Search Results

Search found 68249 results on 2730 pages for 'sudo work'.

Page 282/2730 | < Previous Page | 278 279 280 281 282 283 284 285 286 287 288 289  | Next Page >

• Is there an equivalent of RDP?

  - by detly

  The "Desktop Sharing" settings that come installed by default seem to use VNC. VNC is a bit of a bandwidth hog, can only work at the resolution of whatever screen is attached to the host, and mirrors every action on the host. (It also seems to work poorly with compositing, but maybe that's been fixed.) I know about X tunnelling, but that's annoying to use and doesn't always work properly (or, more accurately, some apps don't work properly). Is there any kind of protocol in between the two, similar to RDP used for Windows? Specifically, something that can run at a different resolution to the host screen and is a little lighter on the network? (Ideally, the more the protocol could have in common with RDP, the better.)

  Read the article

• Fixing a bug while working on a different part of the code base

  - by imgx64

  This happened at least once to me. I'm working on some part of the code base and find a small bug in a different part, and the bug stops me from completing what I'm currently trying to do. Fixing the bug could be as simple as changing a single statement. What do you do in that situation? Fix the bug and commit it together with your current work Save your current work elsewhere, fix the bug in a separate commit, then continue your work [1] Continue what you're supposed to do, commit the code (even if it breaks the build fails some tests), then fix the bug (and the build make tests pass) in a separate commit [1] In practice, this would mean: clone the original repository elsewhere, fix the bug, commit/push the changes, pull the commit to the repository you're working on, merge the changes, and continue your work. Edit: I changed number three to reflect what I really meant.

  Read the article

• Ubuntu 12.04 + Bluetooth problems with maintaining connection

  - by ifndefx

  I have an Apple Wireless Mouse, Apple Wireless Keyboard, Apple Wireless Trackpad, and a pair of headphones. All of these connect to the computer via a Bluetooth dongle. After several hours I've managed to get the devices paired (all except the headphones) with the computer. However, I'm facing a problem where a connection can't be maintained even with one device. I first started with the gnome bluetooth applet, this couldn't even pick up devices, so I installed blueman, and this at least detected the devices but it couldn't keep the connection going. Then i read somewhere to get hidd installed and use the command line using the mac address of the device (which I got from blueman) and this worked really well, the connection was stable, but there's still a couple of issues that I need help with: If I am to use hidd I need to execute this via terminal, which means for the keyboard and the mouse I need to have two of each. This doesn't make sense. I need the bluetooth daemon to have started with the devices picked up a lot earlier, and especially at grub bootloader. The mouse and the trackpad both work if i use hidd, however, I cannot get the right mouse to work nor the scroll wheel to work. The headphones don't work period, they don't work with any of the bluetooth applets, and with hidd either. When I use blueman aplet, it attempts to pair and states 'Authentication Rejected". The headphone is Phillips SHB9100. If someone can help me with this, I would be grateful.

  Read the article

• Continuing Education as a Part of Your Job [closed]

  - by Mike

  I work as a programmer for a mid-sized company (about 500 employees) in the medical industry. Before that I worked at a custom software development/consulting company. At both companies programmers were never officially given time to continue their education through taking classes, reading books or blogs, or doing research relevant to the job. At the software development company we were offered some money to pay for a class, but not offered any time off of work to take the class. I have been wondering, do most employers of programmers give time off of work to take a class, read a book, or do job related research? By time off of work I just mean some period of time where you can stop development; it does not have to mean leaving the office. I would be grateful to hear about everyone's experience with this.

  Read the article

• What technologies are needed to be a freelance web developer / designer targeting small businesses?

  - by Jason Wirth

  First off, I'm not looking for this answer: "learn PHP/MySQL, JQuery, HTML/CSS...." My background, I wear many hats, and do many things. Currently I manage investment accounts with a business partner who is also a friend. He happens to be attending business / law school perusing a joint JD and MBA. As a result, we're putting our into "coast", maintaining our current clients, but not acquiring more. Recently I've picked up some freelance graphic design / web development / online marketing side work (Photoshop, HTML/CSS, WordPress) through some personal networks. The work is enjoyable; now I'm thinking about expanding it into a larger business with these primary goals: augment my finances (I'm shooting for about $1,000-2,000 per mo.), learn new technologies, involve myself with meaningful work. As an entrepreneur I don't mind the aspects of running a business that hassle other freelancers... -- the marketing -- the billing -- etc. I'm considering targeting the small to middle market businesses and organizations where I can contribute in marketing, design, and development building projects from the ground up. Thus I'll have freedom to decide the specific technology (I won't have to work with an existing code base). What kinds of projects should I focus on? What technologies are a good fit for this style of work? For example: It might be fun to develop with Ruby on Rails. However, maybe a lot of projects would be rolling out e-commerce solutions. Thus, I should focus on PHP due to more shopping cart options, skipping ROR entirely.

  Read the article

• When working with contractors/interns how protective should one be of your codebase?

  - by Shizam

  We're considering hiring a contractor or intern to work with us on our iOS project but this causes me to get really paranoid that we're giving somebody who doesn't work for the company access to our precious codebase. I mean, I could just give them access to the classes I'd like them to work on but that seems rude and it would make it more difficult for them to develop what we need them to. How paranoid should one be about people running off with their entire application?

  Read the article

• Ubuntu 13.10 - Wifi not re-connecting after suspend

  - by Matthew

  I upgraded the other day from Ubuntu 13.04 to 13.10 on my ZaReason laptop, and the WiFi doesn't reconnect after suspend. In 13.04, pressing Fn+F2 turned network connections back on after suspend. The 'soft' wireless hardware switch. But in 13.10, that doesn't work, and trying to enable networking in the networking menu doesn't work either. So I have to restart every time. Any suggestions to make it work?

  Read the article

• What non-computer based programming tools do you consider invaluable? [closed]

  - by Schroedinger

  Possible Duplicate: What physical tools do you find useful to work as a programmer? I'm talking about things like whiteboards for process planning, paper for mapping out logic flows, particular books that you've found relevant, things in your workspace that help you think and process what you're working on and how to attack problems. I'm starting out in a corporate environment and want to have an understanding of what tools really work for other programmers to help them work through their problems and solutions.

  Read the article

• Help connecting to Windows Shares on 2003/2008 server Ubuntu 12.04

  - by Murphy1138

  Can anyone help me , I'm really struggling to connect to Windows shares via Nautilus using the Connect to server command - anyone know the correct Syntax to use for user names and domain names or work groups? Its driving me nuts, I would fully use Ubuntu for work if I could make it connect to my Windows 2003/2008/ Windows 7 shares/ File servers seamlessly. How do you do it? I have been googling my a$%S off but can't work it out I have no issues with SSH or FTP

  Read the article

• How to make NFS mounts available while offline?

  - by lpanebr

  Problem: I work on a notebook and while at work I have access to many NFS mounted drives. When I get home they are obviously not available. Windows 7 solution: My business partner uses Windows 7 and maps the folders via samba. Windows 7 has a very nice feature that let's he make these folders available offline. So when when he connects to the work network the changes get synchronized! Question: Is there a way to mimic that in ubuntu? What I have now: Server to local sync: I have added rsync entries on my crontab to copy server folders => local folders every five minutes. When at work I used the NFS mapped folders and while outside work I use the local copies. When I get at work I manually run a script that syncs local folders => server folders. Problems with my setup: slow startup when not at work (I guess do to the fstab trying to map the server folders) no conflict checking/managing I have to remember to sync manually and be careful because of the different file locations recent files do not work between work and home

  Read the article

• Firefox released beta 9 for version 4

  - by anirudha

  Firefox yesterday released 9 beta of Firefox 4. now many plug-in work inside Firefox 4. How to get plugin work in Firefox 4 :- if you like any plug-in who you like to use but not worked in Firefox 4 go to developer site may be they work on them so it’s chance you can get the version who is in development from developer site. because many of plugin wait for a time period for testing or making same version for all plateform like linux or mac osX. How to work with stable and Beta version [standalone] sometime we fell that it’s time wasting because something we need the beta version can’t give. so you can use Firefox 3.6.13 or Firefox 4 both by a trick. you need to install them in other directory. when you trying to install another version beware because they also override the last installation directory so you need to choose the directory where you want to get another version manually.

  Read the article

• Changing the Operating System with only Ubuntu installed

  - by Games Brainiac

  I really wanted to dive into the world of Open Source operating systems, so I downloaded the latest version of Ubuntu (13.10), and installed it on a clean(no operating system installed, absolutely nothing) Lenovo ThinkPad machine. After a few days, I wanted to try out a different Operating System (Elementary OS). I downloaded the ISO file, burned it to a USB, tested that the USB booted from a different computer (I have 2, one is the Lenovo, the other a HP). I was able to get the bootscreen, and everything worked like a charm after I set the BIOS to boot from USB Disk Drive instead of HD. After this, I went back to Lenovo, and tried to open up the boot menu, by pressing F12, so that I could load from a temporary device. To my surprise, nothing but the HD was listed. There was no Optical Drive, No USB Drive, absolutely nothing. So, I thought that these devices were probably disabled. So I went into my BIOS and checked to see what was the case. I saw that all my devices were enabled. USB and all the other devices such as network cable and the rest were all enabled. So, I thought this probably had something to do wit UEFI and Legacy Boot options. So, I made sure that both were enabled. This did not solve the problem either. Again, I got nothing but the option to boot from my Hard Disk. I thought the USB had to be at fault. I tried different ports, but to no avail. Next, I tried with a Live CD, which had Ubuntu on it. This failed too. I simply could not boot from anything other than my hard disk. Okay, so at this point, I was pretty desperate, so I installed Boot-Repair through: sudo add-apt-repository ppa:yannubuntu/boot-repair sudo apt-get update sudo apt-get install boot-repair What this did is lead me to GRUB. Ideally, its just a screen that gives me the option to load from Ubuntu or Advanced Settings. The Advanced settings had nothing but Ubuntu options in it. So, I kept on pressing ESC and that led me to the the grub console, and thats where I am right now with my Lenovo. I've also tried updating the BIOS, but Lenovo only has packages for Red Hat and Windows. So, a dead end there too. Right now, I need to know if there is any way that I can just delete everything from my Lenovo? I want to revert it back to its blank factory condition. How can I achieve this? I have tried to elaborate my problem as best I could. If there is any important information that I've missed out, please do not hesitate to leave a comment. I would have included some screen shots, but BIOS screen shots are a little hard to manage. However, I can provide a camera Image of the boot screen if needed (doing that as we speak).

  Read the article

• How to create a custom keyboardlayout that works in all applications

  - by rubo77

  I created a custom keyboardlayout Coffee++ for Linux that works fine in most apps: http://coffeeplusplus.z11.de/phpBB3/viewtopic.php?f=6&t=13&p=34#p34 I created a debian package, that adds the needed changes to the keyboard configuration and creates a shell script which which you can swap between normal QWERTY and Coffee++ The letters all work fine, only the arrow-combinations, for ex. AltGr+J for "Arrow-left" and so on don't work in all applications. How can I make the Arrow-key-combinations work in all apps?

  Read the article

• Team Development Using Subversion - Best Practices

  Learn how to work with Subversion and JDeveloper

  Read the article

• IEEE SRS documents: lightweight version when working with outside contractors?

  - by maple_shaft

  Typically we follow an Agile development process that tends not to put an emphasis on writing requirements and technical documents that nobody will read. We tend to focus our limited manpower to development and testing activities with collaborative design and whiteboarding as a key focus. There is a mostly standalone web component that will take quite a few weeks to develop, but this work can be mostly parallel with other project work going on. To try and catch up time I was given a budget for hiring a developer on oDesk to complete this work. While my team isn't accustomed to working off of a firm SRS document, I realize that with outsourced development that it is a good idea to be as firm and specific as possible so I realize that I need to provide a detailed Requirements and Technical Specification document for this work to be done correctly. When I do write a Requirements document I typically utilize the standard IEEE SRS document template but I think this is too verbose and probably overkill for what I need to communicate to a developer. Is there another requirements document that is more lightweight and also accepted by a major standards organization like the IEEE? Further, as what will be developed as a software module that will interact with other software modules, my requirements really need to delve into technical specifications for things to work correctly. In this scenario does it make sense to merge technical and requirements specifications into a single document, and if not, what is a viable alternative?

  Read the article

• New HP dm4 - No audio on ubuntu 11.10 64bits

  - by Haze1

  I just got a new laptop, HP dm4, and I'm having problems getting the audio to work properly on it. http://www.alsa-project.org/db/?f=7b697a35465a9f7236fb94deb9ff97fa65e55489 I tried to edit /etc/modprobe.d/alsa-base.conf and added: option snd-hda-intel model=ref this caused the audio to work, but it's muffled. I'm wondering if anybody knows what would be the correct options to get this POS to work. Thanks in advance

  Read the article

• How to maintain Motivation and enthusiasm once you have figured out the solution needed

  - by Pocket_Pie

  I am currently undertaking a software project on my own time. When I first got the project I put in many hours working out how to do the "tricky" parts of the solution. I spent many hours googling and reading up on classes available on MSDN that I could use for the project. I was madly excited and passionate about doing this work. However once, I got a working samples of how I could get around the "tricky" parts and got to the part where all that I needed to do was "grunt" work to finish the project, I lost all interest and desire to work on he project. Suddenly instead of looking forward to sitting down and working on this project it became a chore and a major hassle to motivate myself. I am now fast approaching the deadline and I am getting the work done now, but it is under very high pressure as I have left it almost too close to the deadline! I will manage to get it done but it will involve several all-nighters. (BTW I completely despise doing these all-nighters and would love to eliminate these by maintaining my motivation and working at the project continuously.) So my questions are is this normal? Does everyone else notice such spikes and troughs in their enthusiasm for projects? Anyone more experienced have any advice on how to keep the motivation going? Or am I just not designed to work on a full project lifecycle, should i and people like me being doing an R&D type role where I can do the fun figuring out part of the projects and leave it for someone else to finish the "les interesting/mundane" coding?

  Read the article

• Can I properly use NS2 on latest Ubuntu 12.04?

  - by Nwe Nwe

  I am a novice user for Ubuntu because I need to work with NS2. NS2 properly work only in Linux, Ubuntu. Now, I wonder if it can properly work on latest version Ubuntu. My current window is Window7. I also want to use Window7. So, which one will be better on parallel booting of Window 7 and Ubuntu or using VMWare on Window7 to install Ubuntu. My RAM is 4G. My processor speed is 1.6G. If I use VMWare, my C drive is only extra 10G. So, does VMWare can work properly in D drive? Please let me know which way is the best for my situation. I hope the advice from anyone. Thanks in advance.

  Read the article

• How do you assign resources and keep begin, end and duration of a task intact?

  - by Random

  I have problems with assigning more than one resource to a group of tasks. The idea is simple, my tasks are in one group and are manually scheduled to particular begin and end dates. I want to assign more than one resource to keep task duration and dates (fixed duration) and increase work. For top level tasks it works fine but as long tasks are grouped, the duration of each is extended to reach group end date and work remains. For the problematic tasks, the Gantt chart looks like this: One resource attached (good) ( Task 1.1 ) ( Task 1.2 ) (Task 1.3) More than one resource attached (wrong) ( Task 1.1 )....................... ( Task 1.2 ).......... (Task 1.3) So for tasks like that, I want to have a fixed schedule and just increase work by adding resources that work in the same time, but sometimes MS Project does leveling to do resources work sequentialy.

  Read the article

• No audio on an HP dm4

  - by Haze1

  I just got a new laptop, HP dm4, and I'm having problems getting the audio to work properly on it. http://www.alsa-project.org/db/?f=7b697a35465a9f7236fb94deb9ff97fa65e55489 I tried to edit /etc/modprobe.d/alsa-base.conf and added: option snd-hda-intel model=ref this caused the audio to work, but it's muffled. I'm wondering if anybody knows what would be the correct options to get this POS to work.

  Read the article

• What would be your thoughts in this situation?

  - by craftsman

  Imagine there's a businessman who has a solid internet idea. He hires you as his first employee and pays you well (also offers you shares if product generates revenue). You code the product for more than a year. But the product is not properly marketed and left almost unused. You get frustrated by loneliness and working endlessly on an unused product. You start applying at different companies. You don't get response from anyone. After a month, you realize that things are beginning to change. The businessman has started to give more attention to the product. You have defined a solid marketing plan with him. Things are almost sure to work. Suddenly one day, you get a call from a big company you yearned to work for. They invite you for an interview. If you respond to the interview and get the job, you will lose the chance to earn something from your one year hard-work. If you don't respond, you will obviously miss the chance to work in the big company. If your marketing doesn't work as well (you will know it in a couple of months), you will probably not get a chance in the big company again. What would be your decision?

  Read the article

• Code Completion -- Aptana Eclipse Plugin

  - by jwmajors81

  I have been doing javascript development for the last couple weeks and have tried JSDT and Aptana to assist in code completion. JSDT wasn't very good at all, but I did have more luck with Aptana (used as eclipse plug-in, not the standalone product). The problem I'm encountering is that when I create javascript classes I cannot get code completion to work. For example, if I use the following then code completion doesn't work: var foo = new function(value){ this.myMethod= function(){ } } I have also verified that the following won't work: function foo(value){ this.myMethod= function(){ } } I have found that using a JSON style does work: var foo = { myMethod: function(){ } } Does anyone know why Aptana works for the last style, but not the first? Using the JSON style won't work for me because I have to have seperate instances of the class in question. Also, I am not very successful in getting code completion to work across files. For example, if I have 3 files in the javascript directory then I usually cannot get Aptana to pick up the JSON style markup in the other two classes. This DID work at one point (for the first 2 classes I created), but since then whenever I add new classes they aren't picked up. Thank you very much for you assistance. Jeremy

  Read the article

• Error installing geoip_city gem

  - by tomeara

  I keep getting an error when trying to install the geoip_city gem. I've already installed the GeoIP C library to /opt/GeoIP, but the gem doesn't seem to pick it up. I've tried: sudo gem install geoip_city -- --with-geoip-dir=/opt/GeoIP sudo gem install geoip_city -- --with-geoip-lib=/opt/GeoIP/lib sudo gem install geoip_city -- --with-geoip-dir=/opt/GeoIP --with-geoip-lib=/opt/GeoIP/lib all of which output this error: Building native extensions. This could take a while... ERROR: Error installing geoip_city: ERROR: Failed to build gem native extension. /System/Library/Frameworks/Ruby.framework/Versions/1.8/usr/bin/ruby extconf.rb --with-geoip-lib=/opt/GeoIP/lib checking for GeoIP_record_by_ipnum() in -lGeoIP... no you must have geoip c library installed! *** extconf.rb failed *** Could not create Makefile due to some reason, probably lack of necessary libraries and/or headers. Check the mkmf.log file for more details. You may need configuration options. Provided configuration options: --with-opt-dir --without-opt-dir --with-opt-include --without-opt-include=${opt-dir}/include --with-opt-lib --without-opt-lib=${opt-dir}/lib --with-make-prog --without-make-prog --srcdir=. --curdir --ruby=/System/Library/Frameworks/Ruby.framework/Versions/1.8/usr/bin/ruby --with-geoip-dir --without-geoip-dir --with-geoip-include --without-geoip-include=${geoip-dir}/include --with-geoip-lib=${geoip-dir}/lib --with-GeoIPlib --without-GeoIPlib Gem files will remain installed in /Library/Ruby/Gems/1.8/gems/geoip_city-0.2.0 for inspection. Results logged to /Library/Ruby/Gems/1.8/gems/geoip_city-0.2.0/gem_make.out

  Read the article

• Django Querying Relation of Relation

  - by Brent

  I'm stuck on a Django ORM issue that is bugging me. I have a set of models linked by a foreign key but the requirements are a bit odd. I need to list items by their relation's relation. This is hard to explain so I've tried to depict this below, given: Work ManyToMany(Award) Award ForeignKey(AwardCategory) AwardCategory I need to list work items so they are listed by the award category. Desired output would be: Work Instance A Award Instance A that belongs to Award Category Instance A Award Instance C that belongs to Award Category Instance A Award Instance G that belongs to Award Category Instance A Work Instance A (same instance as above, but listed by different award__category) Award Instance F that belongs to Award Category Instance B Award Instance R that belongs to Award Category Instance B Award Instance Z that belongs to Award Category Instance B Work Instance B Award Instance B that belongs to Award Category Instance A Award Instance A that belongs to Award Category Instance A Essentially I want to list all work by the award category. I can get this to work in part but my solution is filthy and gross. I'm wondering if there is a better way. I considered using a ManyToMany and a through attribute but I'm not certain if I'm utilizing it correctly.

  Read the article

• Query on simple C++ threadpool implementation

  - by ticketman

  Stackoverflow has been a tremendous help to me and I'd to give something back to the community. I have been implementing a simple threadpool using the tinythread C++ portable thread library, using what I have learnt from Stackoverflow. I am new to thread programming, so not that comfortable with mutexes, etc. I have a question best asked after presenting the code (which runs quite well under Linux): // ThreadPool.h class ThreadPool { public: ThreadPool(); ~ThreadPool(); // Creates a pool of threads and gets them ready to be used void CreateThreads(int numOfThreads); // Assigns a job to a thread in the pool, but doesn't start the job // Each SubmitJob call will use up one thread of the pool. // This operation can only be undone by calling StartJobs and // then waiting for the jobs to complete. On completion, // new jobs may be submitted. void SubmitJob( void (*workFunc)(void *), void *workData ); // Begins execution of all the jobs in the pool. void StartJobs(); // Waits until all jobs have completed. // The wait will block the caller. // On completion, new jobs may be submitted. void WaitForJobsToComplete(); private: enum typeOfWorkEnum { e_work, e_quit }; class ThreadData { public: bool ready; // thread has been created and is ready for work bool haveWorkToDo; typeOfWorkEnum typeOfWork; // Pointer to the work function each thread has to call. void (*workFunc)(void *); // Pointer to work data void *workData; ThreadData() : ready(false), haveWorkToDo(false) { }; }; struct ThreadArgStruct { ThreadPool *threadPoolInstance; int threadId; }; // Data for each thread ThreadData *m_ThreadData; ThreadPool(ThreadPool const&); // copy ctor hidden ThreadPool& operator=(ThreadPool const&); // assign op. hidden // Static function that provides the function pointer that a thread can call // By including the ThreadPool instance in the void * parameter, // we can use it to access other data and methods in the ThreadPool instance. static void ThreadFuncWrapper(void *arg) { ThreadArgStruct *threadArg = static_cast<ThreadArgStruct *>(arg); threadArg->threadPoolInstance->ThreadFunc(threadArg->threadId); } // The function each thread calls void ThreadFunc( int threadId ); // Called by the thread pool destructor void DestroyThreadPool(); // Total number of threads available // (fixed on creation of thread pool) int m_numOfThreads; int m_NumOfThreadsDoingWork; int m_NumOfThreadsGivenJobs; // List of threads std::vector<tthread::thread *> m_ThreadList; // Condition variable to signal each thread has been created and executing tthread::mutex m_ThreadReady_mutex; tthread::condition_variable m_ThreadReady_condvar; // Condition variable to signal each thread to start work tthread::mutex m_WorkToDo_mutex; tthread::condition_variable m_WorkToDo_condvar; // Condition variable to signal the main thread that // all threads in the pool have completed their work tthread::mutex m_WorkCompleted_mutex; tthread::condition_variable m_WorkCompleted_condvar; }; cpp file: // // ThreadPool.cpp // #include "ThreadPool.h" // This is the thread function for each thread. // All threads remain in this function until // they are asked to quit, which only happens // when terminating the thread pool. void ThreadPool::ThreadFunc( int threadId ) { ThreadData *myThreadData = &m_ThreadData[threadId]; std::cout << "Hello world: Thread " << threadId << std::endl; // Signal that this thread is ready m_ThreadReady_mutex.lock(); myThreadData->ready = true; m_ThreadReady_condvar.notify_one(); // notify the main thread m_ThreadReady_mutex.unlock(); while(true) { //tthread::lock_guard<tthread::mutex> guard(m); m_WorkToDo_mutex.lock(); while(!myThreadData->haveWorkToDo) // check for work to do m_WorkToDo_condvar.wait(m_WorkToDo_mutex); // if no work, wait here myThreadData->haveWorkToDo = false; // need to do this before unlocking the mutex m_WorkToDo_mutex.unlock(); // Do the work switch(myThreadData->typeOfWork) { case e_work: std::cout << "Thread " << threadId << ": Woken with work to do\n"; // Do work myThreadData->workFunc(myThreadData->workData); std::cout << "#Thread " << threadId << ": Work is completed\n"; break; case e_quit: std::cout << "Thread " << threadId << ": Asked to quit\n"; return; // ends the thread } // Now to signal the main thread that my work is completed m_WorkCompleted_mutex.lock(); m_NumOfThreadsDoingWork--; // Unsure if this 'if' would make the program more efficient // if(NumOfThreadsDoingWork == 0) m_WorkCompleted_condvar.notify_one(); // notify the main thread m_WorkCompleted_mutex.unlock(); } } ThreadPool::ThreadPool() { m_numOfThreads = 0; m_NumOfThreadsDoingWork = 0; m_NumOfThreadsGivenJobs = 0; } ThreadPool::~ThreadPool() { if(m_numOfThreads) { DestroyThreadPool(); delete [] m_ThreadData; } } void ThreadPool::CreateThreads(int numOfThreads) { // Check a thread pool has already been created if(m_numOfThreads > 0) return; m_NumOfThreadsGivenJobs = 0; m_NumOfThreadsDoingWork = 0; m_numOfThreads = numOfThreads; m_ThreadData = new ThreadData[m_numOfThreads]; ThreadArgStruct threadArg; for(int i=0; i<m_numOfThreads; ++i) { threadArg.threadId = i; threadArg.threadPoolInstance = this; // Creates the thread and save in a list so we can destroy it later m_ThreadList.push_back( new tthread::thread( ThreadFuncWrapper, (void *)&threadArg ) ); // It takes a little time for a thread to get established. // Best wait until it gets established before creating the next thread. m_ThreadReady_mutex.lock(); while(!m_ThreadData[i].ready) // Check if thread is ready m_ThreadReady_condvar.wait(m_ThreadReady_mutex); // If not, wait here m_ThreadReady_mutex.unlock(); } } // Adds a job to the batch, but doesn't start the job void ThreadPool::SubmitJob(void (*workFunc)(void *), void *workData) { // Check that the thread pool has been created if(!m_numOfThreads) return; if(m_NumOfThreadsGivenJobs >= m_numOfThreads) return; m_ThreadData[m_NumOfThreadsGivenJobs].workFunc = workFunc; m_ThreadData[m_NumOfThreadsGivenJobs].workData = workData; std::cout << "Submitted job " << m_NumOfThreadsGivenJobs << std::endl; m_NumOfThreadsGivenJobs++; } void ThreadPool::StartJobs() { // Check that the thread pool has been created // and some jobs have been assigned if(!m_numOfThreads || !m_NumOfThreadsGivenJobs) return; // Set 'haveworkToDo' flag for all threads m_WorkToDo_mutex.lock(); for(int i=0; i<m_NumOfThreadsGivenJobs; ++i) m_ThreadData[i].haveWorkToDo = true; m_NumOfThreadsDoingWork = m_NumOfThreadsGivenJobs; // Reset this counter so we can resubmit jobs later m_NumOfThreadsGivenJobs = 0; // Notify all threads they have work to do m_WorkToDo_condvar.notify_all(); m_WorkToDo_mutex.unlock(); } void ThreadPool::WaitForJobsToComplete() { // Check that a thread pool has been created if(!m_numOfThreads) return; m_WorkCompleted_mutex.lock(); while(m_NumOfThreadsDoingWork > 0) // Check if all threads have completed their work m_WorkCompleted_condvar.wait(m_WorkCompleted_mutex); // If not, wait here m_WorkCompleted_mutex.unlock(); } void ThreadPool::DestroyThreadPool() { std::cout << "Ask threads to quit\n"; m_WorkToDo_mutex.lock(); for(int i=0; i<m_numOfThreads; ++i) { m_ThreadData[i].haveWorkToDo = true; m_ThreadData[i].typeOfWork = e_quit; } m_WorkToDo_condvar.notify_all(); m_WorkToDo_mutex.unlock(); // As each thread terminates, catch them here for(int i=0; i<m_numOfThreads; ++i) { tthread::thread *t = m_ThreadList[i]; // Wait for thread to complete t->join(); } m_numOfThreads = 0; } Example of usage: (this calculates pi-squared/6) struct CalculationDataStruct { int inputVal; double outputVal; }; void LongCalculation( void *theSums ) { CalculationDataStruct *sums = (CalculationDataStruct *)theSums; int terms = sums->inputVal; double sum; for(int i=1; i<terms; i++) sum += 1.0/( double(i)*double(i) ); sums->outputVal = sum; } int main(int argc, char** argv) { int numThreads = 10; // Create pool ThreadPool threadPool; threadPool.CreateThreads(numThreads); // Create thread workspace CalculationDataStruct sums[numThreads]; // Set up jobs for(int i=0; i<numThreads; i++) { sums[i].inputVal = 3000*(i+1); threadPool.SubmitJob(LongCalculation, &sums[i]); } // Run the jobs threadPool.StartJobs(); threadPool.WaitForJobsToComplete(); // Print results for(int i=0; i<numThreads; i++) std::cout << "Sum of " << sums[i].inputVal << " terms is " << sums[i].outputVal << std::endl; return 0; } Question: In the ThreadPool::ThreadFunc method, would better performance be obtained if the following if statement if(NumOfThreadsDoingWork == 0) was included? Also, I'd be grateful of criticisms and ways to improve the code. At the same time, I hope the code is of use to others.

  Read the article

< Previous Page | 278 279 280 281 282 283 284 285 286 287 288 289  | Next Page >