Search Results

Search found 9667 results on 387 pages for 'parallel loop'.

Page 8/387 | < Previous Page | 4 5 6 7 8 9 10 11 12 13 14 15  | Next Page >

  • Unified Parallel C - examples and list of extensions

    - by osgx
    Hello Where can I find examples of code, written in "Unified Parallel C"? I also interested in normative documents about this language (standards, reference manuals, online-accessible books and courses). What extensions were added to C to get UPC? Is this dialect alive or dead?

    Read the article

  • Send raw data to USB parallel port after upgrading to 11.10

    - by zaphod
    I have a laser cutter connected via a generic USB to parallel adapter. The laser cutter speaks HPGL, as it happens, but since this is a laser cutter and not a plotter, I usually want to generate the HPGL myself, since I care about the ordering, speed, and direction of cuts and so on. In previous versions of Ubuntu, I was able to print to the cutter by copying an HPGL file directly to the corresponding USB "lp" device. For example: cp foo.plt /dev/usblp1 Well, I just upgraded to Ubuntu 11.10 oneiric, and I can't find any "lp" devices in /dev anymore. D'oh! What's the preferred way to send raw data to a parallel port in Ubuntu? I've tried System Settings Printing + Add, hoping that I might be able to associate my device with some kind of "raw printer" driver and print to it with a command like lp -d LaserCutter foo.plt But my USB to parallel adapter doesn't seem to show up in the list. What I do see are my HP Color LaserJet, two USB-to-serial adapters, "Enter URI", and "Network Printer". Meanwhile, over in /dev, I do see /dev/ttyUSB0 and /dev/ttyUSB1 devices for the 2 USB-to-serial adapters. I don't see anything obvious corresponding to the HP printer (which was /dev/usblp0 prior to the upgrade), except for generic USB stuff. For example, sudo find /dev | grep lp produces no output. I do seem to be able to print to the HP printer just fine, though. The printer setup GUI gives it a device URI starting with "hp:" which isn't much help for the parallel adapter. The CUPS administrator's guide makes it sound like I might need to feed it a device URI of the form parallel:/dev/SOMETHING, but of course if I had a /dev/SOMETHING I'd probably just go on writing to it directly. Here's what dmesg says after I disconnect and reconnect the device from the USB port: [ 924.722906] usb 1-1.1.4: USB disconnect, device number 7 [ 959.993002] usb 1-1.1.4: new full speed USB device number 8 using ehci_hcd And here's how it shows up in lsusb -v: Bus 001 Device 008: ID 1a86:7584 QinHeng Electronics CH340S Device Descriptor: bLength 18 bDescriptorType 1 bcdUSB 1.10 bDeviceClass 0 (Defined at Interface level) bDeviceSubClass 0 bDeviceProtocol 0 bMaxPacketSize0 8 idVendor 0x1a86 QinHeng Electronics idProduct 0x7584 CH340S bcdDevice 2.52 iManufacturer 0 iProduct 2 USB2.0-Print iSerial 0 bNumConfigurations 1 Configuration Descriptor: bLength 9 bDescriptorType 2 wTotalLength 32 bNumInterfaces 1 bConfigurationValue 1 iConfiguration 0 bmAttributes 0x80 (Bus Powered) MaxPower 96mA Interface Descriptor: bLength 9 bDescriptorType 4 bInterfaceNumber 0 bAlternateSetting 0 bNumEndpoints 2 bInterfaceClass 7 Printer bInterfaceSubClass 1 Printer bInterfaceProtocol 2 Bidirectional iInterface 0 Endpoint Descriptor: bLength 7 bDescriptorType 5 bEndpointAddress 0x82 EP 2 IN bmAttributes 2 Transfer Type Bulk Synch Type None Usage Type Data wMaxPacketSize 0x0020 1x 32 bytes bInterval 0 Endpoint Descriptor: bLength 7 bDescriptorType 5 bEndpointAddress 0x02 EP 2 OUT bmAttributes 2 Transfer Type Bulk Synch Type None Usage Type Data wMaxPacketSize 0x0020 1x 32 bytes bInterval 0 Device Status: 0x0000 (Bus Powered)

    Read the article

  • Send raw data to USB parallel port after upgrading to 11.10 oneiric

    - by zaphod
    I have a laser cutter connected via a generic USB to parallel adapter. The laser cutter speaks HPGL, as it happens, but since this is a laser cutter and not a plotter, I usually want to generate the HPGL myself, since I care about the ordering, speed, and direction of cuts and so on. In previous versions of Ubuntu, I was able to print to the cutter by copying an HPGL file directly to the corresponding USB "lp" device. For example: cp foo.plt /dev/usblp1 Well, I just upgraded to Ubuntu 11.10 oneiric, and I can't find any "lp" devices in /dev anymore. D'oh! What's the preferred way to send raw data to a parallel port in Ubuntu? I've tried System Settings Printing + Add, hoping that I might be able to associate my device with some kind of "raw printer" driver and print to it with a command like lp -d LaserCutter foo.plt But my USB to parallel adapter doesn't seem to show up in the list. What I do see are my HP Color LaserJet, two USB-to-serial adapters, "Enter URI", and "Network Printer". Meanwhile, over in /dev, I do see /dev/ttyUSB0 and /dev/ttyUSB1 devices for the 2 USB-to-serial adapters. I don't see anything obvious corresponding to the HP printer (which was /dev/usblp0 prior to the upgrade), except for generic USB stuff. For example, sudo find /dev | grep lp produces no output. I do seem to be able to print to the HP printer just fine, though. The printer setup GUI gives it a device URI starting with "hp:" which isn't much help for the parallel adapter. The CUPS administrator's guide makes it sound like I might need to feed it a device URI of the form parallel:/dev/SOMETHING, but of course if I had a /dev/SOMETHING I'd probably just go on writing to it directly. Here's what dmesg says after I disconnect and reconnect the device from the USB port: [ 924.722906] usb 1-1.1.4: USB disconnect, device number 7 [ 959.993002] usb 1-1.1.4: new full speed USB device number 8 using ehci_hcd And here's how it shows up in lsusb -v: Bus 001 Device 008: ID 1a86:7584 QinHeng Electronics CH340S Device Descriptor: bLength 18 bDescriptorType 1 bcdUSB 1.10 bDeviceClass 0 (Defined at Interface level) bDeviceSubClass 0 bDeviceProtocol 0 bMaxPacketSize0 8 idVendor 0x1a86 QinHeng Electronics idProduct 0x7584 CH340S bcdDevice 2.52 iManufacturer 0 iProduct 2 USB2.0-Print iSerial 0 bNumConfigurations 1 Configuration Descriptor: bLength 9 bDescriptorType 2 wTotalLength 32 bNumInterfaces 1 bConfigurationValue 1 iConfiguration 0 bmAttributes 0x80 (Bus Powered) MaxPower 96mA Interface Descriptor: bLength 9 bDescriptorType 4 bInterfaceNumber 0 bAlternateSetting 0 bNumEndpoints 2 bInterfaceClass 7 Printer bInterfaceSubClass 1 Printer bInterfaceProtocol 2 Bidirectional iInterface 0 Endpoint Descriptor: bLength 7 bDescriptorType 5 bEndpointAddress 0x82 EP 2 IN bmAttributes 2 Transfer Type Bulk Synch Type None Usage Type Data wMaxPacketSize 0x0020 1x 32 bytes bInterval 0 Endpoint Descriptor: bLength 7 bDescriptorType 5 bEndpointAddress 0x02 EP 2 OUT bmAttributes 2 Transfer Type Bulk Synch Type None Usage Type Data wMaxPacketSize 0x0020 1x 32 bytes bInterval 0 Device Status: 0x0000 (Bus Powered)

    Read the article

  • Asynchronous event loop design and issues.

    - by Artyom
    Hello, I'm designing event loop for asynchronous socket IO using epoll/devpoll/kqueue/poll/select (including windows-select). I have two options of performing, IO operation: Non-blocking mode, poll on EAGAIN Set socket to non-blocking mode. Read/Write to socket. If operation succeeds, post completion notification to event loop. If I get EAGAIN, add socket to "select list" and poll socket. Polling mode: poll and then execute Add socket to select list and poll it. Wait for notification that it is readable writable read/write Post completion notification to event loop of sucseeds To me it looks like first would require less system calls when using in normal mode, especially for writing to socket (buffers are quite big). Also it looks like that it would be possible to reduce the overhead over number of "select" executions, especially it is nice when you do not have something that scales well as epoll/devpoll/kqueue. Questions: Are there any advantages of the second approach? Are there any portability issues with non-blocking operations on sockets/file descriptors over numerous operating systems: Linux, FreeBSD, Solaris, MacOSX, Windows. Notes: Please do not suggest using existing event-loop/socket-api implementations

    Read the article

  • loop prematurely quitting

    - by Nick Gibson
    This loop works fine but prematurely quits at times. I set a piece of code in it so that I can view the random number. It only closes prematurely when the random number is equal to the highest numbered question the user inputs (Example...a user wants 10 questions, if the random number is 10 the program quits.) I have no idea why since i have it set to if(random number <= the number of questions) for ( int loop = 1; loop < loopCount; loop++ ) { aa = r.nextInt ( 10 + 1 ); abc = ( int ) aa; String[] userAnswer = new String[x]; JOptionPane.showMessageDialog ( null, abc ); if ( abc <= x ) { for ( overValue = 1; overValue < forLoop; overValue++ ); { userAnswer[j] = JOptionPane.showInputDialog ( null, "Question " + quesNum + "\n" + questions[abc] + "\n\nA: " + a[abc] + "\nB: " + b[abc] + "\nC: " + c[abc] + "\nD: " + d[abc] ); if ( userAnswer[j].equals ( answers[j] ) ) { JOptionPane.showMessageDialog ( null, "Correct. \nThe Correct Answer is " + answers[abc] ); } else { JOptionPane.showMessageDialog ( null, "Wrong. \n The Correct Answer is " + answers[abc] ); }//else }//for }//if }//for

    Read the article

  • seriouosly elusive for loop (racking my brains!)

    - by user1693359
    I've got a loop issue in Python 2.72 that's really frustrating me. Basically the loop is not iterating fast the first index (j), and I've tried all sorts of ways to fix it with no luck. def learn(dataSet): for i in dataSet.getNext(): recall = raw_input("Enter all members of %s you are able to recall >>> (separated by commas) " % (i.getName())) missed = i.getMembers() missedString = [] for a in missed: missedString.append(a.getName()) Here is the loop I can't get to iterate. The first for loop only goes through the first iteration of 'j' in the split string list, then removes it from 'missedString'. I would like for all members of the split-string 'recall' to be removed from 'missedString'. for j in string.split(recall, ','): if j in missedString: missedString.remove(j) continue for b in missed: if b.getName() not in missedString: missed.remove(b) print 'You missed %d. ' % (len(missed)) if (len(missed)) > 0: print 'Maybe a hint or two will help...' for miss in missed: remind(miss.getSecs(), i.getName(), missed) I really have no clue, help would be appreciated!

    Read the article

  • Using TPL and PLINQ to raise performance of feed aggregator

    - by DigiMortal
    In this posting I will show you how to use Task Parallel Library (TPL) and PLINQ features to boost performance of simple RSS-feed aggregator. I will use here only very basic .NET classes that almost every developer starts from when learning parallel programming. Of course, we will also measure how every optimization affects performance of feed aggregator. Feed aggregator Our feed aggregator works as follows: Load list of blogs Download RSS-feed Parse feed XML Add new posts to database Our feed aggregator is run by task scheduler after every 15 minutes by example. We will start our journey with serial implementation of feed aggregator. Second step is to use task parallelism and parallelize feeds downloading and parsing. And our last step is to use data parallelism to parallelize database operations. We will use Stopwatch class to measure how much time it takes for aggregator to download and insert all posts from all registered blogs. After every run we empty posts table in database. Serial aggregation Before doing parallel stuff let’s take a look at serial implementation of feed aggregator. All tasks happen one after other. internal class FeedClient {     private readonly INewsService _newsService;     private const int FeedItemContentMaxLength = 255;       public FeedClient()     {          ObjectFactory.Initialize(container =>          {              container.PullConfigurationFromAppConfig = true;          });           _newsService = ObjectFactory.GetInstance<INewsService>();     }       public void Execute()     {         var blogs = _newsService.ListPublishedBlogs();           for (var index = 0; index <blogs.Count; index++)         {              ImportFeed(blogs[index]);         }     }       private void ImportFeed(BlogDto blog)     {         if(blog == null)             return;         if (string.IsNullOrEmpty(blog.RssUrl))             return;           var uri = new Uri(blog.RssUrl);         SyndicationContentFormat feedFormat;           feedFormat = SyndicationDiscoveryUtility.SyndicationContentFormatGet(uri);           if (feedFormat == SyndicationContentFormat.Rss)             ImportRssFeed(blog);         if (feedFormat == SyndicationContentFormat.Atom)             ImportAtomFeed(blog);                 }       private void ImportRssFeed(BlogDto blog)     {         var uri = new Uri(blog.RssUrl);         var feed = RssFeed.Create(uri);           foreach (var item in feed.Channel.Items)         {             SaveRssFeedItem(item, blog.Id, blog.CreatedById);         }     }       private void ImportAtomFeed(BlogDto blog)     {         var uri = new Uri(blog.RssUrl);         var feed = AtomFeed.Create(uri);           foreach (var item in feed.Entries)         {             SaveAtomFeedEntry(item, blog.Id, blog.CreatedById);         }     } } Serial implementation of feed aggregator downloads and inserts all posts with 25.46 seconds. Task parallelism Task parallelism means that separate tasks are run in parallel. You can find out more about task parallelism from MSDN page Task Parallelism (Task Parallel Library) and Wikipedia page Task parallelism. Although finding parts of code that can run safely in parallel without synchronization issues is not easy task we are lucky this time. Feeds import and parsing is perfect candidate for parallel tasks. We can safely parallelize feeds import because importing tasks doesn’t share any resources and therefore they don’t also need any synchronization. After getting the list of blogs we iterate through the collection and start new TPL task for each blog feed aggregation. internal class FeedClient {     private readonly INewsService _newsService;     private const int FeedItemContentMaxLength = 255;       public FeedClient()     {          ObjectFactory.Initialize(container =>          {              container.PullConfigurationFromAppConfig = true;          });           _newsService = ObjectFactory.GetInstance<INewsService>();     }       public void Execute()     {         var blogs = _newsService.ListPublishedBlogs();                var tasks = new Task[blogs.Count];           for (var index = 0; index <blogs.Count; index++)         {             tasks[index] = new Task(ImportFeed, blogs[index]);             tasks[index].Start();         }           Task.WaitAll(tasks);     }       private void ImportFeed(object blogObject)     {         if(blogObject == null)             return;         var blog = (BlogDto)blogObject;         if (string.IsNullOrEmpty(blog.RssUrl))             return;           var uri = new Uri(blog.RssUrl);         SyndicationContentFormat feedFormat;           feedFormat = SyndicationDiscoveryUtility.SyndicationContentFormatGet(uri);           if (feedFormat == SyndicationContentFormat.Rss)             ImportRssFeed(blog);         if (feedFormat == SyndicationContentFormat.Atom)             ImportAtomFeed(blog);                }       private void ImportRssFeed(BlogDto blog)     {          var uri = new Uri(blog.RssUrl);          var feed = RssFeed.Create(uri);           foreach (var item in feed.Channel.Items)          {              SaveRssFeedItem(item, blog.Id, blog.CreatedById);          }     }     private void ImportAtomFeed(BlogDto blog)     {         var uri = new Uri(blog.RssUrl);         var feed = AtomFeed.Create(uri);           foreach (var item in feed.Entries)         {             SaveAtomFeedEntry(item, blog.Id, blog.CreatedById);         }     } } You should notice first signs of the power of TPL. We made only minor changes to our code to parallelize blog feeds aggregating. On my machine this modification gives some performance boost – time is now 17.57 seconds. Data parallelism There is one more way how to parallelize activities. Previous section introduced task or operation based parallelism, this section introduces data based parallelism. By MSDN page Data Parallelism (Task Parallel Library) data parallelism refers to scenario in which the same operation is performed concurrently on elements in a source collection or array. In our code we have independent collections we can process in parallel – imported feed entries. As checking for feed entry existence and inserting it if it is missing from database doesn’t affect other entries the imported feed entries collection is ideal candidate for parallelization. internal class FeedClient {     private readonly INewsService _newsService;     private const int FeedItemContentMaxLength = 255;       public FeedClient()     {          ObjectFactory.Initialize(container =>          {              container.PullConfigurationFromAppConfig = true;          });           _newsService = ObjectFactory.GetInstance<INewsService>();     }       public void Execute()     {         var blogs = _newsService.ListPublishedBlogs();                var tasks = new Task[blogs.Count];           for (var index = 0; index <blogs.Count; index++)         {             tasks[index] = new Task(ImportFeed, blogs[index]);             tasks[index].Start();         }           Task.WaitAll(tasks);     }       private void ImportFeed(object blogObject)     {         if(blogObject == null)             return;         var blog = (BlogDto)blogObject;         if (string.IsNullOrEmpty(blog.RssUrl))             return;           var uri = new Uri(blog.RssUrl);         SyndicationContentFormat feedFormat;           feedFormat = SyndicationDiscoveryUtility.SyndicationContentFormatGet(uri);           if (feedFormat == SyndicationContentFormat.Rss)             ImportRssFeed(blog);         if (feedFormat == SyndicationContentFormat.Atom)             ImportAtomFeed(blog);                }       private void ImportRssFeed(BlogDto blog)     {         var uri = new Uri(blog.RssUrl);         var feed = RssFeed.Create(uri);           feed.Channel.Items.AsParallel().ForAll(a =>         {             SaveRssFeedItem(a, blog.Id, blog.CreatedById);         });      }        private void ImportAtomFeed(BlogDto blog)      {         var uri = new Uri(blog.RssUrl);         var feed = AtomFeed.Create(uri);           feed.Entries.AsParallel().ForAll(a =>         {              SaveAtomFeedEntry(a, blog.Id, blog.CreatedById);         });      } } We did small change again and as the result we parallelized checking and saving of feed items. This change was data centric as we applied same operation to all elements in collection. On my machine I got better performance again. Time is now 11.22 seconds. Results Let’s visualize our measurement results (numbers are given in seconds). As we can see then with task parallelism feed aggregation takes about 25% less time than in original case. When adding data parallelism to task parallelism our aggregation takes about 2.3 times less time than in original case. More about TPL and PLINQ Adding parallelism to your application can be very challenging task. You have to carefully find out parts of your code where you can safely go to parallel processing and even then you have to measure the effects of parallel processing to find out if parallel code performs better. If you are not careful then troubles you will face later are worse than ones you have seen before (imagine error that occurs by average only once per 10000 code runs). Parallel programming is something that is hard to ignore. Effective programs are able to use multiple cores of processors. Using TPL you can also set degree of parallelism so your application doesn’t use all computing cores and leaves one or more of them free for host system and other processes. And there are many more things in TPL that make it easier for you to start and go on with parallel programming. In next major version all .NET languages will have built-in support for parallel programming. There will be also new language constructs that support parallel programming. Currently you can download Visual Studio Async to get some idea about what is coming. Conclusion Parallel programming is very challenging but good tools offered by Visual Studio and .NET Framework make it way easier for us. In this posting we started with feed aggregator that imports feed items on serial mode. With two steps we parallelized feed importing and entries inserting gaining 2.3 times raise in performance. Although this number is specific to my test environment it shows clearly that parallel programming may raise the performance of your application significantly.

    Read the article

  • Explaining Explain Plan Notes for Auto DOP

    - by jean-pierre.dijcks
    I've recently gotten some questions around "why do I not see a parallel plan" while Auto DOP is on (I think)...? It is probably worthwhile to quickly go over some of the ways to find out what Auto DOP was thinking. In general, there is no need to go tracing sessions and look under the hood. The thing to start with is to do an explain plan on your statement and to look at the parameter settings on the system. Parameter Settings to Look At First and foremost, make sure that parallel_degree_policy = AUTO. If you have that parameter set to LIMITED you will not have queuing and we will only do the auto magic if your objects are set to default parallel (so no degree specified). Next you want to look at the value of parallel_degree_limit. It is typically set to CPU, which in default settings equates to the Default DOP of the system. If you are testing Auto DOP itself and the impact it has on performance you may want to leave it at this CPU setting. If you are running concurrent statements you may want to give this some more thoughts. See here for more information. In general, do stick with either CPU or with a specific number. For now avoid the IO setting as I've seen some mixed results with that... In 11.2.0.2 you should also check that IO Calibrate has been run. Best to simply do a: SQL> select * from V$IO_CALIBRATION_STATUS; STATUS        CALIBRATION_TIME ------------- ---------------------------------------------------------------- READY         04-JAN-11 10.04.13.104 AM You should see that your IO Calibrate is READY and therefore Auto DOP is ready. In any case, if you did not run the IO Calibrate step you will get the following note in the explain plan: Note -----    - automatic DOP: skipped because of IO calibrate statistics are missing One more note on calibrate_io, if you do not have asynchronous IO enabled you will see:  ERROR at line 1: ORA-56708: Could not find any datafiles with asynchronous i/o capability ORA-06512: at "SYS.DBMS_RMIN", line 463 ORA-06512: at "SYS.DBMS_RESOURCE_MANAGER", line 1296 ORA-06512: at line 7 While this is changed in some fixes to the calibrate procedure, you should really consider switching asynchronous IO on for your data warehouse. Explain Plan Explanation To see the notes that are shown and explained here (and the above little snippet ) you can use a simple explain plan mechanism. There should  be no need to add +parallel etc. explain plan for <statement> SELECT PLAN_TABLE_OUTPUT FROM TABLE(DBMS_XPLAN.DISPLAY()); Auto DOP The note structure displaying why Auto DOP did not work (with the exception noted above on IO Calibrate) is like this: Automatic degree of parallelism is disabled: <reason> These are the reason codes: Parameter -  parallel_degree_policy = manual which will not allow Auto DOP to kick in  Hint - One of the following hints are used NOPARALLEL, PARALLEL(1), PARALLEL(MANUAL) Outline - A SQL outline of an older version (before 11.2) is used SQL property restriction - The statement type does not allow for parallel processing Rule-based mode - Instead of the Cost Based Optimizer the system is using the RBO Recursive SQL statement - The statement type does not allow for parallel processing pq disabled/pdml disabled/pddl disabled - For some reason (alter session?) parallelism is disabled Limited mode but no parallel objects referenced - your parallel_degree_policy = LIMITED and no objects in the statement are decorated with the default PARALLEL degree. In most cases all objects have a specific degree in which case Auto DOP will honor that degree. Parallel Degree Limited When Auto DOP does it works you may see the cap you imposed with parallel_degree_limit showing up in the note section of the explain plan: Note -----    - automatic DOP: Computed Degree of Parallelism is 16 because of degree limit This is an obvious indication that your are being capped for this statement. There is one quite interesting one that happens when you are being capped at DOP = 1. First of you get a serial plan and the note changes slightly in that it does not indicate it is being capped (we hope to update the note at some point in time to be more specific). It right now looks like this: Note -----    - automatic DOP: Computed Degree of Parallelism is 1 Dynamic Sampling With 11.2.0.2 you will start seeing another interesting change in parallel plans, and since we are talking about the note section here, I figured we throw this in for good measure. If we deem the parallel (!) statement complex enough, we will enact dynamic sampling on your query. This happens as long as you did not change the default for dynamic sampling on the system. The note looks like this: Note ----- - dynamic sampling used for this statement (level=5)

    Read the article

  • How do I best remove an entity from my game loop when it is dead?

    - by Iain
    Ok so I have a big list of all my entities which I loop through and update. In AS3 I can store this as an Array (dynamic length, untyped), a Vector (typed) or a linked list (not native). At the moment I'm using Array but I plan to change to Vector or linked list if it is faster. Anyway, my question, when an Entity is destroyed, how should I remove it from the list? I could null its position, splice it out or just set a flag on it to say "skip over me, I'm dead." I'm pooling my entities, so an Entity that is dead is quite likely to be alive again at some point. For each type of collection what is my best strategy, and which combination of collection type and removal method will work best?

    Read the article

  • Software to create a virtual parallel port in Windows XP?

    - by drknexus
    I am writing a program that will eventually be used on a computer with a physical parallel port and will need to set certain pins high or low in order to signal to an external device. However, the development laptop I am using does not have any physical parallel ports and is too low powered to run a virtual machine. Is there any option available that will create a virtual parallel port within Windows XP? Ideally it would include a debug mode that would allow me to see what values have been pushed out on the parallel port.

    Read the article

  • Parallel Computing Platform Developer Lab

    - by Daniel Moth
    This is an exciting announcement that I must share: "Microsoft Developer & Platform Evangelism, in collaboration with the Microsoft Parallel Computing Platform product team, is hosting a developer lab at the Platform Adoption Center on April 12-15, 2010.  This event is for Microsoft Partners and Customers seeking to incorporate either .NET Framework 4 or Visual C++ 2010 parallelism features into their new or existing applications, and to gain expertise with new Visual Studio 2010 tools including the Parallel Tasks and Parallel Stacks debugger toolwindows, and the Concurrency Visualizer in the profiler. Opportunities for attendees include: Gain expert design assistance with your Parallel Computing Platform based solution. Develop a solution prototype in collaboration with Microsoft Software Engineers. Attend topical presentations and “chalk-talk” sessions. Your team will be assigned private, secure offices for confidential collaboration activities. The event has limited capacity, thus enrollment is based on an application process.   Please download and complete the application form then return it to the event management team per instructions included within the form.  Applications will be evaluated based upon the technical solution scenario along with indicated project readiness timelines.  Microsoft event management team members may contact you directly for additional clarification and discussion of your project scenario during the nomination process." Comments about this post welcome at the original blog.

    Read the article

  • Unity: parallel vectors and cross product, how to compare vectors

    - by Heisenbug
    I read this post explaining a method to understand if the angle between 2 given vectors and the normal to the plane described by them, is clockwise or anticlockwise: public static AngleDir GetAngleDirection(Vector3 beginDir, Vector3 endDir, Vector3 upDir) { Vector3 cross = Vector3.Cross(beginDir, endDir); float dot = Vector3.Dot(cross, upDir); if (dot > 0.0f) return AngleDir.CLOCK; else if (dot < 0.0f) return AngleDir.ANTICLOCK; return AngleDir.PARALLEL; } After having used it a little bit, I think it's wrong. If I supply the same vector as input (beginDir equal to endDir), the cross product is zero, but the dot product is a little bit more than zero. I think that to fix that I can simply check if the cross product is zero, means that the 2 vectors are parallel, but my code doesn't work. I tried the following solution: Vector3 cross = Vector3.Cross(beginDir, endDir); if (cross == Vector.zero) return AngleDir.PARALLEL; And it doesn't work because comparison between Vector.zero and cross is always different from zero (even if cross is actually [0.0f, 0.0f, 0.0f]). I tried also this: Vector3 cross = Vector3.Cross(beginDir, endDir); if (cross.magnitude == 0.0f) return AngleDir.PARALLEL; it also fails because magnitude is slightly more than zero. So my question is: given 2 Vector3 in Unity, how to compare them? I need the elegant equivalent version of this: if (beginDir.x == endDir.x && beginDir.y == endDir.y && beginDir.z == endDir.z) return true;

    Read the article

  • Fireing Android Dialogs from another thread without Message Loop

    - by Jox
    In a SurfaceView, I'm dispatching new thread that draws on canvas within standard "LockCanvas-Draw-unlockCanvasAndPost" loop. (note that thread doesn't contains message loop). How to show Android standard Dialog from that thread? As thread doesn't have msg loop, following code doesn't work: Builder builder = new AlertDialog.Builder(this); builder.setTitle("Alert"); builder.setMessage("Stackoverflow!"); builder.setNegativeButton("cancel", null); builder.show();

    Read the article

  • Parallel EntityFramework

    - by mehanik
    Is it possible to make some work in parallel with entity framework for following example? using (var dbContext = new DB()) { var res = (from c in dbContext.Customers orderby c.Name select new { c.Id, c.Name, c.Role } ).ToDictionary(c => c.Id, c => new Dictionary<string, object> { { "Name",c.Name }, { "Role", c.Role } }); } For exampe what will be changed if I add AsParrallel? using (var dbContext = new DB()) { var res = (from c in dbContext.Customers orderby c.Name select new { c.Id, c.Name, c.Role } ).AsParallel().ToDictionary(c => c.Id, c => new Dictionary<string, object> { { "Name",c.Name }, { "Role", c.Role } }); }

    Read the article

  • What parallel programming model do you recommend today to take advantage of the manycore processors

    - by Doctor J
    If you were writing a new application from scratch today, and wanted it to scale to all the cores you could throw at it tomorrow, what parallel programming model/system/language/library would you choose? Why? I am particularly interested in answers along these axes: Programmer productivity / ease of use (can mortals successfully use it?) Target application domain (what problems is it (not) good at?) Concurrency style (does it support tasks, pipelines, data parallelism, messages...?) Maintainability / future-proofing (will anybody still be using it in 20 years?) Performance (how does it scale on what kinds of hardware?) I am being deliberately vauge on the nature of the application in anticipation of getting good general answers useful for a variety of applications.

    Read the article

  • PHP: Infinity loop and Time Limit!

    - by Jonathan
    Hi, I have a piece of code that fetches data by giving it an ID. If I give it an ID of 1230 for example, the code fetches an article data with an ID of 1230 from a web site (external) and insert it into a DB. Now, the problem is that I need to fetch all the articles, lets say from ID 00001 to 99999. If a do a 'for' loop, after 60 seconds the PHP internal time limit stops the loop. If a use some kind of header("Location: code.php?id=00001") or header("Location: code.php?id=".$ID) and increase $ID++ and then redirect to the same page the browser stops me because of the infinite loop or redirection problem. Please HELP!

    Read the article

  • Flash AS 3 Loader OnComplete Inside a Loop

    - by meengla
    Hi, I think I posted my question as an answer elsewhere (http://stackoverflow.com/questions/2338317/how-to-get-associated-urlrequest-from-event-complete-fired-by-urlloader/2776515#2776515) . Sorry. Here is my question again: Hi, How can I make your function work for loader object in a loop? Thanks! Meengla Here is my existing (rough) code; I always get the mylabel from the last element of the array. var _loader = new Loader(); for (j = 0; j < 5; j++) { //mylabel variable is correct setup in the loop _loader.contentLoaderInfo.addEventListener(Event.COMPLETE, function(e:Event):void { doneLoad(e, mylabel); }); _loader.load(new URLRequest(encodeURI(recAC[j].url))); }//for loop

    Read the article

  • Can Parallel.ForEach be used safely with CloudTableQuery

    - by knightpfhor
    I have a reasonable number of records in an Azure Table that I'm attempting to do some one time data encryption on. I thought that I could speed things up by using a Parallel.ForEach. Also because there are more than 1K records and I don't want to mess around with continuation tokens myself I'm using a CloudTableQuery to get my enumerator. My problem is that some of my records have been double encrypted and I realised that I'm not sure how thread safe the enumerator returned by CloudTableQuery.Execute() is. Has anyone else out there had any experience with this combination?

    Read the article

  • Nested loop with dependent bounds trip count

    - by aaa
    hello. just out of curiosity I tried to do the following, which turned out to be not so obvious to me; Suppose I have nested loops with runtime bounds, for example: t = 0 // trip count for l in 0:N for k in 0:N for j in max(l,k):N for i in k:j+1 t += 1 t is loop trip count is there a general algorithm/way (better than N^4 obviously) to calculate loop trip count? I am working on the assumption that the iteration bounds depend only on constant or previous loop variables.

    Read the article

  • Flash AS3 for loop query

    - by jimbo
    I was hoping for a way that I could save on code by creating a loop for a few lines of code. Let me explain a little, with-out loop: icon1.button.iconLoad.load(new URLRequest("icons/icon1.jpg")); icon2.button.iconLoad.load(new URLRequest("icons/icon2.jpg")); icon3.button.iconLoad.load(new URLRequest("icons/icon3.jpg")); icon4.button.iconLoad.load(new URLRequest("icons/icon4.jpg")); etc... But with a loop could I have something like: for (var i:uint = 0; i < 4; i++) { icon+i+.button.iconLoad.load(new URLRequest("icons/icon"+i+"jpg")); } Any ideas welcome...

    Read the article

  • for loop vs while loop

    - by Atul
    We can use for loop and while loop for same purpose. in what means they effect our code if I use for instead of while? same question arises between if-else and switch-case? how to decide what to use? for example which one you would prefer? This code: int main() { int n; cin>>n; for(int i=0;i<n;i++) { do_something(); } return 0; } Or this code: int main() { int n,i=0; cin>>n; while(i<n) { do_something(); i++; } return 0; } if using for or while loop does not effect the code by any means then may I know What was the need to make 2 solution for same problem?

    Read the article

  • Javascript Replace Child/Loop issue

    - by Charles John Thompson III
    I have this really bizarre issue where I have a forloop that is supposed to replace all divs with the class of "original" to text inputs with a class of "new". When I run the loop, it only replaces every-other div with an input, but if I run the loop to just replace the class of the div and not change the tag to input, it does every single div, and doesn't only do every-other. Here is my loop code, and a link to the live version: live version here function divChange() { var divs = document.getElementsByTagName("div"); for (var i=0; i<divs.length; i++) { if (divs[i].className == 'original') { var textInput = document.createElement('input'); textInput.className = 'new'; textInput.type = 'text'; textInput.value = divs[i].innerHTML; var parent = divs[i].parentNode; parent.replaceChild(textInput, divs[i]); } } }

    Read the article

  • can i changed for loop index variable inside the loop in Matlab???

    - by shawana
    hi every one. i need to change my loop variable inside the iterationa as ihave to acces array elements in the loop which is changing w.r.t size inside the loop. ashort piece of code is that: que=[]; que=[2,3,4]; global len; len=size(que,2) x=4; for i=1:len if x<=10 que(x)= 5; len=size(que,2) x=x+1; end end que array should print like: 2 3 4 5 5 5 5 5 5 5 but it is printed in such a way: 2 3 4 5 5 5. how shuould it be accomplished in matlab? in visual c++ it happen correctly and print whole array of 10 elements which increases at run time plz reply me if have any ideaabout this as i m new into matlab

    Read the article

  • jquery for loop on click image swap

    - by user2939914
    I've created a for loop of images. I would like each image to swap with another image on click individually. Here's the jQuery I've written so far: for ( var i = 1; i < 50; i++) { $('article').append('<div class="ps-block" id="' + i + '"><img src="img/bw/' + i + 'bw.png"></div>'); } $('img').click(function() { var imgid = $(this).attr('id'); $(this).attr("src", "img/color/" + imgid + ".png"); }); I also attempted to use this code inside the for loop after the append, but i ends up returning 50 every time you click since the loop has already ran: $('img[src="img/bw/' + i + 'bw.png"]').click(function() { $(this).attr("src", "img/color/" + this.id + ".png"); }); Thanks!

    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 | 4 5 6 7 8 9 10 11 12 13 14 15  | Next Page >