Search Results

Search found 1631 results on 66 pages for 'optimize'.

Page 14/66 | < Previous Page | 10 11 12 13 14 15 16 17 18 19 20 21  | Next Page >

  • How to make a single image in to different(optimize) sizes once image uploaded?

    - by Shakil
    Hi All i am working on mvc4(webapi) i need create for single image into different sizes for example i had loaded a image for file upload control once press the button of upload the single image should divide in to different sizes (like thumbnail large size medium size ) actually i am new for webapi i don't have any idea about it but i need to make this by writing hard code i will provide a path for it to store images could any body plz help me to do this work thank in advance

    Read the article

  • How to optimize a postgreSQL server for a "write once, read many"-type infrastructure ?

    - by mhu
    Greetings, I am working on a piece of software that logs entries (and related tagging) in a PostgreSQL database for storage and retrieval. We never update any data once it has been inserted; we might remove it when the entry gets too old, but this is done at most once a day. Stored entries can be retrieved by users. The insertion of new entries can happen rather fast and regularly, thus the database will commonly hold several millions elements. The tables used are pretty simple : one table for ids, raw content and insertion date; and one table storing tags and their values associated to an id. User search mostly concern tags values, so SELECTs usually consist of JOIN queries on ids on the two tables. To sum it up : 2 tables Lots of INSERT no UPDATE some DELETE, once a day at most some user-generated SELECT with JOIN huge data set What would an optimal server configuration (software and hardware, I assume for example that RAID10 could help) be for my PostgreSQL server, given these requirements ? By optimal, I mean one that allows SELECT queries taking a reasonably little amount of time. I can provide more information about the current setup (like tables, indexes ...) if needed.

    Read the article

  • How to optimize 'class name' in Android XML Layout ?

    - by Mac
    Hello, I am using some custom components in my project for that I am using following code. <view class="com.android.mypackage.myclass" id="@+id/button" android:layout_width="fill_parent" android:layout_height="fill_parent" android:background="@android:drawable/button" android:padding="10dip" /> Its working perfectly fine, I am using this code near about 35 times in my application. so while creating new clone application from same project, I need to update package name in 35 places. Is there any way to reduce these efforts? I had tried with "class="@string/class_name" but its not working.

    Read the article

  • How do you optimize database performance when providing results for autocomplete/iterative search?

    - by Howiecamp
    Note: In this question I'm using the term "autocomplete" (or "iterative search") to refer to returning search-as-you-type results, e.g. like Google Search gives you. Also my question is not specific to web applications vs. fat client apps. How are SQL SELECT queries normally constructed to provide decent performance for this type of query, especially over arbitrarily large data sets? In the case where the search will only query based on the first n characters (easiest case) am I still issuing a new SELECT result FROM sometable WHERE entry LIKE... on each keypress. Even with various forms of caching this seems like it might result in poor performance. In cases where you want your search string to return results with prefix matches, substring matches, etc. it's an even more difficult problem. Looking at a case of searching a list of contacts, you might return results that match FirstName + LastName, LastName + FirstName, or any other substring.

    Read the article

  • How optimize queries with fully qualified names in t-sql?

    - by tomaszs
    Whe I call: select * from Database.dbo.Table where NAME = 'cat' It takes: 200 ms And when I change database to Database in Management Studio and call it without fully qualified name it's much faster: select * from Table where NAME = 'cat' It takes: 17 ms Is there any way to make fully qualified queries faster without changing database?

    Read the article

  • Does the compiler optimize the function parameters passed by value?

    - by Naveen
    Lets say I have a function where the parameter is passed by value instead of const-reference. Further, lets assume that only the value is used inside the function i.e. the function doesn't try to modify it. In that case will the compiler will be able to figure out that it can pass the value by const-reference (for performance reasons) and generate the code accordingly? Is there any compiler which does that?

    Read the article

  • Does a c/c++ compiler optimize constant divisions by power-of-two value into shifts?

    - by porgarmingduod
    Question says it all. Does anyone know if the following... size_t div(size_t value) { const size_t x = 64; return value / x; } ...is optimized into? size_t div(size_t value) { return value >> 6; } Do compilers do this? (My interest lies in GCC). Are there situations where it does and others where it doesn't? I would really like to know, because every time I write a division that could be optimized like this I spend some mental energy wondering about whether precious nothings of a second is wasted doing a division where a shift would suffice.

    Read the article

  • Try to fill the GAE datastore but the code consumes to much cpu time. How to optimize this?

    - by Neverland
    I try to get the list of images in Amazon EC2 inside the Google datastore. I want to realize this with a cron job inside the GAE. class AmazonEC2uswest(db.Model): ami = db.StringProperty(required=True) mani = db.StringProperty() typ = db.StringProperty() arch = db.StringProperty() state = db.StringProperty() owner = db.StringProperty() class CronAMIsAmazonUS_WEST(webapp.RequestHandler): def get(self): aws_access_key_id_admin = "<secret>" aws_secret_access_key_admin = "<secret>" conn_us_west = boto.ec2.connect_to_region('us-west-1', aws_access_key_id=aws_access_key_id_admin, aws_secret_access_key=aws_secret_access_key_admin, is_secure = False) liste_images_us_west = conn_us_west.get_all_images() laenge_liste_images_us_west = len(liste_images_us_west) for i in range(laenge_liste_images_us_west): datastore_uswest_AMIs = AmazonEC2uswest(ami=liste_images_us_west[i].id, mani=str(liste_images_us_west[i].location), typ=liste_images_us_west[i].type, arch=liste_images_us_west[i].architecture, state=liste_images_us_west[i].state, owner=liste_images_us_west[i].ownerId) datastore_uswest_AMIs.put() The problem: Getting the list with get_all_images() lasts only a few seconds. But writing the data to the Google datastore needs way too much CPU time. My IBM T42p (P4M with 2GHz) needs for that piece of code approx. 1 Minute! Is it possible to optimize my code in a way that it needs fewer CPU time?

    Read the article

  • Lab Ops 2–The Lee-Robinson Script

    Marcus Robinson adapted PowerShell scripts by Thomas Lee to build a set of VMs to run a course in a reliable and repeatable way. With Marcus’s permission, Andrew Fryer has put that Setup Script on SkyDrive, and provided notes on the script. Optimize SQL Server performance“With SQL Monitor, we can be proactive in our optimization process, instead of waiting until a customer reports a problem,” John Trumbul, Sr. Software Engineer. Optimize your servers with a free trial.

    Read the article

  • Talend vs. SSIS: A Simple Performance Comparison

    With all of the ETL tools in the marketplace, which one is best? Jeff Singleton brings us simple performance comparison pitting SSIS against open source powerhouse Talend. Optimize SQL Server performance“With SQL Monitor, we can be proactive in our optimization process, instead of waiting until a customer reports a problem,” John Trumbul, Sr. Software Engineer. Optimize your servers with a free trial.

    Read the article

  • Premature-Optimization and Performance Anxiety

    - by James Michael Hare
    While writing my post analyzing the new .NET 4 ConcurrentDictionary class (here), I fell into one of the classic blunders that I myself always love to warn about.  After analyzing the differences of time between a Dictionary with locking versus the new ConcurrentDictionary class, I noted that the ConcurrentDictionary was faster with read-heavy multi-threaded operations.  Then, I made the classic blunder of thinking that because the original Dictionary with locking was faster for those write-heavy uses, it was the best choice for those types of tasks.  In short, I fell into the premature-optimization anti-pattern. Basically, the premature-optimization anti-pattern is when a developer is coding very early for a perceived (whether rightly-or-wrongly) performance gain and sacrificing good design and maintainability in the process.  At best, the performance gains are usually negligible and at worst, can either negatively impact performance, or can degrade maintainability so much that time to market suffers or the code becomes very fragile due to the complexity. Keep in mind the distinction above.  I'm not talking about valid performance decisions.  There are decisions one should make when designing and writing an application that are valid performance decisions.  Examples of this are knowing the best data structures for a given situation (Dictionary versus List, for example) and choosing performance algorithms (linear search vs. binary search).  But these in my mind are macro optimizations.  The error is not in deciding to use a better data structure or algorithm, the anti-pattern as stated above is when you attempt to over-optimize early on in such a way that it sacrifices maintainability. In my case, I was actually considering trading the safety and maintainability gains of the ConcurrentDictionary (no locking required) for a slight performance gain by using the Dictionary with locking.  This would have been a mistake as I would be trading maintainability (ConcurrentDictionary requires no locking which helps readability) and safety (ConcurrentDictionary is safe for iteration even while being modified and you don't risk the developer locking incorrectly) -- and I fell for it even when I knew to watch out for it.  I think in my case, and it may be true for others as well, a large part of it was due to the time I was trained as a developer.  I began college in in the 90s when C and C++ was king and hardware speed and memory were still relatively priceless commodities and not to be squandered.  In those days, using a long instead of a short could waste precious resources, and as such, we were taught to try to minimize space and favor performance.  This is why in many cases such early code-bases were very hard to maintain.  I don't know how many times I heard back then to avoid too many function calls because of the overhead -- and in fact just last year I heard a new hire in the company where I work declare that she didn't want to refactor a long method because of function call overhead.  Now back then, that may have been a valid concern, but with today's modern hardware even if you're calling a trivial method in an extremely tight loop (which chances are the JIT compiler would optimize anyway) the results of removing method calls to speed up performance are negligible for the great majority of applications.  Now, obviously, there are those coding applications where speed is absolutely king (for example drivers, computer games, operating systems) where such sacrifices may be made.  But I would strongly advice against such optimization because of it's cost.  Many folks that are performing an optimization think it's always a win-win.  That they're simply adding speed to the application, what could possibly be wrong with that?  What they don't realize is the cost of their choice.  For every piece of straight-forward code that you obfuscate with performance enhancements, you risk the introduction of bugs in the long term technical debt of the application.  It will become so fragile over time that maintenance will become a nightmare.  I've seen such applications in places I have worked.  There are times I've seen applications where the designer was so obsessed with performance that they even designed their own memory management system for their application to try to squeeze out every ounce of performance.  Unfortunately, the application stability often suffers as a result and it is very difficult for anyone other than the original designer to maintain. I've even seen this recently where I heard a C++ developer bemoaning that in VS2010 the iterators are about twice as slow as they used to be because Microsoft added range checking (probably as part of the 0x standard implementation).  To me this was almost a joke.  Twice as slow sounds bad, but it almost never as bad as you think -- especially if you're gaining safety.  The only time twice is really that much slower is when once was too slow to begin with.  Think about it.  2 minutes is slow as a response time because 1 minute is slow.  But if an iterator takes 1 microsecond to move one position and a new, safer iterator takes 2 microseconds, this is trivial!  The only way you'd ever really notice this would be in iterating a collection just for the sake of iterating (i.e. no other operations).  To my mind, the added safety makes the extra time worth it. Always favor safety and maintainability when you can.  I know it can be a hard habit to break, especially if you started out your career early or in a language such as C where they are very performance conscious.  But in reality, these type of micro-optimizations only end up hurting you in the long run. Remember the two laws of optimization.  I'm not sure where I first heard these, but they are so true: For beginners: Do not optimize. For experts: Do not optimize yet. This is so true.  If you're a beginner, resist the urge to optimize at all costs.  And if you are an expert, delay that decision.  As long as you have chosen the right data structures and algorithms for your task, your performance will probably be more than sufficient.  Chances are it will be network, database, or disk hits that will be your slow-down, not your code.  As they say, 98% of your code's bottleneck is in 2% of your code so premature-optimization may add maintenance and safety debt that won't have any measurable impact.  Instead, code for maintainability and safety, and then, and only then, when you find a true bottleneck, then you should go back and optimize further.

    Read the article

  • Stairway to PowerPivot and DAX - Level 3: The DAX DISTINCT() Function and Basic Distinct Counts

    Bill Pearson, Business Intelligence architect and author, exposes the DAX DISTINCT() function, and then provides some hands-on exposure to its use in generating distinct counts. Moreover, he further explores working with measures in the PivotTable in this, the third Level of our new Stairway to PowerPivot and DAX series. Optimize SQL Server performance“With SQL Monitor, we can be proactive in our optimization process, instead of waiting until a customer reports a problem,” John Trumbul, Sr. Software Engineer. Optimize your servers with a free trial.

    Read the article

  • How to implement SHA-2 in SQL Server 2005 or 2008 with a CLR assembly

    SQL Server 2012 supports SHA-256 and SHA-512 through the HASHBYTES() function, but earlier versions of SQL Server do not. SHA-256, SHA-384 and SHA-512 can, however, be implemented in SQL Server 2005 or SQL Server 2008 with the CLR assembly described in this article. Optimize SQL Server performance“With SQL Monitor, we can be proactive in our optimization process, instead of waiting until a customer reports a problem,” John Trumbul, Sr. Software Engineer. Optimize your servers with a free trial.

    Read the article

  • How to Become a Marketing Ninja

    Are you interested in discovering the best ways to optimize your websites, blogs and online CPA campaigns? There's a very handy suite of tolls available at an a modest price the can help you get organic traffic through SEO. With this useful software suite, you can track your page rank, optimize your content for SEO, even automatically import relevant content to your sites.

    Read the article

  • Identifying Incompatibility Issues When Migrating SQL Server Database to Windows Azure

    In this article, Marcin Policht looks at migrating existing SQL Server databases to Windows Azure, starting with identifying obstacles associated with such migrations. Optimize SQL Server performance“With SQL Monitor, we can be proactive in our optimization process, instead of waiting until a customer reports a problem,” John Trumbul, Sr. Software Engineer. Optimize your servers with a free trial.

    Read the article

  • Exporting Execution Plans - SQL Spackle

    A short SQL Spackle article to fill in your knowledge of SQL Server. In this one, Jason Brimhall shows how to export execution plans when you ask for query tuning help. Optimize SQL Server performance“With SQL Monitor, we can be proactive in our optimization process, instead of waiting until a customer reports a problem,” John Trumbul, Sr. Software Engineer. Optimize your servers with a free trial.

    Read the article

  • SQL Relay 2013R2

    Come join in the SQL Relay in 2013, attending a user group the week of Nov 11, 2013 in the UK. There are some amazing speakers, so be sure to register and attend. Optimize SQL Server performance“With SQL Monitor, we can be proactive in our optimization process, instead of waiting until a customer reports a problem,” John Trumbul, Sr. Software Engineer. Optimize your servers with a free trial.

    Read the article

  • SQL Saturday #237 - Charlotte

    The Charlotte BI Group (CBIG) is excited to bring the 2nd annual SQL Saturday, BI Edition, to Charlotte on October 19. This SQL Saturday training event is focused on Microsoft Business Intelligence, Analytics, and Data Administration topics. Optimize SQL Server performance“With SQL Monitor, we can be proactive in our optimization process, instead of waiting until a customer reports a problem,” John Trumbul, Sr. Software Engineer. Optimize your servers with a free trial.

    Read the article

  • Execute As

    Learn how you can use Execute As in a stored procedure to control permissions. This article includes an example that limits access to encryption routines for users. Optimize SQL Server performance“With SQL Monitor, we can be proactive in our optimization process, instead of waiting until a customer reports a problem,” John Trumbul, Sr. Software Engineer. Optimize your servers with a free trial.

    Read the article

  • Meet SQLBI at PASS Summit 2012 #sqlpass

    - by Marco Russo (SQLBI)
    Next week I and Alberto Ferrari will be in Seattle at PASS Summit 2012. You can meet us at our sessions, at a book signing and hopefully watching some other session during the conference. Here are our appointments: Thursday, November 08, 2012, 10:15 AM - 11:45 AM – Alberto Ferrari – Room 606-607 Querying and Optimizing DAX (BIA-321-S) Do you want to learn how to write DAX queries and how to optimize them? Don’t miss this session! Thursday, November 08, 2012, 12:00 PM - 12:30 PM – Bookstore Book signing event at the Bookstore corner with Alberto Ferrari, Marco Russo and Chris Webb Visit the bookstore and sign your copy of our Microsoft SQL Server 2012 Analysis Services: The BISM Tabular Model book. Thursday, November 08, 2012, 1:30 PM - 2:45 PM – Marco Russo – Room 611 Near Real-Time Analytics with xVelocity (without DirectQuery) (BIA-312) What’s the latency you can tolerate for your data? Discover what is the limit in Tabular without using DirectQuery and learn how to optimize your data model and your queries for a near real-time analytical system. Not a trivial task, but more affordable than you might think. Friday, November 09, 2012, 9:45 AM - 11:00 AM Parent-Child Hierarchies in Tabular (BIA-301) Multidimensional has a more advanced support for hierarchies than Tabular, but in reality you can do almost the same things by using data modeling, DAX functions and BIDS Helper!  Friday, November 09, 2012, 1:00 PM - 2:15 PM – Marco Russo – Room 612 Inside DAX Query Plans (BIA-403) Discover the query plan for your DAX query and learn how to read it and how to optimize a DAX query by using these information. If you meet us at the conference, stop us and say hello: it’s always nice to know our readers!

    Read the article

  • Flex 4 + Apache Ant, Cannot Load FlashPunk Libraries

    - by SquareCrow
    I have been searching google, Apache Docs*, and FlashPunk forums looking for an answer to this: I cannot get Ant/Flex to find and compile the FlashPunk libraries. Here is my build.xml. [code] <!-- Fetch the JAR full of Flex tasks if it is not already in the source directory --> <copy file="${FLEX_HOME}/ant/lib/flexTasks.jar" todir="${SOURCE_PATH}"/> <!-- Add flextasks to the project --> <taskdef resource="flexTasks.tasks" classpath="${SOURCE_PATH}/flexTasks.jar"></taskdef> <!-- Release build Flash Player 10.1 --> <target name="build"> <!-- Build the FlashPunk library --> <echo message="building swc..." /> <compc output="FlashPunk.swc" keep-generated-actionscript="false" incremental="false" optimize="false" debug="true" use-network="false"> <include-sources dir="${FLASHPUNK_PATH}/net" includes="**/* flashpunk/utils/* flashpunk/masks/*" excludes="**/*.TTF **/*.png"/> <load-config filename="${FLEX_HOME}/frameworks/flex-config.xml"/> </compc> <echo message="building swf..." /> <mxmlc file="${SOURCE_PATH}/epOne.as" output="${OUTPUT_PATH}/epOne.swf" debug="false" incremental="false" strict="true" accessible="false" link-report="link_report.xml" static-link-runtime-shared-libraries="true"> <optimize>true</optimize> </mxmlc> </target> [/code] Results in many errors of the type "Definition net.flashpunk.masks:Grid could not be found" even though when I open the directories I can see the *.AS files right there. Sorry if this is very basic. I am piecing together knowledge of Ant from docs and tutorials. *I decided to use Ant because neither FlashDevelop for Windows nor Eclipse for Linux seemto work for me.

    Read the article

  • How to reduce the CPU load on a hosting with WordPress installed as a CMS? [on hold]

    - by Akky Awesøme
    I have been using hostgators hatchling plan for three months. I got an email from the hosting that my website is creating an over load on CPU. They said that I am eating up their processor and as a precaution, they have temporarily suspended my account. When I contacted their customer support, they said: You have to optimize your database and use some sort of caching mechanism, where the script does not need to generate a new page with every request, helps to lower the over load that a script will cause. I am not a technical geek, I am wondering how I will do this thing. I don't have any resource to hire a web developer to do this job. My website is down for 48hours. I was using wp super cache along with cloudfare's free support. Now I have intalled optimize-db plugin and optimized my database. Please provide me with some more tips on how to optimize my database to reduce CPU usage. Any help would be appreciated.

    Read the article

  • Performance and Optimization Isn’t Evil

    - by Reed
    Donald Knuth is a fairly amazing guy.  I consider him one of the most influential contributors to computer science of all time.  Unfortunately, most of the time I hear his name, I cringe.  This is because it’s typically somebody quoting a small portion of one of his famous statements on optimization: “premature optimization is the root of all evil.” I mention that this is only a portion of the entire quote, and, as such, I feel that Knuth is being quoted out of context.  Optimization is important.  It is a critical part of every software development effort, and should never be ignored.  A developer who ignores optimization is not a professional.  Every developer should understand optimization – know what to optimize, when to optimize it, and how to think about code in a way that is intelligent and productive from day one. I want to start by discussing my own, personal motivation here.  I recently wrote about a performance issue I ran across, and was slammed by multiple comments and emails that effectively boiled down to: “You’re an idiot.  Premature optimization is the root of all evil.  This doesn’t matter.”  It didn’t matter that I discovered this while measuring in a profiler, and that it was a portion of my code base that can take “many hours to complete.”  Even so, multiple people instantly jump to “it’s premature – it doesn’t matter.” This is a common thread I see.  For example, StackOverflow has many pages of posts with answers that boil down to (mis)quoting Knuth.  In fact, just about any question relating to a performance related issue gets this quote thrown at it immediately – whether it deserves it or not.  That being said, I did receive some positive comments and emails as well.  Many people want to understand how to optimize their code, approaches to take, tools and techniques they can use, and any other advice they can discover. First, lets get back to Knuth – I mentioned before that Knuth is being quoted out of context.  Lets start by looking at the entire quote from his 1974 paper Structured Programming with go to Statements: “We should forget about small efficiencies, say about 97% of the time: premature optimization is the root of all evil. Yet we should not pass up our opportunities in that critical 3%. A good programmer will not be lulled into complacency by such reasoning, he will be wise to look carefully at the critical code; but only after that code has been identified.” Ironically, if you read Knuth’s original paper, this statement was made in the middle of a discussion of how Knuth himself had changed how he approaches optimization.  It was never a statement saying “don’t optimize”, but rather, “optimizing intelligently provides huge advantages.”  His approach had three benefits: “a) it doesn’t take long” … “b) the payoff is real”, c) you can “be less efficient in the other parts of my programs, which therefore are more readable and more easily written and debugged.” Looking at Knuth’s premise here, and reading that section of his paper, really leads to a few observations: Optimization is important  “he will be wise to look carefully at the critical code” Normally, 3% of your code – three lines out of every 100 you write, are “critical code” and will require some optimization: “we should not pass up our opportunities in that critical 3%” Optimization, if done well, should not be time consuming: “it doesn’t take long” Optimization, if done correctly, provides real benefits: “the payoff is real” None of this is new information.  People who care about optimization have been discussing this for years – for example, Rico Mariani’s Designing For Performance (a fantastic article) discusses many of the same issues very intelligently. That being said, many developers seem unable or unwilling to consider optimization.  Many others don’t seem to know where to start.  As such, I’m going to spend some time writing about optimization – what is it, how should we think about it, and what can we do to improve our own code.

    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

< Previous Page | 10 11 12 13 14 15 16 17 18 19 20 21  | Next Page >