Now this is basic data manipulation- and I used Deducer for it.
The best thing is the ability to use GGPlot using a GUI.
I am now trying to create more complicated plots for example with more than one Y variable but it is still a work in progress. Overall Deducer has made impressive improvements and with the JGR GUI seems very very promising. The look and feel also shows a combination of features (from SPSS ‘s variable and data view)
And yes China overtook India in 1985. In GDP per capita. Sigh
GGPLot though overtook Excel graphics as well.
Here is a video which is much better than my screenshots
PALO ALTO, Calif., Sept. 20 — Revolution Analytics, the leading commercial provider of software and support for the popular open source R statistics language, today announced it will deliver Revolution R Enterprise for Microsoft Windows HPC Server 2008 R2, released today, enabling users to analyze very large data sets in high-performance computing environments.
R is a powerful open source statistics language and the modern system for predictive analytics. Revolution Analytics recently introduced RevoScaleR, new “Big Data” analysis capabilities, to its R distribution, Revolution R Enterprise. RevoScaleR solves the performance and capacity limitations of the R language by with parallelized algorithms that stream data across multiple cores on a laptop, workstation or server. Users can now process, visualize and model terabyte-class data sets at top speeds — without the need for specialized hardware.
“Revolution Analytics is pleased to support Microsoft’s Technical Computing initiative, whose efforts will benefit scientists, engineers and data analysts,” said David Champagne, CTO at Revolution. “We believe the engineering we have done for Revolution R Enterprise, in particular our work on big-data statistics and multicore computing, along with Microsoft’s HPC platform for technical computing, makes an ideal combination for high-performance large scale statistical computing.”
“Processing and analyzing this ‘big data’ is essential to better prediction and decision making,” said Bill Hamilton, director of technical computing at Microsoft Corp. “Revolution R Enterprise for Windows HPC Server 2008 R2 gives customers an extremely powerful tool that handles analysis of very large data and high workloads.”
REvolution R Enterprise is designed for both novice and experienced R users looking for a production-grade R distribution to perform mission critical predictive analytics tasks right from the desktop and scale across multiprocessor environments. Featuring RPE™ REvolution’s R Productivity Environment for Windows.
Of course R Enterprise is available on Linux but on Red Hat Enterprise Linux- it would be nice to see Amazom Machine Images as well as Ubuntu versions as well.
Like all virtual appliances, the main component of an AMI is a read-only filesystem image which includes an operating system (e.g., Linux, UNIX, or Windows) and any additional software required to deliver a service or a portion of it.[2]
The AMI filesystem is compressed, encrypted, signed, split into a series of 10MB chunks and uploaded into Amazon S3 for storage. An XML manifest file stores information about the AMI, including name, version, architecture, default kernel id, decryption key and digests for all of the filesystem chunks.
An AMI does not include a kernel image, only a pointer to the default kernel id, which can be chosen from an approved list of safe kernels maintained by Amazon and its partners (e.g., RedHat, Canonical, Microsoft). Users may choose kernels other than the default when booting an AMI.[3]
Paid: a for-pay AMI image that is registered with Amazon DevPay and can be used by any one who subscribes for it. DevPay allows developers to mark-up Amazon’s usage fees and optionally add monthly subscription fees.
JMP 9 releases on Oct 12- it is a very good reliable data visualization and analytical tool ( AND available on Mac as well)
AND IT is advertising R Graphics as well (lol- I can visualize the look on some ahem SAS fans in the R Project)
Updated Pricing- note I am not sure why they are charging US academics 495$ when SAS On Demand is free for academics. Shouldnt JMP be free to students- maybe John Sall and his people can do a tradeoff analysis for this given JMP’s graphics are better than Base SAS (which is under some pressure from WPS and R)
Be First in Line for JMP® 9
Save up to $300 when you pre-order a
single-user license by Oct. 11
Make JMP your analytic hub for visual data discovery with this special offer, good through Oct. 11, 2010. Pre-order a single-user license of JMP 9 – for a discount of up to $300 – and get ready for a leap in data interactivity.
Order now and enjoy the compelling new features of JMP 9 when the software is released Oct. 12. New capabilities in JMP 9 let you:
Optimize and simulate using your Microsoft Excel spreadsheets.
Use maps to find patterns in your geographic data.
Enjoy the updated look and flexibility of JMP 9 on Microsoft Windows.
Create and share custom add-ins that extend JMP.
Leverage an expanded array of advanced statistical methodologies.
Display analytic results from R using interactive graphics.
What if I already have a JMP 8 single-user license?
Great news! You can upgrade to JMP 9 for less than half the regular price.
What if I’m an annual license customer?
Don’t worry, we’ve got you covered. Annual license customers enjoy priority access to all the latest JMP releases as soon as they become available. JMP 9 will be shipped to you automatically.
What if I work or study in the academic world?
Call 1-877-594-6567 to learn about significant discounts for students and professors through the JMP Academic Program.
i spend more time with you
than with anything or anyone else
i could leave you behind
but you climb my lap and now have turned mobile
my fingers hurt and my eyes are red
inputting my stuff on you i go on and on instead
this is crazy not just done
no sooner do I finish writing that
I find I have just begun
for what separates the pretenders from the rest
is the actions not their words that make them the best
so my friend my computer and me
together we create
so much work to be done while the haters hate
news to be read, blogs to be done
code to be executed, and sometimes to be undone
email lists, and online games as well,
dreaming online heaven in offline hell
Words can be sublime so much can be told
My friend my computer and me- together we grow old.
His argument of love is not very original though it was first made by these four guys
I am going to argue that “some” R developers should be paid, while the main focus should be volunteers code. These R developers should be paid as per usage of their packages.
Let me expand.
Imagine the following conversation between Ross Ihaka, Norman Nie and Peter Dalgaard.
Norman- Hey Guys, Can you give me some code- I got this new startup.
Ross Ihaka and Peter Dalgaard- Sure dude. Here is 100,000 lines of code, 2000 packages and 2 decades of effort.
Norman- Thanks guys.
Ross Ihaka- Hey, What you gonna do with this code.
Norman- I will better it. Sell it. Finally beat Jim Goodnight and his **** Proc GLM and **** Proc Reg.
Ross- Okay, but what will you give us? Will you give us some code back of what you improve?
Norman – Uh, let me explain this open core …
Peter D- Well how about some royalty?
Norman- Sure, we will throw parties at all conferences, snacks you know at user groups.
Ross – Hmm. That does not sound fair. (walks away in a huff muttering)-He takes our code, sells it and wont share the code
Peter D- Doesnt sound fair. I am back to reading Hamlet, the great Dane, and writing the next edition of my book. I am glad I wrote a book- Ross didnt even write that.
Norman-Uh Oh. (picks his phone)- Hey David Smith, We need to write some blog articles pronto – these open source guys ,man…
———–I think that sums what has been going on in the dynamics of R recently. If Ross Ihaka and R Gentleman had adopted an open core strategy- meaning you can create packages to R but not share the original where would we all be?
At this point if he is reading this, David Smith , long suffering veteran of open source flameouts is rolling his eyes while Tal G is wondering if he will publish this on R Bloggers and if so when or something.
Lets bring in another R veteran- Hadley Wickham who wrote a book on R and also created ggplot. Thats the best quality, most often used graphics package.
In terms of economic utilty to end user- the ggplot package may be as useful if not more as the foreach package developed by Revolution Computing/Analytics.
However lets come to open core licensing ( read it here http://alampitt.typepad.com/lampitt_or_leave_it/2008/08/open-core-licen.html ) which is where the debate is- Revolution takes code- enhances it (in my opinion) substantially with new formats XDF for better efficieny, web services API, and soon coming next year a GUI (thanks in advance , Dr Nie and guys)
and sells this advanced R code to businesses happy to pay ( they are currently paying much more to DR Goodnight and HIS guys)
Why would any sane customer buy it from Revolution- if he could download exactly the same thing from http://r-project.org
Hence the business need for Revolution Analytics to have an enhanced R- as they are using a product based software model not software as a service model.
If Revolution gives away source code of these new enhanced codes to R core team- how will R core team protect the above mentioned intelectual property- given they have 2 decades experience of giving away free code , and back and forth on just code.
Now Revolution also has a marketing budget- and thats how they sponsor some R Core events, conferences, after conference snacks.
How would people decide if they are being too generous or too stingy in their contribution (compared to the formidable generosity of SAS Institute to its employees, stakeholders and even third party analysts).
Would it not be better- IF Revolution can shift that aspect of relationship to its Research and Development budget than it’s marketing budget- come with some sort of incentive for “SOME” developers – even researchers need grants and assistantships, scholarships, make a transparent royalty formula say 17.5 % of the NEW R sales goes to R PACKAGE Developers pool, which in turn examines usage rate of packages and need/merit before allocation- that would require Revolution to evolve from a startup to a more sophisticated corporate and R Core can use this the same way as John M Chambers software award/scholarship
Dont pay all developers- it would be an insult to many of them – say Prof Harrell creator of HMisc to accept – but can Revolution expand its dev base (and prospect for future employees) by even sponsoring some R Scholarships.
And I am sure that if Revolution opens up some more code to the community- they would the rest of the world and it’s help useful. If it cant trust people like R Gentleman with some source code – well he is a board member.
——————————————————————————————–
Now to sum up some technical discussions on NeW R
1) An accepted way of benchmarking efficiencies.
2) Code review and incorporation of efficiencies.
3) Multi threading- Multi core usage are trends to be incorporated.
4) GUIs like R Commander E Plugins for other packages, and Rattle for Data Mining to have focussed (or Deducer). This may involve hiring User Interface Designers (like from Apple 😉 who will work for love AND money ( Even the Beatles charge royalty for that song)
5) More support to cloud computing initiatives like Biocep and Elastic R – or Amazon AMI for using cloud computers- note efficiency arguements dont matter if you just use a Chrome Browser and pay 2 cents a hour for an Amazon Instance. Probably R core needs more direct involvement of Google (Cloud OS makers) and Amazon as well as even Salesforce.com (for creating Force.com Apps). Note even more corporates here need to be involved as cloud computing doesnot have any free and open source infrastructure (YET)
“If something goes wrong with Microsoft, I can phone Microsoft up and have it fixed. With Open Source, I have to rely on the community.”
And the community, as much as we may love it, is unpredictable. It might care about your problem and want to fix it, then again, it may not. Anyone who has ever witnessed something online go “viral”, good or bad, will know what I’m talking about.
Just checked out cool new series from NVidia servers.
Now though SAS Inc/ Jim Goodnight thinks HP Blade Servers are the cool thing- the GPU takes hardware high performance computing to another level. It would be interesting to see GPU based cloud computers as well – say for the on Demand SAS (free for academics and students) but which has had some complaints of being slow.
To give users hands-on experience, the program is underpinned by a virtual computing lab (VCL), a remote access service that allows users to reserve a computer configured with a desired set of applications and operating system and then access that computer over the Internet. The lab is powered by an IBM BladeCenter infrastructure, which includes more than 500 blade servers, distributed between two locations. The assignment of the blade servers can be changed to meet shifts in the balance of demand among the various groups of users. Laura Ladrie, MSA Classroom Coordinator and Technical Support Specialist, says, “The virtual computing lab chose IBM hardware because of its quality, reliability and performance. IBM hardware is also energy efficient and lends itself well to high performance/low overhead computing.
Thats interesting since IBM now competes (as owner of SPSS) and also cooperates with SAS Institute
You’re effectively turbo-charging through deployment of many processors within the blade servers?
Yes. We’ve got machines with 192 blades on them. One of them has 202 or 203 blades. We’re using Hewlett-Packard blades with 12 CP cores on each, so it’s a total 2300 CPU cores doing the computation.
Our idea was to give every one of those cores a little piece of work to do, and we came up with a solution. It involved a very small change to the algorithm we were using, and it’s just incredible how fast we can do things now.
I don’t think of it as a grid, I think of it as essentially one computer. Most people will take a blade and make a grid out of it, where everything’s a separate computer running separate jobs.
We just look at it as one big machine that has memory and processors all over the place, so it’s a totally different concept.
GPU servers can be faster than CPU servers, though , Professor G.
TESLA GPU COMPUTING SOLUTIONS FOR DATA CENTERS
Supercharge your cluster with the Tesla family of GPU computing solutions. Deploy 1U systems from NVIDIA or hybrid CPU-GPU servers from OEMs that integrate NVIDIA® Tesla™ GPU computing processors.
When compared to the latest quad-core CPU, Tesla 20-series GPU computing processors deliver equivalent performance at 1/20th the power consumption and 1/10th the cost. Each Tesla GPU features hundreds of parallel CUDA cores and is based on the revolutionary NVIDIA® CUDA™ parallel computing architecture with a rich set of developer tools (compilers, profilers, debuggers) for popular programming languages APIs like C, C++, Fortran, and driver APIs like OpenCL and DirectCompute.
NVIDIA’s partners provide turnkey easy-to-deploy Preconfigured Tesla GPU clusters that are customizable to your needs. For 3D cloud computing applications, our partners offer the Tesla RS clusters that are optimized for running RealityServer with iray.
Usage of accelerated BLAS libraries seems to shrouded in some mystery, judging from somewhat regularly recurring requests for help on lists such as r-sig-hpc(gmane version), the R list dedicated to High-Performance Computing. Yet it doesn’t have to be; installation can be really simple (on appropriate systems).
Another issue that I felt needed addressing was a comparison between the different alternatives available, quite possibly including GPU computing. So a few weeks ago I sat down and wrote a small package to run, collect, analyse and visualize some benchmarks. That package, called gcbd (more about the name below) is now onCRAN as of this morning. The package both facilitates the data collection for the paper it also contains (in the vignette form common among R packages) and provides code to analyse the data—which is also included as a SQLite database. All this is done in the Debian and Ubuntu context by transparently installing and removing suitable packages providing BLAS implementations: that we can fully automate data collection over several competing implementations via a single script (which is also included). Contributions of benchmark results is encouraged—that is the idea of the package.
And from his paper on the same-
Analysts are often eager to reap the maximum performance from their computing platforms.
A popular suggestion in recent years has been to consider optimised basic linear algebra subprograms (BLAS). Optimised BLAS libraries have been included with some (commercial) analysis platforms for a decade (Moler 2000), and have also been available for (at least some) Linux distributions for an equally long time (Maguire 1999). Setting BLAS up can be daunting: the R language and environment devotes a detailed discussion to the topic in its Installation and Administration manual (R Development Core Team 2010b, appendix A.3.1). Among the available BLAS implementations, several popular choices have emerged. Atlas (an acronym for Automatically Tuned Linear Algebra System) is popular as it has shown very good performance due to its automated and CPU-speci c tuning (Whaley and Dongarra 1999; Whaley and Petitet 2005). It is also licensed in such a way that it permits redistribution leading to fairly wide availability of Atlas.1 We deploy Atlas in both a single-threaded and a multi-threaded con guration. Another popular BLAS implementation is Goto BLAS which is named after its main developer, Kazushige Goto (Goto and Van De Geijn 2008). While `free to use’, its license does not permit redistribution putting the onus of con guration, compilation and installation on the end-user. Lastly, the Intel Math Kernel Library (MKL), a commercial product, also includes an optimised BLAS library. A recent addition to the tool chain of high-performance computing are graphical processing units (GPUs). Originally designed for optimised single-precision arithmetic to accelerate computing as performed by graphics cards, these devices are increasingly used in numerical analysis. Earlier criticism of insucient floating-point precision or severe performance penalties for double-precision calculation are being addressed by the newest models. Dependence on particular vendors remains a concern with NVidia’s CUDA toolkit (NVidia 2010) currently still the preferred development choice whereas the newer OpenCL standard (Khronos Group 2008) may become a more generic alternative that is independent of hardware vendors. Brodtkorb et al. (2010) provide an excellent recent survey. But what has been lacking is a comparison of the e ective performance of these alternatives. This paper works towards answering this question. By analysing performance across ve di erent BLAS implementations|as well as a GPU-based solution|we are able to provide a reasonably broad comparison.
Performance is measured as an end-user would experience it: we record computing times from launching commands in the interactive R environment (R Development Core Team 2010a) to their completion.
And
Basic Linear Algebra Subprograms (BLAS) provide an Application Programming Interface
(API) for linear algebra. For a given task such as, say, a multiplication of two conformant
matrices, an interface is described via a function declaration, in this case sgemm for single
precision and dgemm for double precision. The actual implementation becomes interchangeable
thanks to the API de nition and can be supplied by di erent approaches or algorithms. This
is one of the fundamental code design features we are using here to benchmark the di erence
in performance from di erent implementations.
A second key aspect is the di erence between static and shared linking. In static linking,
object code is taken from the underlying library and copied into the resulting executable.
This has several key implications. First, the executable becomes larger due to the copy of
the binary code. Second, it makes it marginally faster as the library code is present and
no additional look-up and subsequent redirection has to be performed. The actual amount
of this performance penalty is the subject of near-endless debate. We should also note that
this usually amounts to only a small load-time penalty combined with a function pointer
redirection|the actual computation e ort is unchanged as the actual object code is identi-
cal. Third, it makes the program more robust as fewer external dependencies are required.
However, this last point also has a downside: no changes in the underlying library will be
reected in the binary unless a new build is executed. Shared library builds, on the other
hand, result in smaller binaries that may run marginally slower|but which can make use of
di erent libraries without a rebuild.
Basic Linear Algebra Subprograms (BLAS) provide an Application Programming Interface(API) for linear algebra. For a given task such as, say, a multiplication of two conformantmatrices, an interface is described via a function declaration, in this case sgemm for singleprecision and dgemm for double precision. The actual implementation becomes interchangeablethanks to the API de nition and can be supplied by di erent approaches or algorithms. Thisis one of the fundamental code design features we are using here to benchmark the di erencein performance from di erent implementations.A second key aspect is the di erence between static and shared linking. In static linking,object code is taken from the underlying library and copied into the resulting executable.This has several key implications. First, the executable becomes larger due to the copy ofthe binary code. Second, it makes it marginally faster as the library code is present andno additional look-up and subsequent redirection has to be performed. The actual amountof this performance penalty is the subject of near-endless debate. We should also note thatthis usually amounts to only a small load-time penalty combined with a function pointerredirection|the actual computation e ort is unchanged as the actual object code is identi-cal. Third, it makes the program more robust as fewer external dependencies are required.However, this last point also has a downside: no changes in the underlying library will bereected in the binary unless a new build is executed. Shared library builds, on the otherhand, result in smaller binaries that may run marginally slower|but which can make use ofdi erent libraries without a rebuild.
And summing up,
reference BLAS to be dominated in all cases. Single-threaded Atlas BLAS improves on the reference BLAS but loses to multi-threaded BLAS. For multi-threaded BLAS we nd the Goto BLAS dominate the Intel MKL, with a single exception of the QR decomposition on the xeon-based system which may reveal an error. The development version of Atlas, when compiled in multi-threaded mode is competitive with both Goto BLAS and the MKL. GPU computing is found to be compelling only for very large matrix sizes. Our benchmarking framework in the gcbd package can be employed by others through the R packaging system which could lead to a wider set of benchmark results. These results could be helpful for next-generation systems which may need to make heuristic choices about when to compute on the CPU and when to compute on the GPU.
Hardware solutions or atleast need to be a part of Revolution Analytic’s thinking as well. SPSS does not have any choice anymore though 😉
It would be interesting to see how the new SAS Cloud Computing/ Server Farm/ Time Sharing facility is benchmarking CPU and GPU for SAS analytics performance – if being done already it would be nice to see a SUGI paper on the same at http://sascommunity.org.
Multi threading needs to be taken care automatically by statistical software to optimize current local computing (including for New R)
Acceptable benchmarks for testing hardware as well as software need to be reinforced and published across vendors, academics and companies.
DECISION STATS 
