale?

I am a little attached to RDBMS and SQL since we worked on an engine called Integra SQL, for a decade starting in 1985. Declarative queries fascinate me. In fact declarative anything fascinates me since I think it is the only we can push complexity under the hood and have application developers deal with higher levels of abstractions.

I have been watching the NoSQL space for a while. I really don’t like that name. I have watched ORMs (object relational mappers), columnar databases, Entity stores (just another name?), document databases, Open Linked Data, RDF stores and the whole galore. There is a common thread. Large data sets, web scale computing, distributed data are some common themes.

So when I saw this link for the panel discussions on VLDB, I thought it deserved a read. I am not sure why a dated article fell into my infostream. Here is an interesting snippet.

Does the full-featured DBMS scale to web scale? Microsoft says the Azure version of SQL server does. Yahoo says they want no SQL but Hadoop and PNUTS.Twitter, Facebook, and other web names got their own discussion. Why do they not go to serious DBMS vendors for their data but make their own, like Facebook with Hive?

Who can divine the mind of the web developer? What makes them go to memcached, manually sharded MySQL, and MapReduce, walking away from the 40 years of technology invested in declarative query and ACID?

A few more interesting quotes from this page:

The appeal exerted by the diverse language/paradigm -isms on their followers seems to be based on hitting a simplification of reality that coincides with a problem in the air. MapReduce is an example of this. PHP is another. A quick fix for a present need: Scripting web servers (PHP) or processing tons of files (MapReduce).

query languages that were ever universally adopted were declarative, i.e., keyword search and SQL

It is an interesting space to watch. If this movement is the real thing, it may change a lot of the way we build data driven apps in the future.

Meta:

I found this article while I was doing some testing of InfoPro, a product we have been working on a for a few months.

A few more things you may want to lookup – couchdb, cassandra, mongodb

What Technologies Will Have Major Impact In the Next 5 Years?

During the lightning talk session at DevCamp #dcc1, I asked the audience to predict “What Technologies Will Have Major Imact in the Next 5 Years”. Here are some answers (in no particular order):

1. Google Wave

2. HTML5 and CSS

3. iPad like devices, interactive books

4. Windows 7 Phone

5. Mobile Applications

6. Mobile Commerce

7. Augmented Reality

Add to this list if your favorite one is not here.

Added four more from Satya (see comments)

8. Small Business applications (services)

9.  Human-Machine Interaction space

10. Online Gaming. (Virtual Reality)

11.  Social Applications

LinkLog: Some Social Thoughts

From Some Social Thoughts:

Remember that as a source of information, social media should be considered as just another source. Information that you find should be verified and validated, just as you would information from mainstream media. As far as social technologies go, don’t wait to long to embrace them.  Social is rapidly causing a fundamental shift in the way we communicate and share knowledge, and discover and navigate information. The sooner you begin to build your network, and communicate with your customers, the better off you will be positioned as adoption and usage of these technologies and networks continues to explode.

I totally agree. I also think that there is an opportunity for tools and services. How do you verify/validate the information?

  • Check the source (of the original post) – an obvious one
  • Look at their typical biases – pretty easy to do with a simple word/phrase cloud of their posts (more about this in another post)
  • Check the information from an independent source (more difficult when we are living in an echo chamber)

Do you know of any social media tools that can help you do this easily?

LinkLog: Measuring the Story Telling of 10s

From Horace Luke on Design As Story Telling:

I tell my teams to measure a storytelling of 10s. In the first 10 seconds they have the phone, do they understand what they are getting? Then the first 10 minutes: Is it easy to engage with the product, do a simple task? At 10 days, are they frustrated with the product or telling their friends about it? What are they starting to discover? At 10 weeks, what are the powerful features they didn’t use before? Are they putting folders and widgets for their own stuff in the phone? At 10 months, they may be ready for a change. Did your brand fulfill its promise, so they stay with you?

I think we need a model like this for software. Any ideas?

Freemium Model – Our Experience

Here is some notes from the Freemium session at Google I/O. I wanted to annotate it with our experience with InfoMinder.

Why would you offer something for free?

To give the user time to learn about the product

User can provide distribution benefit (inviting other users to the product)

Network benefit – adding value in some way

The scale of conversions is in a 1-10% range.

In our case, we sent email message to about 20 friends and had a network of about 3000 users in 6 months. We provided the product free for a year. Then we started charging for a professional version but kept a basic version free for some time. Our conversion rate, initially was slightly more than 10%. We had a much bigger payoff when two companies OEMed our product.

In the past no business app, except Intuit, has penetrated the 10 employee or less market.

The model is going after new segments and new opportunities with a scale that can make it work.

Direct vs. indirect revenue models.

Feedburner – free version and $5/month version with some extra features.

Our basic version is still less than $2.5/month.

Product segmentation of free versus paid side

Viral/growth oriented things should be on the free side

Things that engage users in deeper behavior should be on the paid side

Can be difficult to draw the line within a company of what should be free vs. paid

The pricing was a bit difficult for us. We initially priced it at $14.99 per year. Gradually moved it up to $30/year. Our current products range from $30-$5000 per year based on the configuration. We have a lot more customers in the mid range $10/month than in others.

Start from the beginning with the model you hope to put in place – put your business model in beta at the same time you put your product into beta

This is easier said than done for two reasons. One is that product evolves and also your knowledge of customers. It took us a couple of years to find the sweet spots.

If the free product is so good that there’s no reason to pay, then an alternative is to limit capacity. Where do you draw the line? On the selection of features or on the capacity? – find your fanatical users and have them help you segment where the paywall should be

This is exactly what we did. Our trial product is almost as good as our full product. We charge by capacity. Others like WordPress do the same.

Establishing the value of your product is probably more important than establishing a pricing model right away.

This is very true. We continue to get business and keep customers even though there are some free products that provide some of the functionality. One of our customers told us that after trying out a few competitive products, they decided to stay with us due to a combination of quality and support.

Your instinct about what your customer will/will not pay for is likely wrong. Be flexible early in your business to be able to listen to your user feedback. Have the right premise – if you need 100M people to use your product and it’s not viral, it probably won’t work, for example.

Question: Freemium seems to be when dealing with the direct consumer. What is the balance between different models?

In mobile gaming space — collecting pennies per user over a lot of users can make a big difference. A mix of revenue types (direct/indirect) can work

Conversion rates will probably be between 2% and 5%, realistically. Most people won’t pay you for features (you may think they’re more valuable than the users do). Make people feel like they’re getting more value than just additional stuff than they’re already users.

Go ahead and read the wave. It has lots of information. Please feel free to ask us questions. You can email me or contact me directly (my co-ordinates are in the About link).

Thank You Twitter

Some times I find some thing beautiful (like the Wabi-Sabi) and some times I find some thing useful (like how to design a Home Page of your Website).

I have the urge to blog, but don’t find the time. So I simply book mark it and move along. But since Twitter came along, I can do something better – Tweet it. Putting some context around the link hopefully helps. It helps me remember too. And then I can come back and blog a bit if I feel like it.

So Thank You Twitter. You help me remember things and keep a log  and share it as well.

Wish List for a New Tablet Like Device

The industry is buzzing about a tablet like device Apple may introduce. So I thought a bit about what I need from a tablet. One note of warning. My tastes are a bit geeky. That is reflected a lot in this list. If you are an author or musician, you may want an entirely different set of capabilities. I am assuming that it will have all the functionality of current laptops but may have a smaller size (a paper back), higher battery life and an extremely good display and a pen like device.

Essential

  1. I am assuming that it will be a general purpose device and hoping it to be the next laptop and do many things my current laptop based tablets don’t do well.
  2. A Reader – Any content (books, magazines, news papers, blogs, tweets) in any of the popular formats
  3. A Sketching Tool – I doodle a lot. I draw mind maps, simple sketches, flow diagrams in free hand. These are my thinking tools. So an easy to use, simple sketching tool would be an ideal companion.
  4. Annotator – When I read, I also markup and make notes. A tool for annotating would be compelling. Ideally if there is annotation framework for plugging in any annotation tool and support popular annotation protocols it may be even better.
  5. Mark And Invoke – A framework for launching applications that take a piece of marked text and invoke a mini-app will be great. Why? This is what I do when I read serious stuff. I jot down quotes, I mark items for future research, I Google a few, I check the dictionary/thesaurus and some times Tweet. Since I do not know what I may do in future, I would rather have a framework for plugging these mini-apps than bolted in functionality.
  6. A simple scripting tool. The device may replace my laptop. So some simple scripting capability to pull together applications may help me do my own.
  7. Ability to add my own apps to my device and give a few to my friends. I should have the freedom to write and add my own apps like I do on my laptop now, in a language I choose (currently Python).
  8. A good presentation tool – I am slowly shifting to handwritten sketch based presentations instead of the powerpoint style sanitized ones. So I am ok with drawing a few sketches, some notes and linking them up as a presentation. Would love simple wiki style hyper-linking across pictures, videos, sketches, images.
  9. A web synchronized wiki tool – Most of my learn logs, daily logs and idea logs all reside in desktop wiki. I would love to have one that automatically synchronizes between my desktop and my server (hosted) account.
  10. I don’t need this now but having some external plugins for various devices like webcams, sensors would be great 

Some Random Thoughts (Just thinking aloud):

  1. Cell phones brought in single finger typing. The current generations seem reasonably comfortable sending text messages using these tiny keyboards at the speed of light. Blackberry and other copycats ushered in thumb typing. Will the new tablets bring in a different typing model (optional). Holding it with one hand and typing with the other. Will they become palm holds?
  2. With location based everything, the new device has to have GPS. That may make it a great navigation tool, some thing that replaces my current automobile driving assistant, as well. Why should assitance stop with driving? Why can’t it be extended to visual/audio instructions when I am walking around?
  3. A paper back would be an awkward thing to carry around. What new types of pockets or jackets will emerge to fashionably carry this device (I am not looking at wearing the conventional jackets).
  4. Just like email, Facebook and Twitter made every one write more, will this new device make people sketch more? Think More?

Eager to see what Apple will bring out. Even if it does not do all the items on my list, Apple is capable of starting a new revolution in computing devices and some of the competitors may jump in and do some of these, hopefully. If not, please give us a cool device with all the device drivers and an app framework. We the geeks, will take care of the rest.

Posted via web from Dorai’s Tech Log