Search Results

Search found 347 results on 14 pages for 'academic'.

Page 3/14 | < Previous Page | 1 2 3 4 5 6 7 8 9 10 11 12  | Next Page >

  • Why is Software Engineering not the typical major for future software developers?

    - by FarmBoy
    While most agree that a certain level of Computer Science is essential to being a good programmer, it seems to me that the principles of good software development is even more important, though not as fundamental. Just like mechanical engineers take physics classes, but far more engineering classes, I would expect, now that software is over a half century old, that software development would begin to dominate the undergraduate curriculum. But I don't see much evidence of this. Is there a reason that Software Engineering hasn't taken hold as an academic discipline?

    Read the article

  • Why (not) logic programming?

    - by Anto
    I have not yet heard about any uses of a logical programming language (such as Prolog) in the software industry, nor do I know of usage of it in hobby programming or open source projects. It (Prolog) is used as an academic language to some extent, though (why is it used in academia?). This makes me wonder, why should you use logic programming, and why not? Why is it not getting any detectable industry usage?

    Read the article

  • Book Review: Brownfield Application Development in .NET

    - by DotNetBlues
    I recently finished reading the book Brownfield Application Development in .NET by Kyle Baley and Donald Belcham.  The book is available from Manning.  First off, let me say that I'm a huge fan of Manning as a publisher.  I've found their books to be top-quality, over all.  As a Kindle owner, I also appreciate getting an ebook copy along with the dead tree copy.  I find ebooks to be much more convenient to read, but hard-copies are easier to reference. The book covers, surprisingly enough, working with brownfield applications.  Which is well and good, if that term has meaning to you.  It didn't for me.  Without retreading a chunk of the first chapter, the authors break code bases into three broad categories: greenfield, brownfield, and legacy.  Greenfield is, essentially, new development that hasn't had time to rust and is (hopefully) being approached with some discipline.  Legacy applications are those that are more or less stable and functional, that do not expect to see a lot of work done to them, and are more likely to be replaced than reworked. Brownfield code is the gray (brown?) area between the two and the authors argue, quite effectively, that it is the most likely state for an application to be in.  Brownfield code has, in some way, been allowed to tarnish around the edges and can be difficult to work with.  Although I hadn't realized it, most of the code I've worked on has been brownfield.  Sometimes, there's talk of scrapping and starting over.  Sometimes, the team dismisses increased discipline as ivory tower nonsense.  And, sometimes, I've been the ignorant culprit vexing my future self. The book is broken into two major sections, plus an introduction chapter and an appendix.  The first section covers what the authors refer to as "The Ecosystem" which consists of version control, build and integration, testing, metrics, and defect management.  The second section is on actually writing code for brownfield applications and discusses object-oriented principles, architecture, external dependencies, and, of course, how to deal with these when coming into an existing code base. The ecosystem section is just shy of 140 pages long and brings some real meat to the matter.  The focus on "pain points" immediately sets the tone as problem-solution, rather than academic.  The authors also approach some of the topics from a different angle than some essays I've read on similar topics.  For example, the chapter on automated testing is on just that -- automated testing.  It's all well and good to criticize a project as conflating integration tests with unit tests, but it really doesn't make anyone's life better.  The discussion on testing is more focused on the "right" level of testing for existing projects.  Sometimes, an integration test is the best you can do without gutting a section of functional code.  Even if you can sell other developers and/or management on doing so, it doesn't actually provide benefit to your customers to rewrite code that works.  This isn't to say the authors encourage sloppy coding.  Far from it.  Just that they point out the wisdom of ignoring the sleeping bear until after you deal with the snarling wolf. The other sections take a similarly real-world, workable approach to the pain points they address.  As the section moves from technical solutions like version control and continuous integration (CI) to the softer, process issues of metrics and defect tracking, the authors begin to gently suggest moving toward a zero defect count.  While that really sounds like an unreasonable goal for a lot of ongoing projects, it's quite apparent that the authors have first-hand experience with taming some gruesome projects.  The suggestions are grounded and workable, and the difficulty of some situations is explicitly acknowledged. I have to admit that I started getting bored by the end of the ecosystem section.  No matter how valuable I think a good project manager or business analyst is to a successful ALM, at the end of the day, I'm a gear-head.  Also, while I agreed with a lot of the ecosystem ideas, in theory, I didn't necessarily feel that a lot of the single-developer projects that I'm often involved in really needed that level of rigor.  It's only after reading the sidebars and commentary in the coding section that I had the context for the arguments made in favor of a strong ecosystem supporting the development process.  That isn't to say that I didn't support good product management -- indeed, I've probably pushed too hard, on occasion, for a strong ALM outside of just development.  This book gave me deeper insight into why some corners shouldn't be cut and how damaging certain sins of omission can be. The code section, though, kept me engaged for its entirety.  Many technical books can be used as reference material from day one.  The authors were clear, however, that this book is not one of these.  The first chapter of the section (chapter seven, over all) addresses object oriented (OO) practices.  I've read any number of definitions, discussions, and treatises on OO.  None of the chapter was new to me, but it was a good review, and I'm of the opinion that it's good to review the foundations of what you do, from time to time, so I didn't mind. The remainder of the book is really just about how to apply OOP to existing code -- and, just because all your code exists in classes does not mean that it's object oriented.  That topic has the potential to be extremely condescending, but the authors miraculously managed to never once make me feel like a dolt or that they were wagging their finger at me for my prior sins.  Instead, they continue the "pain points" and problem-solution presentation to give concrete examples of how to apply some pretty academic-sounding ideas.  That's a point worth emphasizing, as my experience with most OO discussions is that they stay in the academic realm.  This book gives some very, very good explanations of why things like the Liskov Substitution Principle exist and why a corporate programmer should even care.  Even if you know, with absolute certainty, that you'll never have to work on an existing code-base, I would recommend this book just for the clarity it provides on OOP. This book goes beyond just theory, or even real-world application.  It presents some methods for fixing problems that any developer can, and probably will, encounter in the wild.  First, the authors address refactoring application layers and internal dependencies.  Then, they take you through those layers from the UI to the data access layer and external dependencies.  Finally, they come full circle to tie it all back to the overall process.  By the time the book is done, you're left with a lot of ideas, but also a reasonable plan to begin to improve an existing project structure. Throughout the book, it's apparent that the authors have their own preferred methodology (TDD and domain-driven design), as well as some preferred tools.  The "Our .NET Toolbox" is something of a neon sign pointing to that latter point.  They do not beat the reader over the head with anything resembling a "One True Way" mentality.  Even for the most emphatic points, the tone is quite congenial and helpful.  With some of the near-theological divides that exist within the tech community, I found this to be one of the more remarkable characteristics of the book.  Although the authors favor tools that might be considered Alt.NET, there is no reason the advice and techniques given couldn't be quite successful in a pure Microsoft shop with Team Foundation Server.  For that matter, even though the book specifically addresses .NET, it could be applied to a Java and Oracle shop, as well.

    Read the article

  • You're invited : Oracle Solaris Forum, June 19th, Petah Tikva

    - by Frederic Pariente
    The local ISV Engineering will be attending and speaking at the Oracle and ilOUG Solaris Forum next week in Israel. Come meet us there! This free event requires registration, thanks. YOU'RE INVITED Oracle Solaris Forum Date : Tuesday, June 19th, 2012 Time : 14:00 Location :  Dan Academic CenterPetach TikvaIsrael Agenda : Enterprise Manager OPS Center and Oracle Exalogic Elastic CloudSolaris 11NetworkingCustomer Case Study : BMCOpen Systems Curriculum See you there!

    Read the article

  • You're invited : Oracle Solaris Forum, Dec 18th, Petah Tikva

    - by Frederic Pariente
    The local ISV Engineering will be attending and speaking at the Oracle and ilOUG Solaris Forum next week in Israel. Come meet us there! This free event requires registration, thanks. YOU'RE INVITED Oracle Solaris Forum Date : Tuesday, December 18th, 2012 Time : 14:00 Location :  Dan Academic CenterPetach TikvaIsrael Agenda : New Features in Solaris 11.1SPARC T4 & T5Solaris 11 Serviceability See you there!

    Read the article

  • Reporting on common code smells : A POC

    - by Dave Ballantyne
    Over the past few blog entries, I’ve been looking at parsing TSQL scripts in a variety of ways for a variety of tasks.  In my last entry ‘How to prevent ‘Select *’ : The elegant way’, I looked at parsing SQL to report upon uses of SELECT *.  The obvious question leading on from this is, “Great, what about other code smells ?”  Well, using the language service parser to do that was turning out to be a bit of a hard job,  sure I was getting tokens but no real context.  I wasn't even being told when an end of statement had been reached. One of the other parsing options available from Microsoft is exposed in the assembly ‘Microsoft.SqlServer.TransactSql.ScriptDom’,  this is ,I believe, installed with the client development tools with SQLServer.  It is much more feature rich than the original parser I had used and breaks a TSQL script into intuitive classes for analysis. So, what sort of smells can I now find using it ?  Well, for an opening gambit quite a nice little list. Use of NOLOCK Set of READ UNCOMMITTED Use of SELECT * Insert without column references Explicit datatype conversion on Sargs Cross server selects Non use of two-part naming convention Table and Query hint usage Changes in set options Use of single line comments Use of ordinal column positions in ORDER BY clause Now, lets not argue the point that “It depends” as smells on some of these, but as an academic exercise it is quite interesting.  The code is available from this link :https://www.dropbox.com/s/rfk32sou4fzl2cw/TSQLDomTest.zip  All the usual disclaimers apply to this code, I cannot be held responsible for anything ranging from mild annoyance through to universe destruction due to the use of this code or examples. The zip file contains a powershell script and my test cases.  The assembly used requires .Net 4 to run, which means that you will need powershell 3 ( though im running through PowerGUI and all works ok ) .  The code searches for all .sql files in the folder hierarchy for the workingpath,  you can override this if you want by simply changing the $Folder variable, and processes each in turn for the smells.  Feedback is not great at the moment, all it does is output to an xml file (Smells.xml) the offset position and a description of the smell found. Right now, I am interested in your feedback.  What do you think ?  Is this (or should it be) more than an academic exercise ?  Can tooling such as this be used as some form of code quality measure ?  Does it Work ? Do you have a case listed above which is not being reported ? Do you have a case that you would love to be reported ? Let me know , please mailto: [email protected]. Thanks

    Read the article

  • What are the essential things one needs to know about UML?

    - by Hanno Fietz
    I want my scribbles of a program's design and behaviour to become more streamlined and have a common language with other developers. I looked at UML and in principle it seems to be what I'm looking for, but it seems to be overkill. The information I found online also seems very bloated and academic. How can I understand UML in plain-English way, enough to be able to explain it to my colleagues? What are the canonical resources for understanding UML at a ground level?

    Read the article

  • What's the most useful 10% of UML and is there a quick tutorial on it?

    - by Hanno Fietz
    I want my scribbles of a program's design and behaviour to become more streamlined and have a common language with other developers. I looked at UML and in principle it seems to be what I'm looking for, just way overkill. The information I found online also seems very bloated and academic. Is there a no-bullshit, 15-minutes introduction to the handful of UML symbols I'll need when discussing the architecture of some garden variety software on a whiteboard with my colleagues?

    Read the article

  • is there a formal algebra method to analyze programs?

    - by Gabriel
    Is there a formal/academic connection between an imperative program and algebra, and if so where would I learn about it? The example I'm thinking of is: if(C1) { A1(); A2(); } if(C2) { A1(); A2(); } Represented as a sum of terms: (C1)(A1) + (C1)(A2) + (C2)(A1) + (C2)(A2) = (C1+C2)(A1+A2) The idea being that manipulation could lead to programatic refactoring - "factoring" being the common concept in this example.

    Read the article

  • how can i allow user to create posts in website using ASP.NET [closed]

    - by Sana
    I am making a website "Online voting system" a part of my academic project ... besides allowing the registered voters to vote on the posts that I have created while developing the voting system ... I want to allow users to create polls by themselves too regarding any thing How can I implement this scenerio using ASP.NET and c# in VS 2010 .. with the user posting polls having: Post title label Description about poll Radio buttons for displaying various options to allow voting process to be carried out when user selects one option and submit his vote...

    Read the article

  • Download Singularity Source Code

    - by Editor
    The Singularity Research Development Kit (RDK) is based on the Microsoft Research Singularity project. It includes source code, build tools, test suites, design notes, and other background materials. The Singularity RDK is for academic non-commercial use and is governed by this license. Singularity is a research project focused on the construction of dependable [...]

    Read the article

  • Strangling the life out of Software Testing

    - by MarkPearl
    I recently did a course at the local university on Software Engineering. At the beginning of the course I looked over the outline of the subject and there seemed to be some really good content. It covered traditional & agile project methodologies, some general communication and modelling chapters and finished off with testing. I was particularly excited to see the section on testing as this was something I learnt on my own and see great value in. The course has now just ended and I am very disappointed. I now know one of the reasons why so few people i.e. in my region do Test Driven Development, or perform even basic testing methodologies. The topic was to academic! Yes, you might be able to list 4 different types of black box test approaches vs. white box test approaches and describe the characteristics of Smoke Tests, but never during course did we see an example of an actual test or how it might be implemented! In fact, if I did not have personal experience of applying testing in actual projects, I wouldn’t even know what a unit test looked like. Now, what worries me is the following… It took us 6 months to cover the course material, other students more than likely came out of that course with little appreciation of the subject – in fact they now have a very complex view of what a test is – so complex that I think most of them will never attempt it again on their own. Secondly, imagine studying to be a dentist without ever actually seeing a tooth? Yes, you might be able to describe a tooth, and know what it is made out of – but nobody would want a dentist who has never seen a tooth to operate on them. Yet somehow we expect people studying software engineering to do the same? This is not right. Now, before I finish my rant let me say that I know this is not the same everywhere in the world, and that there needs to be a balance on practical implementation and academic understanding – I am just disappointed that this does not seem to be happening at the institution that I am currently studying at ;-( Please, if you happen to be a lecturer or teacher reading this post – a combination of theory and practical's goes a long way. We need to up the quality of software being produced and that starts at learner level!

    Read the article

  • Cisco ASA 5510 Time of Day Based Policing

    - by minamhere
    I have a Cisco ASA 5510 setup at a boarding school. I determined that many (most?) of the students were downloading files, watching movies, etc, during the day and this was causing the academic side of our network to suffer. The students should not even be in their rooms during the day, so I configured the ASA to police their network segment and limit their outbound bandwidth. This resolved all of our academic issues, and everyone was happy. Except the resident students. I have been asked to change/remove the policing policy at the end of the day, to allow the residents access to the unused bandwidth at night. There's no reason to let bandwidth sit unused at night just because it would be abused during the day. Is there a way to setup Time of Day based Policies on the ASA? Ideally I'd like to be able to open up the network at night and all day during weekends. If I can't set Time based policies, is is possible to schedule the ASA to load a set of commands at a specific time? I suppose I could just setup a scheduled task on one of our servers to log in and make the changes with a simple script, but this seems like a hack, and I'm hoping there is a better or more standard way to accomplish this. Thanks. Edit: If there is a totally different solution that would accomplish a similar goal, I'd be interested in that as well. Free/Cheap would be ideal, but if a separate internet connection is my only other option, it might be worth fighting for money for hardware or software to do this better or more efficiently.

    Read the article

  • Software development is (mostly) a trade, and what to do about it

    - by Jeff
    (This is another cross-post from my personal blog. I don’t even remember when I first started to write it, but I feel like my opinion is well enough baked to share.) I've been sitting on this for a long time, particularly as my opinion has changed dramatically over the last few years. That I've encountered more crappy code than maintainable, quality code in my career as a software developer only reinforces what I'm about to say. Software development is just a trade for most, and not a huge academic endeavor. For those of you with computer science degrees readying your pitchforks and collecting your algorithm interview questions, let me explain. This is not an assault on your way of life, and if you've been around, you know I'm right about the quality problem. You also know the HR problem is very real, or we wouldn't be paying top dollar for mediocre developers and importing people from all over the world to fill the jobs we can't fill. I'm going to try and outline what I see as some of the problems, and hopefully offer my views on how to address them. The recruiting problem I think a lot of companies are doing it wrong. Over the years, I've had two kinds of interview experiences. The first, and right, kind of experience involves talking about real life achievements, followed by some variation on white boarding in pseudo-code, drafting some basic system architecture, or even sitting down at a comprooder and pecking out some basic code to tackle a real problem. I can honestly say that I've had a job offer for every interview like this, save for one, because the task was to debug something and they didn't like me asking where to look ("everyone else in the company died in a plane crash"). The other interview experience, the wrong one, involves the classic torture test designed to make the candidate feel stupid and do things they never have, and never will do in their job. First they will question you about obscure academic material you've never seen, or don't care to remember. Then they'll ask you to white board some ridiculous algorithm involving prime numbers or some kind of string manipulation no one would ever do. In fact, if you had to do something like this, you'd Google for a solution instead of waste time on a solved problem. Some will tell you that the academic gauntlet interview is useful to see how people respond to pressure, how they engage in complex logic, etc. That might be true, unless of course you have someone who brushed up on the solutions to the silly puzzles, and they're playing you. But here's the real reason why the second experience is wrong: You're evaluating for things that aren't the job. These might have been useful tactics when you had to hire people to write machine language or C++, but in a world dominated by managed code in C#, or Java, people aren't managing memory or trying to be smarter than the compilers. They're using well known design patterns and techniques to deliver software. More to the point, these puzzle gauntlets don't evaluate things that really matter. They don't get into code design, issues of loose coupling and testability, knowledge of the basics around HTTP, or anything else that relates to building supportable and maintainable software. The first situation, involving real life problems, gives you an immediate idea of how the candidate will work out. One of my favorite experiences as an interviewee was with a guy who literally brought his work from that day and asked me how to deal with his problem. I had to demonstrate how I would design a class, make sure the unit testing coverage was solid, etc. I worked at that company for two years. So stop looking for algorithm puzzle crunchers, because a guy who can crush a Fibonacci sequence might also be a guy who writes a class with 5,000 lines of untestable code. Fashion your interview process on ways to reveal a developer who can write supportable and maintainable code. I would even go so far as to let them use the Google. If they want to cut-and-paste code, pass on them, but if they're looking for context or straight class references, hire them, because they're going to be life-long learners. The contractor problem I doubt anyone has ever worked in a place where contractors weren't used. The use of contractors seems like an obvious way to control costs. You can hire someone for just as long as you need them and then let them go. You can even give them the work that no one else wants to do. In practice, most places I've worked have retained and budgeted for the contractor year-round, meaning that the $90+ per hour they're paying (of which half goes to the person) would have been better spent on a full-time person with a $100k salary and benefits. But it's not even the cost that is an issue. It's the quality of work delivered. The accountability of a contractor is totally transient. They only need to deliver for as long as you keep them around, and chances are they'll never again touch the code. There's no incentive for them to get things right, there's little incentive to understand your system or learn anything. At the risk of making an unfair generalization, craftsmanship doesn't matter to most contractors. The education problem I don't know what they teach in college CS courses. I've believed for most of my adult life that a college degree was an essential part of being successful. Of course I would hold that bias, since I did it, and have the paper to show for it in a box somewhere in the basement. My first clue that maybe this wasn't a fully qualified opinion comes from the fact that I double-majored in journalism and radio/TV, not computer science. Eventually I worked with people who skipped college entirely, many of them at Microsoft. Then I worked with people who had a masters degree who sucked at writing code, next to the high school diploma types that rock it every day. I still think there's a lot to be said for the social development of someone who has the on-campus experience, but for software developers, college might not matter. As I mentioned before, most of us are not writing compilers, and we never will. It's actually surprising to find how many people are self-taught in the art of software development, and that should reveal some interesting truths about how we learn. The first truth is that we learn largely out of necessity. There's something that we want to achieve, so we do what I call just-in-time learning to meet those goals. We acquire knowledge when we need it. So what about the gaps in our knowledge? That's where the most valuable education occurs, via our mentors. They're the people we work next to and the people who write blogs. They are critical to our professional development. They don't need to be an encyclopedia of jargon, but they understand the craft. Even at this stage of my career, I probably can't tell you what SOLID stands for, but you can bet that I practice the principles behind that acronym every day. That comes from experience, augmented by my peers. I'm hell bent on passing that experience to others. Process issues If you're a manager type and don't do much in the way of writing code these days (shame on you for not messing around at least), then your job is to isolate your tradespeople from nonsense, while bringing your business into the realm of modern software development. That doesn't mean you slap up a white board with sticky notes and start calling yourself agile, it means getting all of your stakeholders to understand that frequent delivery of quality software is the best way to deal with change and evolving expectations. It also means that you have to play technical overlord to make sure the education and quality issues are dealt with. That's why I make the crack about sticky notes, because without the right technique being practiced among your code monkeys, you're just a guy with sticky notes. You're asking your business to accept frequent and iterative delivery, now make sure that the folks writing the code can handle the same thing. This means unit testing, the right instrumentation, integration tests, automated builds and deployments... all of the stuff that makes it easy to see when change breaks stuff. The prognosis I strongly believe that education is the most important part of what we do. I'm encouraged by things like The Starter League, and it's the kind of thing I'd love to see more of. I would go as far as to say I'd love to start something like this internally at an existing company. Most of all though, I can't emphasize enough how important it is that we mentor each other and share our knowledge. If you have people on your staff who don't want to learn, fire them. Seriously, get rid of them. A few months working with someone really good, who understands the craftsmanship required to build supportable and maintainable code, will change that person forever and increase their value immeasurably.

    Read the article

  • How John Got 15x Improvement Without Really Trying

    - by rchrd
    The following article was published on a Sun Microsystems website a number of years ago by John Feo. It is still useful and worth preserving. So I'm republishing it here.  How I Got 15x Improvement Without Really Trying John Feo, Sun Microsystems Taking ten "personal" program codes used in scientific and engineering research, the author was able to get from 2 to 15 times performance improvement easily by applying some simple general optimization techniques. Introduction Scientific research based on computer simulation depends on the simulation for advancement. The research can advance only as fast as the computational codes can execute. The codes' efficiency determines both the rate and quality of results. In the same amount of time, a faster program can generate more results and can carry out a more detailed simulation of physical phenomena than a slower program. Highly optimized programs help science advance quickly and insure that monies supporting scientific research are used as effectively as possible. Scientific computer codes divide into three broad categories: ISV, community, and personal. ISV codes are large, mature production codes developed and sold commercially. The codes improve slowly over time both in methods and capabilities, and they are well tuned for most vendor platforms. Since the codes are mature and complex, there are few opportunities to improve their performance solely through code optimization. Improvements of 10% to 15% are typical. Examples of ISV codes are DYNA3D, Gaussian, and Nastran. Community codes are non-commercial production codes used by a particular research field. Generally, they are developed and distributed by a single academic or research institution with assistance from the community. Most users just run the codes, but some develop new methods and extensions that feed back into the general release. The codes are available on most vendor platforms. Since these codes are younger than ISV codes, there are more opportunities to optimize the source code. Improvements of 50% are not unusual. Examples of community codes are AMBER, CHARM, BLAST, and FASTA. Personal codes are those written by single users or small research groups for their own use. These codes are not distributed, but may be passed from professor-to-student or student-to-student over several years. They form the primordial ocean of applications from which community and ISV codes emerge. Government research grants pay for the development of most personal codes. This paper reports on the nature and performance of this class of codes. Over the last year, I have looked at over two dozen personal codes from more than a dozen research institutions. The codes cover a variety of scientific fields, including astronomy, atmospheric sciences, bioinformatics, biology, chemistry, geology, and physics. The sources range from a few hundred lines to more than ten thousand lines, and are written in Fortran, Fortran 90, C, and C++. For the most part, the codes are modular, documented, and written in a clear, straightforward manner. They do not use complex language features, advanced data structures, programming tricks, or libraries. I had little trouble understanding what the codes did or how data structures were used. Most came with a makefile. Surprisingly, only one of the applications is parallel. All developers have access to parallel machines, so availability is not an issue. Several tried to parallelize their applications, but stopped after encountering difficulties. Lack of education and a perception that parallelism is difficult prevented most from trying. I parallelized several of the codes using OpenMP, and did not judge any of the codes as difficult to parallelize. Even more surprising than the lack of parallelism is the inefficiency of the codes. I was able to get large improvements in performance in a matter of a few days applying simple optimization techniques. Table 1 lists ten representative codes [names and affiliation are omitted to preserve anonymity]. Improvements on one processor range from 2x to 15.5x with a simple average of 4.75x. I did not use sophisticated performance tools or drill deep into the program's execution character as one would do when tuning ISV or community codes. Using only a profiler and source line timers, I identified inefficient sections of code and improved their performance by inspection. The changes were at a high level. I am sure there is another factor of 2 or 3 in each code, and more if the codes are parallelized. The study’s results show that personal scientific codes are running many times slower than they should and that the problem is pervasive. Computational scientists are not sloppy programmers; however, few are trained in the art of computer programming or code optimization. I found that most have a working knowledge of some programming language and standard software engineering practices; but they do not know, or think about, how to make their programs run faster. They simply do not know the standard techniques used to make codes run faster. In fact, they do not even perceive that such techniques exist. The case studies described in this paper show that applying simple, well known techniques can significantly increase the performance of personal codes. It is important that the scientific community and the Government agencies that support scientific research find ways to better educate academic scientific programmers. The inefficiency of their codes is so bad that it is retarding both the quality and progress of scientific research. # cacheperformance redundantoperations loopstructures performanceimprovement 1 x x 15.5 2 x 2.8 3 x x 2.5 4 x 2.1 5 x x 2.0 6 x 5.0 7 x 5.8 8 x 6.3 9 2.2 10 x x 3.3 Table 1 — Area of improvement and performance gains of 10 codes The remainder of the paper is organized as follows: sections 2, 3, and 4 discuss the three most common sources of inefficiencies in the codes studied. These are cache performance, redundant operations, and loop structures. Each section includes several examples. The last section summaries the work and suggests a possible solution to the issues raised. Optimizing cache performance Commodity microprocessor systems use caches to increase memory bandwidth and reduce memory latencies. Typical latencies from processor to L1, L2, local, and remote memory are 3, 10, 50, and 200 cycles, respectively. Moreover, bandwidth falls off dramatically as memory distances increase. Programs that do not use cache effectively run many times slower than programs that do. When optimizing for cache, the biggest performance gains are achieved by accessing data in cache order and reusing data to amortize the overhead of cache misses. Secondary considerations are prefetching, associativity, and replacement; however, the understanding and analysis required to optimize for the latter are probably beyond the capabilities of the non-expert. Much can be gained simply by accessing data in the correct order and maximizing data reuse. 6 out of the 10 codes studied here benefited from such high level optimizations. Array Accesses The most important cache optimization is the most basic: accessing Fortran array elements in column order and C array elements in row order. Four of the ten codes—1, 2, 4, and 10—got it wrong. Compilers will restructure nested loops to optimize cache performance, but may not do so if the loop structure is too complex, or the loop body includes conditionals, complex addressing, or function calls. In code 1, the compiler failed to invert a key loop because of complex addressing do I = 0, 1010, delta_x IM = I - delta_x IP = I + delta_x do J = 5, 995, delta_x JM = J - delta_x JP = J + delta_x T1 = CA1(IP, J) + CA1(I, JP) T2 = CA1(IM, J) + CA1(I, JM) S1 = T1 + T2 - 4 * CA1(I, J) CA(I, J) = CA1(I, J) + D * S1 end do end do In code 2, the culprit is conditionals do I = 1, N do J = 1, N If (IFLAG(I,J) .EQ. 0) then T1 = Value(I, J-1) T2 = Value(I-1, J) T3 = Value(I, J) T4 = Value(I+1, J) T5 = Value(I, J+1) Value(I,J) = 0.25 * (T1 + T2 + T5 + T4) Delta = ABS(T3 - Value(I,J)) If (Delta .GT. MaxDelta) MaxDelta = Delta endif enddo enddo I fixed both programs by inverting the loops by hand. Code 10 has three-dimensional arrays and triply nested loops. The structure of the most computationally intensive loops is too complex to invert automatically or by hand. The only practical solution is to transpose the arrays so that the dimension accessed by the innermost loop is in cache order. The arrays can be transposed at construction or prior to entering a computationally intensive section of code. The former requires all array references to be modified, while the latter is cost effective only if the cost of the transpose is amortized over many accesses. I used the second approach to optimize code 10. Code 5 has four-dimensional arrays and loops are nested four deep. For all of the reasons cited above the compiler is not able to restructure three key loops. Assume C arrays and let the four dimensions of the arrays be i, j, k, and l. In the original code, the index structure of the three loops is L1: for i L2: for i L3: for i for l for l for j for k for j for k for j for k for l So only L3 accesses array elements in cache order. L1 is a very complex loop—much too complex to invert. I brought the loop into cache alignment by transposing the second and fourth dimensions of the arrays. Since the code uses a macro to compute all array indexes, I effected the transpose at construction and changed the macro appropriately. The dimensions of the new arrays are now: i, l, k, and j. L3 is a simple loop and easily inverted. L2 has a loop-carried scalar dependence in k. By promoting the scalar name that carries the dependence to an array, I was able to invert the third and fourth subloops aligning the loop with cache. Code 5 is by far the most difficult of the four codes to optimize for array accesses; but the knowledge required to fix the problems is no more than that required for the other codes. I would judge this code at the limits of, but not beyond, the capabilities of appropriately trained computational scientists. Array Strides When a cache miss occurs, a line (64 bytes) rather than just one word is loaded into the cache. If data is accessed stride 1, than the cost of the miss is amortized over 8 words. Any stride other than one reduces the cost savings. Two of the ten codes studied suffered from non-unit strides. The codes represent two important classes of "strided" codes. Code 1 employs a multi-grid algorithm to reduce time to convergence. The grids are every tenth, fifth, second, and unit element. Since time to convergence is inversely proportional to the distance between elements, coarse grids converge quickly providing good starting values for finer grids. The better starting values further reduce the time to convergence. The downside is that grids of every nth element, n > 1, introduce non-unit strides into the computation. In the original code, much of the savings of the multi-grid algorithm were lost due to this problem. I eliminated the problem by compressing (copying) coarse grids into continuous memory, and rewriting the computation as a function of the compressed grid. On convergence, I copied the final values of the compressed grid back to the original grid. The savings gained from unit stride access of the compressed grid more than paid for the cost of copying. Using compressed grids, the loop from code 1 included in the previous section becomes do j = 1, GZ do i = 1, GZ T1 = CA(i+0, j-1) + CA(i-1, j+0) T4 = CA1(i+1, j+0) + CA1(i+0, j+1) S1 = T1 + T4 - 4 * CA1(i+0, j+0) CA(i+0, j+0) = CA1(i+0, j+0) + DD * S1 enddo enddo where CA and CA1 are compressed arrays of size GZ. Code 7 traverses a list of objects selecting objects for later processing. The labels of the selected objects are stored in an array. The selection step has unit stride, but the processing steps have irregular stride. A fix is to save the parameters of the selected objects in temporary arrays as they are selected, and pass the temporary arrays to the processing functions. The fix is practical if the same parameters are used in selection as in processing, or if processing comprises a series of distinct steps which use overlapping subsets of the parameters. Both conditions are true for code 7, so I achieved significant improvement by copying parameters to temporary arrays during selection. Data reuse In the previous sections, we optimized for spatial locality. It is also important to optimize for temporal locality. Once read, a datum should be used as much as possible before it is forced from cache. Loop fusion and loop unrolling are two techniques that increase temporal locality. Unfortunately, both techniques increase register pressure—as loop bodies become larger, the number of registers required to hold temporary values grows. Once register spilling occurs, any gains evaporate quickly. For multiprocessors with small register sets or small caches, the sweet spot can be very small. In the ten codes presented here, I found no opportunities for loop fusion and only two opportunities for loop unrolling (codes 1 and 3). In code 1, unrolling the outer and inner loop one iteration increases the number of result values computed by the loop body from 1 to 4, do J = 1, GZ-2, 2 do I = 1, GZ-2, 2 T1 = CA1(i+0, j-1) + CA1(i-1, j+0) T2 = CA1(i+1, j-1) + CA1(i+0, j+0) T3 = CA1(i+0, j+0) + CA1(i-1, j+1) T4 = CA1(i+1, j+0) + CA1(i+0, j+1) T5 = CA1(i+2, j+0) + CA1(i+1, j+1) T6 = CA1(i+1, j+1) + CA1(i+0, j+2) T7 = CA1(i+2, j+1) + CA1(i+1, j+2) S1 = T1 + T4 - 4 * CA1(i+0, j+0) S2 = T2 + T5 - 4 * CA1(i+1, j+0) S3 = T3 + T6 - 4 * CA1(i+0, j+1) S4 = T4 + T7 - 4 * CA1(i+1, j+1) CA(i+0, j+0) = CA1(i+0, j+0) + DD * S1 CA(i+1, j+0) = CA1(i+1, j+0) + DD * S2 CA(i+0, j+1) = CA1(i+0, j+1) + DD * S3 CA(i+1, j+1) = CA1(i+1, j+1) + DD * S4 enddo enddo The loop body executes 12 reads, whereas as the rolled loop shown in the previous section executes 20 reads to compute the same four values. In code 3, two loops are unrolled 8 times and one loop is unrolled 4 times. Here is the before for (k = 0; k < NK[u]; k++) { sum = 0.0; for (y = 0; y < NY; y++) { sum += W[y][u][k] * delta[y]; } backprop[i++]=sum; } and after code for (k = 0; k < KK - 8; k+=8) { sum0 = 0.0; sum1 = 0.0; sum2 = 0.0; sum3 = 0.0; sum4 = 0.0; sum5 = 0.0; sum6 = 0.0; sum7 = 0.0; for (y = 0; y < NY; y++) { sum0 += W[y][0][k+0] * delta[y]; sum1 += W[y][0][k+1] * delta[y]; sum2 += W[y][0][k+2] * delta[y]; sum3 += W[y][0][k+3] * delta[y]; sum4 += W[y][0][k+4] * delta[y]; sum5 += W[y][0][k+5] * delta[y]; sum6 += W[y][0][k+6] * delta[y]; sum7 += W[y][0][k+7] * delta[y]; } backprop[k+0] = sum0; backprop[k+1] = sum1; backprop[k+2] = sum2; backprop[k+3] = sum3; backprop[k+4] = sum4; backprop[k+5] = sum5; backprop[k+6] = sum6; backprop[k+7] = sum7; } for one of the loops unrolled 8 times. Optimizing for temporal locality is the most difficult optimization considered in this paper. The concepts are not difficult, but the sweet spot is small. Identifying where the program can benefit from loop unrolling or loop fusion is not trivial. Moreover, it takes some effort to get it right. Still, educating scientific programmers about temporal locality and teaching them how to optimize for it will pay dividends. Reducing instruction count Execution time is a function of instruction count. Reduce the count and you usually reduce the time. The best solution is to use a more efficient algorithm; that is, an algorithm whose order of complexity is smaller, that converges quicker, or is more accurate. Optimizing source code without changing the algorithm yields smaller, but still significant, gains. This paper considers only the latter because the intent is to study how much better codes can run if written by programmers schooled in basic code optimization techniques. The ten codes studied benefited from three types of "instruction reducing" optimizations. The two most prevalent were hoisting invariant memory and data operations out of inner loops. The third was eliminating unnecessary data copying. The nature of these inefficiencies is language dependent. Memory operations The semantics of C make it difficult for the compiler to determine all the invariant memory operations in a loop. The problem is particularly acute for loops in functions since the compiler may not know the values of the function's parameters at every call site when compiling the function. Most compilers support pragmas to help resolve ambiguities; however, these pragmas are not comprehensive and there is no standard syntax. To guarantee that invariant memory operations are not executed repetitively, the user has little choice but to hoist the operations by hand. The problem is not as severe in Fortran programs because in the absence of equivalence statements, it is a violation of the language's semantics for two names to share memory. Codes 3 and 5 are C programs. In both cases, the compiler did not hoist all invariant memory operations from inner loops. Consider the following loop from code 3 for (y = 0; y < NY; y++) { i = 0; for (u = 0; u < NU; u++) { for (k = 0; k < NK[u]; k++) { dW[y][u][k] += delta[y] * I1[i++]; } } } Since dW[y][u] can point to the same memory space as delta for one or more values of y and u, assignment to dW[y][u][k] may change the value of delta[y]. In reality, dW and delta do not overlap in memory, so I rewrote the loop as for (y = 0; y < NY; y++) { i = 0; Dy = delta[y]; for (u = 0; u < NU; u++) { for (k = 0; k < NK[u]; k++) { dW[y][u][k] += Dy * I1[i++]; } } } Failure to hoist invariant memory operations may be due to complex address calculations. If the compiler can not determine that the address calculation is invariant, then it can hoist neither the calculation nor the associated memory operations. As noted above, code 5 uses a macro to address four-dimensional arrays #define MAT4D(a,q,i,j,k) (double *)((a)->data + (q)*(a)->strides[0] + (i)*(a)->strides[3] + (j)*(a)->strides[2] + (k)*(a)->strides[1]) The macro is too complex for the compiler to understand and so, it does not identify any subexpressions as loop invariant. The simplest way to eliminate the address calculation from the innermost loop (over i) is to define a0 = MAT4D(a,q,0,j,k) before the loop and then replace all instances of *MAT4D(a,q,i,j,k) in the loop with a0[i] A similar problem appears in code 6, a Fortran program. The key loop in this program is do n1 = 1, nh nx1 = (n1 - 1) / nz + 1 nz1 = n1 - nz * (nx1 - 1) do n2 = 1, nh nx2 = (n2 - 1) / nz + 1 nz2 = n2 - nz * (nx2 - 1) ndx = nx2 - nx1 ndy = nz2 - nz1 gxx = grn(1,ndx,ndy) gyy = grn(2,ndx,ndy) gxy = grn(3,ndx,ndy) balance(n1,1) = balance(n1,1) + (force(n2,1) * gxx + force(n2,2) * gxy) * h1 balance(n1,2) = balance(n1,2) + (force(n2,1) * gxy + force(n2,2) * gyy)*h1 end do end do The programmer has written this loop well—there are no loop invariant operations with respect to n1 and n2. However, the loop resides within an iterative loop over time and the index calculations are independent with respect to time. Trading space for time, I precomputed the index values prior to the entering the time loop and stored the values in two arrays. I then replaced the index calculations with reads of the arrays. Data operations Ways to reduce data operations can appear in many forms. Implementing a more efficient algorithm produces the biggest gains. The closest I came to an algorithm change was in code 4. This code computes the inner product of K-vectors A(i) and B(j), 0 = i < N, 0 = j < M, for most values of i and j. Since the program computes most of the NM possible inner products, it is more efficient to compute all the inner products in one triply-nested loop rather than one at a time when needed. The savings accrue from reading A(i) once for all B(j) vectors and from loop unrolling. for (i = 0; i < N; i+=8) { for (j = 0; j < M; j++) { sum0 = 0.0; sum1 = 0.0; sum2 = 0.0; sum3 = 0.0; sum4 = 0.0; sum5 = 0.0; sum6 = 0.0; sum7 = 0.0; for (k = 0; k < K; k++) { sum0 += A[i+0][k] * B[j][k]; sum1 += A[i+1][k] * B[j][k]; sum2 += A[i+2][k] * B[j][k]; sum3 += A[i+3][k] * B[j][k]; sum4 += A[i+4][k] * B[j][k]; sum5 += A[i+5][k] * B[j][k]; sum6 += A[i+6][k] * B[j][k]; sum7 += A[i+7][k] * B[j][k]; } C[i+0][j] = sum0; C[i+1][j] = sum1; C[i+2][j] = sum2; C[i+3][j] = sum3; C[i+4][j] = sum4; C[i+5][j] = sum5; C[i+6][j] = sum6; C[i+7][j] = sum7; }} This change requires knowledge of a typical run; i.e., that most inner products are computed. The reasons for the change, however, derive from basic optimization concepts. It is the type of change easily made at development time by a knowledgeable programmer. In code 5, we have the data version of the index optimization in code 6. Here a very expensive computation is a function of the loop indices and so cannot be hoisted out of the loop; however, the computation is invariant with respect to an outer iterative loop over time. We can compute its value for each iteration of the computation loop prior to entering the time loop and save the values in an array. The increase in memory required to store the values is small in comparison to the large savings in time. The main loop in Code 8 is doubly nested. The inner loop includes a series of guarded computations; some are a function of the inner loop index but not the outer loop index while others are a function of the outer loop index but not the inner loop index for (j = 0; j < N; j++) { for (i = 0; i < M; i++) { r = i * hrmax; R = A[j]; temp = (PRM[3] == 0.0) ? 1.0 : pow(r, PRM[3]); high = temp * kcoeff * B[j] * PRM[2] * PRM[4]; low = high * PRM[6] * PRM[6] / (1.0 + pow(PRM[4] * PRM[6], 2.0)); kap = (R > PRM[6]) ? high * R * R / (1.0 + pow(PRM[4]*r, 2.0) : low * pow(R/PRM[6], PRM[5]); < rest of loop omitted > }} Note that the value of temp is invariant to j. Thus, we can hoist the computation for temp out of the loop and save its values in an array. for (i = 0; i < M; i++) { r = i * hrmax; TEMP[i] = pow(r, PRM[3]); } [N.B. – the case for PRM[3] = 0 is omitted and will be reintroduced later.] We now hoist out of the inner loop the computations invariant to i. Since the conditional guarding the value of kap is invariant to i, it behooves us to hoist the computation out of the inner loop, thereby executing the guard once rather than M times. The final version of the code is for (j = 0; j < N; j++) { R = rig[j] / 1000.; tmp1 = kcoeff * par[2] * beta[j] * par[4]; tmp2 = 1.0 + (par[4] * par[4] * par[6] * par[6]); tmp3 = 1.0 + (par[4] * par[4] * R * R); tmp4 = par[6] * par[6] / tmp2; tmp5 = R * R / tmp3; tmp6 = pow(R / par[6], par[5]); if ((par[3] == 0.0) && (R > par[6])) { for (i = 1; i <= imax1; i++) KAP[i] = tmp1 * tmp5; } else if ((par[3] == 0.0) && (R <= par[6])) { for (i = 1; i <= imax1; i++) KAP[i] = tmp1 * tmp4 * tmp6; } else if ((par[3] != 0.0) && (R > par[6])) { for (i = 1; i <= imax1; i++) KAP[i] = tmp1 * TEMP[i] * tmp5; } else if ((par[3] != 0.0) && (R <= par[6])) { for (i = 1; i <= imax1; i++) KAP[i] = tmp1 * TEMP[i] * tmp4 * tmp6; } for (i = 0; i < M; i++) { kap = KAP[i]; r = i * hrmax; < rest of loop omitted > } } Maybe not the prettiest piece of code, but certainly much more efficient than the original loop, Copy operations Several programs unnecessarily copy data from one data structure to another. This problem occurs in both Fortran and C programs, although it manifests itself differently in the two languages. Code 1 declares two arrays—one for old values and one for new values. At the end of each iteration, the array of new values is copied to the array of old values to reset the data structures for the next iteration. This problem occurs in Fortran programs not included in this study and in both Fortran 77 and Fortran 90 code. Introducing pointers to the arrays and swapping pointer values is an obvious way to eliminate the copying; but pointers is not a feature that many Fortran programmers know well or are comfortable using. An easy solution not involving pointers is to extend the dimension of the value array by 1 and use the last dimension to differentiate between arrays at different times. For example, if the data space is N x N, declare the array (N, N, 2). Then store the problem’s initial values in (_, _, 2) and define the scalar names new = 2 and old = 1. At the start of each iteration, swap old and new to reset the arrays. The old–new copy problem did not appear in any C program. In programs that had new and old values, the code swapped pointers to reset data structures. Where unnecessary coping did occur is in structure assignment and parameter passing. Structures in C are handled much like scalars. Assignment causes the data space of the right-hand name to be copied to the data space of the left-hand name. Similarly, when a structure is passed to a function, the data space of the actual parameter is copied to the data space of the formal parameter. If the structure is large and the assignment or function call is in an inner loop, then copying costs can grow quite large. While none of the ten programs considered here manifested this problem, it did occur in programs not included in the study. A simple fix is always to refer to structures via pointers. Optimizing loop structures Since scientific programs spend almost all their time in loops, efficient loops are the key to good performance. Conditionals, function calls, little instruction level parallelism, and large numbers of temporary values make it difficult for the compiler to generate tightly packed, highly efficient code. Conditionals and function calls introduce jumps that disrupt code flow. Users should eliminate or isolate conditionls to their own loops as much as possible. Often logical expressions can be substituted for if-then-else statements. For example, code 2 includes the following snippet MaxDelta = 0.0 do J = 1, N do I = 1, M < code omitted > Delta = abs(OldValue ? NewValue) if (Delta > MaxDelta) MaxDelta = Delta enddo enddo if (MaxDelta .gt. 0.001) goto 200 Since the only use of MaxDelta is to control the jump to 200 and all that matters is whether or not it is greater than 0.001, I made MaxDelta a boolean and rewrote the snippet as MaxDelta = .false. do J = 1, N do I = 1, M < code omitted > Delta = abs(OldValue ? NewValue) MaxDelta = MaxDelta .or. (Delta .gt. 0.001) enddo enddo if (MaxDelta) goto 200 thereby, eliminating the conditional expression from the inner loop. A microprocessor can execute many instructions per instruction cycle. Typically, it can execute one or more memory, floating point, integer, and jump operations. To be executed simultaneously, the operations must be independent. Thick loops tend to have more instruction level parallelism than thin loops. Moreover, they reduce memory traffice by maximizing data reuse. Loop unrolling and loop fusion are two techniques to increase the size of loop bodies. Several of the codes studied benefitted from loop unrolling, but none benefitted from loop fusion. This observation is not too surpising since it is the general tendency of programmers to write thick loops. As loops become thicker, the number of temporary values grows, increasing register pressure. If registers spill, then memory traffic increases and code flow is disrupted. A thick loop with many temporary values may execute slower than an equivalent series of thin loops. The biggest gain will be achieved if the thick loop can be split into a series of independent loops eliminating the need to write and read temporary arrays. I found such an occasion in code 10 where I split the loop do i = 1, n do j = 1, m A24(j,i)= S24(j,i) * T24(j,i) + S25(j,i) * U25(j,i) B24(j,i)= S24(j,i) * T25(j,i) + S25(j,i) * U24(j,i) A25(j,i)= S24(j,i) * C24(j,i) + S25(j,i) * V24(j,i) B25(j,i)= S24(j,i) * U25(j,i) + S25(j,i) * V25(j,i) C24(j,i)= S26(j,i) * T26(j,i) + S27(j,i) * U26(j,i) D24(j,i)= S26(j,i) * T27(j,i) + S27(j,i) * V26(j,i) C25(j,i)= S27(j,i) * S28(j,i) + S26(j,i) * U28(j,i) D25(j,i)= S27(j,i) * T28(j,i) + S26(j,i) * V28(j,i) end do end do into two disjoint loops do i = 1, n do j = 1, m A24(j,i)= S24(j,i) * T24(j,i) + S25(j,i) * U25(j,i) B24(j,i)= S24(j,i) * T25(j,i) + S25(j,i) * U24(j,i) A25(j,i)= S24(j,i) * C24(j,i) + S25(j,i) * V24(j,i) B25(j,i)= S24(j,i) * U25(j,i) + S25(j,i) * V25(j,i) end do end do do i = 1, n do j = 1, m C24(j,i)= S26(j,i) * T26(j,i) + S27(j,i) * U26(j,i) D24(j,i)= S26(j,i) * T27(j,i) + S27(j,i) * V26(j,i) C25(j,i)= S27(j,i) * S28(j,i) + S26(j,i) * U28(j,i) D25(j,i)= S27(j,i) * T28(j,i) + S26(j,i) * V28(j,i) end do end do Conclusions Over the course of the last year, I have had the opportunity to work with over two dozen academic scientific programmers at leading research universities. Their research interests span a broad range of scientific fields. Except for two programs that relied almost exclusively on library routines (matrix multiply and fast Fourier transform), I was able to improve significantly the single processor performance of all codes. Improvements range from 2x to 15.5x with a simple average of 4.75x. Changes to the source code were at a very high level. I did not use sophisticated techniques or programming tools to discover inefficiencies or effect the changes. Only one code was parallel despite the availability of parallel systems to all developers. Clearly, we have a problem—personal scientific research codes are highly inefficient and not running parallel. The developers are unaware of simple optimization techniques to make programs run faster. They lack education in the art of code optimization and parallel programming. I do not believe we can fix the problem by publishing additional books or training manuals. To date, the developers in questions have not studied the books or manual available, and are unlikely to do so in the future. Short courses are a possible solution, but I believe they are too concentrated to be much use. The general concepts can be taught in a three or four day course, but that is not enough time for students to practice what they learn and acquire the experience to apply and extend the concepts to their codes. Practice is the key to becoming proficient at optimization. I recommend that graduate students be required to take a semester length course in optimization and parallel programming. We would never give someone access to state-of-the-art scientific equipment costing hundreds of thousands of dollars without first requiring them to demonstrate that they know how to use the equipment. Yet the criterion for time on state-of-the-art supercomputers is at most an interesting project. Requestors are never asked to demonstrate that they know how to use the system, or can use the system effectively. A semester course would teach them the required skills. Government agencies that fund academic scientific research pay for most of the computer systems supporting scientific research as well as the development of most personal scientific codes. These agencies should require graduate schools to offer a course in optimization and parallel programming as a requirement for funding. About the Author John Feo received his Ph.D. in Computer Science from The University of Texas at Austin in 1986. After graduate school, Dr. Feo worked at Lawrence Livermore National Laboratory where he was the Group Leader of the Computer Research Group and principal investigator of the Sisal Language Project. In 1997, Dr. Feo joined Tera Computer Company where he was project manager for the MTA, and oversaw the programming and evaluation of the MTA at the San Diego Supercomputer Center. In 2000, Dr. Feo joined Sun Microsystems as an HPC application specialist. He works with university research groups to optimize and parallelize scientific codes. Dr. Feo has published over two dozen research articles in the areas of parallel parallel programming, parallel programming languages, and application performance.

    Read the article

  • Know your Data Lineage

    - by Simon Elliston Ball
    An academic paper without the footnotes isn’t an academic paper. Journalists wouldn’t base a news article on facts that they can’t verify. So why would anyone publish reports without being able to say where the data has come from and be confident of its quality, in other words, without knowing its lineage. (sometimes referred to as ‘provenance’ or ‘pedigree’) The number and variety of data sources, both traditional and new, increases inexorably. Data comes clean or dirty, processed or raw, unimpeachable or entirely fabricated. On its journey to our report, from its source, the data can travel through a network of interconnected pipes, passing through numerous distinct systems, each managed by different people. At each point along the pipeline, it can be changed, filtered, aggregated and combined. When the data finally emerges, how can we be sure that it is right? How can we be certain that no part of the data collection was based on incorrect assumptions, that key data points haven’t been left out, or that the sources are good? Even when we’re using data science to give us an approximate or probable answer, we cannot have any confidence in the results without confidence in the data from which it came. You need to know what has been done to your data, where it came from, and who is responsible for each stage of the analysis. This information represents your data lineage; it is your stack-trace. If you’re an analyst, suspicious of a number, it tells you why the number is there and how it got there. If you’re a developer, working on a pipeline, it provides the context you need to track down the bug. If you’re a manager, or an auditor, it lets you know the right things are being done. Lineage tracking is part of good data governance. Most audit and lineage systems require you to buy into their whole structure. If you are using Hadoop for your data storage and processing, then tools like Falcon allow you to track lineage, as long as you are using Falcon to write and run the pipeline. It can mean learning a new way of running your jobs (or using some sort of proxy), and even a distinct way of writing your queries. Other Hadoop tools provide a lot of operational and audit information, spread throughout the many logs produced by Hive, Sqoop, MapReduce and all the various moving parts that make up the eco-system. To get a full picture of what’s going on in your Hadoop system you need to capture both Falcon lineage and the data-exhaust of other tools that Falcon can’t orchestrate. However, the problem is bigger even that that. Often, Hadoop is just one piece in a larger processing workflow. The next step of the challenge is how you bind together the lineage metadata describing what happened before and after Hadoop, where ‘after’ could be  a data analysis environment like R, an application, or even directly into an end-user tool such as Tableau or Excel. One possibility is to push as much as you can of your key analytics into Hadoop, but would you give up the power, and familiarity of your existing tools in return for a reliable way of tracking lineage? Lineage and auditing should work consistently, automatically and quietly, allowing users to access their data with any tool they require to use. The real solution, therefore, is to create a consistent method by which to bring lineage data from these data various disparate sources into the data analysis platform that you use, rather than being forced to use the tool that manages the pipeline for the lineage and a different tool for the data analysis. The key is to keep your logs, keep your audit data, from every source, bring them together and use the data analysis tools to trace the paths from raw data to the answer that data analysis provides.

    Read the article

  • Applying for internship

    - by Margus
    At the moment I'm thinking about applying for internship at Eesti Energia. I seem to be eligible, but before contacting them I need to learn how to compile an informative and complete CV and cover letter. I do not consider myself as shallow minded, but also I'm not sure how to convincingly justify the stand of interest and how internship will help me in my future career. Course of life Tallinna Tehnikagümnaasium 2003 - 2006 Tallinna Tehnikaülikool 2006 – 2009 Military service at Signal Batallion Tallinna Tehnikaülikool 2010 – ... I started my academic career as Computer and Systems Engineer, but as I excelled in programming classes, I changed my major to Software Engineer and taken my specialty in web applications and logic. Nowadays I mainly use Java, Mathematica and C# to solve problems. For 2 times, I have taken part in ACM International Collegiate Programming Contest, where my team won the nationals and did pretty well in Europe. Also as part of notable thing in my academic career, my team wrote the Kalah game AI, that won in University's main programming class AI tournament. My hobbies are mind games and occasional problem solving. Few years ago I also competed in International Checkers EM (requires being in top 3 in nationals) as part of cadet and junior age group - I did not come close to winning, but I exceeded about half of the players each time. In high school and gymnasium I took part and later was the captain of team, that passes regionals and made it to top 3 of nationals (and later won) in (blitz) russian checkers. That was impressive because, it was a team effort as we only had (depending on year) 2-3 strong players. Although I started programming exactly 9,5 years ago I have no work experience. Well actually thats not true, as I completed my army duty, I was hired for a year (days still counting) to be apart of communicational (emergency) infrastructure action group where I'm the teams IT specialist (it's more complicated). So I consider myself to be aware of: rough conditions, teamwork, high stress tolerance, being on time and what responsibility means. As negative things I can mention: I do not have drivers licence. Although only Estonian and English are noted as requirements, then Russian is most likely required as well and I barely understand some of it. Reasons why I want to apply there, are: I need to do at least 4-6 week traineeship and it's in the right field I have the requirements and tasks seem easy enough Company is well known and has fairly good reputation Family and friend think, that it would be acceptable place to work Myriad of options to do final thesis about open up Work place is located in the same city I live atm. At moment, I see myself having a hard time explain why I would prefer it or where I see myself in 10 years if I was offered a job there. Question I have some idea how Curriculum Vitæ should look like, or I can google for template, but I'm not sure how to write informative one. Last I did one, it looked like: picture + contact information + education. Vaguely I only remember, that cover letter should be custom tailored for each place you apply containing ...

    Read the article

  • Rails CSS not Loading using Heroku

    - by eWizardII
    I have the following site set up here on Heroku - http://www.peerinstruction.net/users/sign_up the issue is that I have updated the css yet it is not being actively reflected on the site, it just shows a textbox, with some edited/custom fonts. I have attached the css file in the following gist - https://gist.github.com/f74b626c54ecbb60bbde The signup page controller: !!! Strict %html %head %title= yield(:title) || "Untitled" = stylesheet_link_tag 'application', 'web-app-theme/base', 'web-app-theme/themes/activo/style', 'web-app-theme/override' = javascript_include_tag :defaults = csrf_meta_tag = yield(:head) %body #container #header %h1 %a{:href => "/"} Peer Instruction Network #user-navigation %ul.wat-cf %li .content.login .flash - flash.each do |type, message| %div{ :class => "message #{type}" } %p= message = form_for(resource, :as => resource_name, :url => session_path(resource_name), :html => { :class => "form login" }) do |f| .group.wat-cf .left= f.label :email, :class => "label right" .right= f.text_field :email, :class => "text_field" .group.wat-cf .left= f.label :password, :class => "label right" .right= f.password_field :password, :class => "text_field" .group.wat-cf .right %button.button{ :type => "submit" } Login /= link_to "Sign In", destroy_user_session_path #box = yield The signup pages haml file: %h2 .block .content.login .flash - flash.each do |type, message| %div{ :class => "message #{type}" } %p= message = form_for(resource, :as => resource_name, :url => registration_path(resource_name)) do |f| = devise_error_messages! %div = f.label :firstname %br/ = f.text_field :firstname %div = f.label :middlename %br/ = f.text_field :middlename %div = f.label :lastname %br/ = f.text_field :lastname %div = f.label :email %br/ = f.email_field :email %div = f.label :password %br/ = f.password_field :password %div = f.label :academic %br/ = f.text_field :academic %div= f.submit "Continue" = render :partial => "devise/shared/links" I used web-app-theme to create an activo theme and then modify it.

    Read the article

  • In-order tree traversal

    - by Chris S
    I have the following text from an academic course I took a while ago about in-order traversal (they also call it pancaking) of a binary tree (not BST): In-order tree traversal Draw a line around the outside of the tree. Start to the left of the root, and go around the outside of the tree, to end up to the right of the root. Stay as close to the tree as possible, but do not cross the tree. (Think of the tree — its branches and nodes — as a solid barrier.) The order of the nodes is the order in which this line passes underneath them. If you are unsure as to when you go “underneath” a node, remember that a node “to the left” always comes first. Here's the example used (slightly different tree from below) However when I do a search on google, I get a conflicting definition. For example the wikipedia example: Inorder traversal sequence: A, B, C, D, E, F, G, H, I (leftchild,rootnode,right node) But according to (my understanding of) definition #1, this should be A, B, D, C, E, F, G, I, H Can anyone clarify which definition is correct? They might be both describing different traversal methods, but happen to be using the same name. I'm having trouble believing the peer-reviewed academic text is wrong, but can't be certain.

    Read the article

  • Sync my files across multiple computers

    - by EnderMB
    I do a lot of work on my home computer, ranging from programming, writing stored procedures and writing documentation and reporting. A lot of this work is university related and constantly swapping files across several computers is annoying at best. I have a large final-year project coming up and I'm going to be sharing this work amongst home and university and require some kind of online storage that provides version control for my programs, as well as my Word documents, PDF's and saved academic papers. Are there any good solutions for my problem?

    Read the article

  • Bittorrent surveillance/monitoring

    - by Flamewires
    Is there any tool to sniff bittorrent traffic and reassemble data about the torrent? Im looking for file names, peers, tracker address, local IP, etc. This is purely for academic interest in which all parties would be willing participants and therefore please dont upvote responses that talk merely about legal issues with using this kind of approach on a production network. I also am assuming that the torrent connections are unencrypted. Thanks

    Read the article

  • Sandbox on a linux server for group members

    - by mgualt
    I am a member of a large group (academic department) using a central GNU/Linux server. I would like to be able to install web apps like instiki, run version control repositories, and serve content over the web. But the admins won't permit this due to security concerns. Is there a way for them to sandbox me, protecting their servers in case I am hacked? What is the standard solution for a problem like this?

    Read the article

  • Is there any technical info about the youtube network?

    - by Allen
    I'm an IT graduate student trying to get my head around distributed content distribution systems, much like what I assume Youtube uses. I have read Google Research stuff like Bigtable and Google File System academic papers. Is there any such thing for Youtube? Can anyone point me at stuff to learn about the Youtube network and the underlying technology? thanks dbaman

    Read the article

< Previous Page | 1 2 3 4 5 6 7 8 9 10 11 12  | Next Page >