Search Results

Search found 31409 results on 1257 pages for 'app offline htm'.

Page 658/1257 | < Previous Page | 654 655 656 657 658 659 660 661 662 663 664 665  | Next Page >

  • Launching a URL from an OOB Silverlight Application

    So I'm working on this code browser mostly to help me with my fading memory. In the app I want to be able to launch a url. So I do my normal thing and put a line of code that looks like this:System.Windows.Browser.HtmlPage.Window.Navigate(new Uri(ThisURI), "_blank");I was agast when I realized that this didn't work, for that matter it didn't even blow... grr... but with a bit of research I found that the hyper link button worked so a ended up making a little class like this:public class MyHyperLink...Did you know that DotNetSlackers also publishes .net articles written by top known .net Authors? We already have over 80 articles in several categories including Silverlight. Take a look: here.

    Read the article

  • Native, FOSS GUI prototyping tools?

    - by Oli
    As part of my job as a web developer, I spend an amount of time doing UI prototypes to show the client. It's a pain in the behind but sometimes it has to be done. I've seen Shuttleworth (and the design team) pump out images like this: That's made by something called Balsamiq Mockups... Something that balances on top of Adobe Air (yack!) and costs $79. I've tried it but it kept falling over. I think it had something to do with Air not the app itself. My point is if I'm paying out for something, I want it to be native.

    Read the article

  • How to create a "shutdown user" or "shutdown account"

    - by pcapademic
    Red Hat had a feature useful to me at the present time. There was an account, generally called "shutdown", and when you logged in with the account, the system shut down. In my specific case, I have Ubuntu Server running in a VM on my local system. The VM is running a web app, and when I'm done doing work, I want to shut down the VM. Unfortunately, I can't install VMware tools to get the "power button" based shutdown. Currently I login then sudo shutdown -h now, then type my password again, and things shutdown. Really, it's getting annoying all that waiting around and typing things. How do I replicate the "shutdown account" functionality in Ubuntu? A related question, were there any security gotchas that motivated people to stop using this kind of account?

    Read the article

  • Switch interface implementation using configuration

    - by Marcos
    We want to allow the same core service to be either fully implemented or, as other option, to be a proxy toward a client legacy system (via a WSDL for example). In that way, we have both implementation (proxy & full) and we switch which one to use in the configuration of the app. So in a nutshell, Some desired features: Two different implementation (proxy, full) instead of one implementation with a switch inside Switch implementation using configuration: dependency injection? reflection? Nice-to-have: the packaged delivered to the client doesn’t have to change depending on the choice between proxy or full Nice-to-have: Client can develop their custom implementation of the Core Interface and configure the applciation to use that one With this background, the question is: What alternatives we have to choose one implementation or other of an interface just changing configuration? Thanks

    Read the article

  • Bing Desktop Automatically Downloads Bing Wallpapers to Your Computer

    - by Jason Fitzpatrick
    Windows 7: Bing Desktop is a new and lightweight offering from Microsoft that automatically swaps your desktop background every day and offers quick access to the Bing search engine. In addition to downloading the Bing wallpaper, Bing Desktop also includes a small search box that allows you to search Bing from your desktop–although most users will likely grab the app simply to get the daily wallpaper update. Hit up the link below to download a copy. Bing Desktop is free, Windows 7 only. Bing Desktop [via Quick Online Tips] How to Stress Test the Hard Drives in Your PC or Server How To Customize Your Android Lock Screen with WidgetLocker The Best Free Portable Apps for Your Flash Drive Toolkit

    Read the article

  • Google+ Platform Office Hours for March 28, 2012: Hangouts API v1.0

    Google+ Platform Office Hours for March 28, 2012: Hangouts API v1.0 Here's another video from a previous session of our office hours. Watch this video to learn about the Hangouts Apps launch from +Wolff and +Jonathan. Discuss this video on Google+: goo.gl 3:31 - Publishing your hangout app 4:28 - Hangout applications vs extensions 8:00 - The application switcher 9:58 - On the terms of service, privacy policy and support contact fields 12:07 - OAuth client and hangout apps featuring the API console 15:50 - Registering as a Chrome web store developer 17:44 - Linking to your hangout 20:25 - The hangout button 24:33 - How data URIs can make things easier in your apps Q&A 29:00 - What's the status of the REST APIs? 30:41 - How do I set the hangout topic or title? 31:19 - How do those of us in other time zones know when your office hours will be held? 34:04 - Can I use the hangout button with other peoples' hangout apps? From: GoogleDevelopers Views: 2788 28 ratings Time: 35:18 More in Science & Technology

    Read the article

  • tech-ed 2012

    - by foxjazz
    So, am not going to tech-ed this year.I didn't get much benefit from going last year, but I did meet a lot of nice folks.I am working on my first official Silverlight project, and it's going ok.I having a few issues which I may resolve with WCF services.I am still green around the edges with this technology, but I am getting the hang of it slowly.Learning a lot about IQueryable and how to handle databases.Depending on what I am looking todo, I may use some messaging services within the app.It has been a hard study the last month. Learning SL, JQuery, More CSS and website work, code-first.Node.js, SingalRThere seems to be a lot to do to keep up with the technology.Hope to post more often, but am hammering on something new, most of the time.

    Read the article

  • Copying content on webpages in safari. To HTML

    - by Carl Smith
    Hi, is there an easier way to copy and paste website content in html? Want to copy and look like this. Product Information: Length: S / M / L Material: Polyester and Elasthane Brand: Roxana Exclusive Style: Basque But when i paste it into my content box it looks like this- Product Information Length: S / M / L Material: Polyester and Elasthane Brand: Roxana Exclusive Style: Basque Then i need to edit it in the html editor to rearrange it. Is the some sort of app or plugin that i can get so i can turn the text of the page into html so it looks right straight away when i copy it into my content box? If that makes any sense? Thanks Carl Smith :-)

    Read the article

  • Different behaviour with windows authentication on IIS7 websites

    - by amaters
    I need to run a website with just windows authentication. Given the following situation: The location of the default website is: c:\inetpub\wwwroot The location of my code is: c:\Sites\WebApp my hostfile is edited so any .local i use points to 127.0.0.1 I have created a new application called 'AppX' underneath the default website and point it to c:\Sites\WebApp. It will use the DefaultappPool. When I switch off anonymous and switch on windows authentication all works well when I go to localhost/AppX/. What i really want is a new website (No need to question why I want this). So I created Website2 and did exact the same creation of the application. Everything is the same; destination, app pool and authentication. Now when I browse to this website web2.local/AppX/ I get the 401.2 - Unauthorized error. What am I missing here?

    Read the article

  • Any Good Cocos2d Pause Menu Library

    - by Mahbubur R Aaman
    Background : From http://code.google.com/p/cocos2d-iphone/issues/detail?id=173 Scenes/Nodes doesn't support the CocosNodeOpacity protocol. From http://playsnackgames.com/blog/2011/09/cocos2d-tutorial-creating-a-reusable-pause-layer/ Cocos2d offers a simple method to pause and resume itself, but these methods stop the CCDirector (the class that manages most aspects of a Cocos2d’s app lifetime) from running actions and lower the fps to 5 to conserve battery life. Related issues http://www.cocos2d-iphone.org/forum/topic/4368 http://www.cocos2d-iphone.org/forum/topic/151 http://stackoverflow.com/questions/5852354/cocos2d-engine-pause-resume http://stackoverflow.com/questions/11878450/how-to-pause-a-layer-in-cocos2d-2-0 Question : Is there any Good Cocos2d Pause Menu Library solving these tricky issues? This will save many hours of Game Developer's life.

    Read the article

  • Applicability of the Joel Test to web development companies

    - by dreftymac
    QUESTION: How can you re-write the questions of the Joel test to apply to web developers? 1. Do you use source control? (source control for all aspects of your app, including configuration, database and user-based settings?) 2. Can you make a build in one step? (can you deploy a site from staging to prod in 1 step?) ... 10. Do you have testers? (how do you test AJAX and CSS?) BACKGROUND: This is for people who work in a shop that does some web development but also uses some off-the-shelf tools like Drupal and Wordpress, but doing custom development on top of that. RELATED LINKS: http://www.joelonsoftware.com/articles/fog0000000043.html What do you think about the Joel Test?

    Read the article

  • How to setup Thinkpad features on Thinkpad T500

    - by gijoemike
    I have a IBM-Lenovo Thinkpad T500. I was previously a exclusive windows user, but recently installed ubuntu and loving it because of speed and interface. The only thing is that I don't get some features that I came to enjoy in windows. I need help setting these up: Hard-drive protection - active protection software that pauses drive when there is movement My printer doesn't work (can't find the driver for this one): canon Ip2600 A way to change which graphics chip to use while in OS. I have both the integrated and non integrated (dual-graphics). (If not easy to setup, I know there's a way to do it before it boots, but don't know how). CPU performance level - in windows you can pick "high performance", "power saver", etc.. to save batteries. My integrated camera w/light - it works but need an app where I can record videos, take snapshots, etc. can't find one that works. Thanks!

    Read the article

  • Dash is slow and shows irrelevant results

    - by Alexey Frishman
    I currently have the latest Ubuntu 12.10 installed on my laptop. Usually I use Launchy application to have a quick access to any app/config/file etc. Now I'm trying to get used to Dash, which is supposed to be default way to do such things in recent Ubuntu versions. The difference between the usage of Launchy and Dash is following: Launchy: Alt+Space - Launchy shell shown instantly - type your request - open the target Dash: SuperKey - PERIOD - Dash is shown - type your request - PERIOD - navigate with arrow buttons between the results - open the desired result Another problem. When I type the term "ryth" (which is incorrectly spelled part of "Rhythmbox") what is shown in these 2 shells: Launchy: 1 result, which is Rhythmbox. The letters 'r', 'y', 't' and 'h' are highlighted. Dash: 2 results, which are MP3s from Amazon and are completely irrelevant to my request So is there any way to tweak the Dash to allow me to use it as I use Launchy with the same performance and results?

    Read the article

  • Login screen restarts while entering password

    - by Shane L
    I am having a problem that only occurred after installing the fglrx proprietary driver through the additional drivers app. The exact same issue is described in this question, however it's closed. Must login twice before entering Unity; first login screen has graphical anomalies When I boot up my computer, once I get to lightdm login screen I will start typing my password, and upon the entering of two of my password characters "n2" the screen will start displaying some corruption in the very top edge of the screen, then after I hit the number 2, lightdm seems to crash and it will restart and everything is fine, I can login from that point. So the issue is, every time I reboot, I have to login once, allow lightdm to crash and restart, then login as normal. The issue did not occur prior to installing fglrx, and has happened through several reinstalls.

    Read the article

  • PHP Development Environment (Host: Windows 7, Guest: Ubuntu)

    - by Kristian Leiws Jones
    Since editing files live from a remote server slows down development. I use XAMPP on windows to develop then run the web app's on a Linux server. However to avoid environment dependencies I'd like to mirror the live environment and the development environments. What I'm asking is running development server on Ubuntu inside VirtualBox whilst editing the source files via ftp/Dreamweaver is a good idea? If so, and I wanted to view the local website on the host OS (windows) how would I do this? does the guest OS have a LAN/Local IP address? I notice on windows "ipconfig /all" there are "tunneling" adapters which I assume is for VirtualBox, so I guess the guest OS has the same LAN/Local IP address? if so how would I view the websites hosted on the guest OS on the host OS? I'd also need to host FTP server on guest OS. Note: I need windows! I would love to use Linux all the way -.-

    Read the article

  • Bring Gadgets back to Your Desktop in Windows 8 RTM with 8GadgetPack

    - by Asian Angel
    Are you someone who loved using desktop gadgets in Windows 7 and Vista, but felt disappointed when learning they were removed in Windows 8 RTM? Then 8GadgetPack is just the app to put those gadgets back on your desktop! The good folks over at 7 Tutorials have a nice little write-up about 8GadgetPack with all the details you need to get those gadgets up and running once again. Just browse on over using the link below… How to Use Desktop Gadgets in Windows 8 with 8GadgetPack [7 Tutorials] How To Create a Customized Windows 7 Installation Disc With Integrated Updates How to Get Pro Features in Windows Home Versions with Third Party Tools HTG Explains: Is ReadyBoost Worth Using?

    Read the article

  • Url Navigation

    - by russ.bishop
    One of the new features is URL-based navigation which is useful for creating intranet links or auto-generating email links (such as from workflow systems, etc). For IIS 6 and earlier, the format is as follows: http://machine/drm-client/Logon.aspx? app=<appname>&action=go&ver=<version name>&hier=<hier name>&node=<node name> Just replace the fields with their appropriate values (URL-encoded of course). <node name> is optional. If provided it will open the hierarchy and expand directly to the target node. Otherwise the hierarchy is opened to the top node. Note that if the specified version is not loaded it will be loaded automatically.

    Read the article

  • Understanding and developing web services

    - by Pankaj Upadhyay
    This question is in conjuction with How would you approach developing a Hotel Reservation System? The solution to a system with different interfaces(or clients i should say) is to go with developing a Web service and have other systems interact with it. I never had the requirement for developing a Web service so i am bit short on it. All i understand is that A web service is a system or application that performs some operations which may include modifying, sending or receiving data over a network using HTTP protocol. (Let me know if the understanding is wrong) Now, from the other question it's clearly understood that i need to develop a web service but i have no idea as to how should i go about it. My language of choice is C# and .NET Framework. Question:: How do we develop a webservice and which tools,technology and framework should i use for the same using C# language?? Question:: How can i interact with this from a desktop WPF application, Website and Mobile app

    Read the article

  • SilverlightShow for Dec 27-Jan 2, 2011

    - by Dave Campbell
    Check out the Top Five most popular news at SilverlightShow for Dec 27-Jan 2, 2011. The most visited news for last week is Mahesh Sabnis's post on how to use Prism in Silverlight 4. Among the top 5 news is also the announcement for SilverlightShow December Newsletter that you can now read online. Here is SilverlightShow's weekly top 5: Using Prism with Silverlight 4 "What's new in Silverlight 4 demo" app Cinch - A Rich Full Featured WPF/SL MVVM Framework SilverlightShow December Newsletter Now Online Cracking a Microsoft contest or why Silverlight-WCF security is important Visit and bookmark SilverlightShow. Stay in the 'Light

    Read the article

  • Do professional software developers still dream of creating industry/world-changing apps?

    - by Andrew Heath
    I'm a hobby programmer. The absence of real world deadlines, customer feedback, or performance reviews leaves me free to daydream about having and implementing The Next Great Idea That Changes the World. Of course I'm aware I probably have a better chance of winning the lottery, but it's fun to imagine knocking out some fully-homebrewed app that destroys the status quo. I know many professional programmers have side projects, some for profit others not. I was wondering on the way to work this morning (non-IT boring work) if having to code for your food tended to dampen the dreaming? Does greater experience leave you jaded and more focused on the projects at hand? Not trying to be a downer, just interested in the mindset of the real software professional :-)

    Read the article

  • Tips/Process for web-development using Django in a small team

    - by Mridang Agarwalla
    We're developing a web app uing Django and we're a small team of 3-4 programmers — some doing the UI stuff and some doing the Backend stuff. I'd love some tips and suggestions from the people here. This is out current setup: We're using Git as as our SCM tool and following this branching model. We're following the PEP8 for your style guide. Agile is our software development methodology and we're using Jira for that. We're using the Confluence plugin for Jira for documentation and I'm going to be writing a script that also dumps the PyDocs into Confluence. We're using virtualenv for sandboxing We're using zc.buildout for building This is whatever I can think of off the top of my head. Any other suggestions/tips would be welcome. I feel that we have a pretty good set up but I'm also confident that we could do more. Thanks.

    Read the article

  • Why did Apple remove Python support in Mavericks, aka Mac OS X 10.9?

    - by alex gray
    Apple has removed Python support (at least on the Developer level) in 10.9. Python IS still on the machine in /System/Library/Frameworks/Python.framework... but trying to link to Python using the 10.9 SDK fails. /Applications/Xcode.app/Contents/Developer/Platforms/MacOSX.platform/Developer/SDKs/MacOSX10.9.sdk/System/Library/Frameworks does not have Python. I'm not a Pythonista, but find it interesting that Apple has made this change. I don't understand why this is done and I'm a bit annoyed that I have to remove Python from my compilation units in order to compile with 10.9 SDK. Is this a statement by Apple, along the lines of "People aren't using Python very much anymore so we're going to phase out support"? Or was something else driving the change?

    Read the article

  • Developing iOS apps as web developer

    - by Keyo
    My Boss has sold a few 'iPhone apps' to clients, we are a web development shop. I have explained to him that I do not know the first thing about them, but it's such a powerful buzz-word and we need to meet clients expectations. I do have some experience in C, Java and Python which should help if I need to use objective-C. I have even done a few Android tutorials. These apps will more or less be HTML, in my mind they are not real apps, but faux apps which have the same functionality as the clients' websites. To me a real app is something that uses the phones hardware inputs and outputs, gps, accelerometer, speaker etc. What resources can I use to get up to speed iOS development and how to build apps in html. I have no idea where to begin.

    Read the article

  • Taking the training wheels off: Accelerating the Business with Oracle IAM by Brian Mozinski (Accenture)

    - by Greg Jensen
    Today, technical requirements for IAM are evolving rapidly, and the bar is continuously raised for high performance IAM solutions as organizations look to roll out high volume use cases on the back of legacy systems.  Existing solutions were often designed and architected to support offline transactions and manual processes, and the business owners today demand globally scalable infrastructure to support the growth their business cases are expected to deliver. To help IAM practitioners address these challenges and make their organizations and themselves more successful, this series we will outline the: • Taking the training wheels off: Accelerating the Business with Oracle IAM The explosive growth in expectations for IAM infrastructure, and the business cases they support to gain investment in new security programs. • "Necessity is the mother of invention": Technical solutions developed in the field Well proven tricks of the trade, used by IAM guru’s to maximize your solution while addressing the requirements of global organizations. • The Art & Science of Performance Tuning of Oracle IAM 11gR2 Real world examples of performance tuning with Oracle IAM • No Where to go but up: Extending the benefits of accelerated IAM Anything is possible, compelling new solutions organizations are unlocking with accelerated Oracle IAM Let’s get started … by talking about the changing dynamics driving these discussions. Big Companies are getting bigger everyday, and increasingly organizations operate across state lines, multiple times zones, and in many countries or continents at the same time.  No longer is midnight to 6am a safe time to take down the system for upgrades, to run recon’s and import or update user accounts and attributes.  Further IT organizations are operating as shared services with SLA’s similar to telephone carrier levels expected by their “clients”.  Workers are moved in and out of roles on a weekly, daily, or even hourly rate and IAM is expected to support those rapid changes.  End users registering for services during business hours in Singapore are expected their access to be green-lighted in custom apps hosted in Portugal within the hour.  Many of the expectations of asynchronous systems and batched updates are not adequate and the number and types of users is growing. When organizations acted more like independent teams at functional or geographic levels it was manageable to have processes that relied on a handful of people who knew how to make things work …. Knew how to get you access to the key systems to get your job done.  Today everyone is expected to do more with less, the finance administrator previously supporting their local Atlanta sales office might now be asked to help close the books for the Johannesburg team, and access certification process once completed monthly by Joan on the 3rd floor is now done by a shared pool of resources in Sao Paulo.   Fragmented processes that rely on institutional knowledge to get access to systems and get work done quickly break down in these scenarios.  Highly robust processes that have automated workflows for connected or disconnected systems give organizations the dynamic flexibility to share work across these lines and cut costs or increase productivity. As the IT industry computing paradigms continue to change with the passing of time, and as mature or proven approaches become clear, it is normal for organizations to adjust accordingly. Businesses must manage identity in an increasingly hybrid world in which legacy on-premises IAM infrastructures are extended or replaced to support more and more interconnected and interdependent services to a wider range of users. The old legacy IAM implementation models we had relied on to manage identities no longer apply. End users expect to self-request access to services from their tablet, get supervisor approval over mobile devices and email, and launch the application even if is hosted on the cloud, or run by a partner, vendor, or service provider. While user expectations are higher, they are also simpler … logging into custom desktop apps to request approvals, or going through email or paper based processes for certification is unacceptable.  Users expect security to operate within the paradigm of the application … i.e. feel like the application they are using. Citizen and customer facing applications have evolved from every where, with custom applications, 3rd party tools, and merging in from acquired entities or 3rd party OEM’s resold to expand your portfolio of services.  These all have their own user stores, authentication models, user lifecycles, session management, etc.  Often the designers/developers are no longer accessible and the documentation is limited.  Bringing together underlying directories to scale for growth, and improve user experience is critical for revenue … but also for operations. Job functions are more dynamic.... take the Olympics for example.  Endless organizations from corporations broadcasting, endorsing, or marketing through the event … to non-profit athletic foundations and public/government entities for athletes and public safety, all operate simultaneously on the world stage.  Each organization needs to spin up short-term teams, often dealing with proprietary information from hot ads to racing strategies or security plans.  IAM is expected to enable team’s to spin up, enable new applications, protect privacy, and secure critical infrastructure.  Then it needs to be disabled just as quickly as users go back to their previous responsibilities. On a more technical level … Optimized system directory; tuning guidelines and parameters are needed by businesses today. Business’s need to be making the right choices (virtual directories) and considerations via choosing the correct architectural patterns (virtual, direct, replicated, and tuning), challenge is that business need to assess and chose the correct architectural patters (centralized, virtualized, and distributed) Today's Business organizations have very complex heterogeneous enterprises that contain diverse and multifaceted information. With today's ever changing global landscape, the strategic end goal in challenging times for business is business agility. The business of identity management requires enterprise's to be more agile and more responsive than ever before. The continued proliferation of networking devices (PC, tablet, PDA's, notebooks, etc.) has caused the number of devices and users to be granted access to these devices to grow exponentially. Business needs to deploy an IAM system that can account for the demands for authentication and authorizations to these devices. Increased innovation is forcing business and organizations to centralize their identity management services. Access management needs to handle traditional web based access as well as handle new innovations around mobile, as well as address insufficient governance processes which can lead to rouge identity accounts, which can then become a source of vulnerabilities within a business’s identity platform. Risk based decisions are providing challenges to business, for an adaptive risk model to make proper access decisions via standard Web single sign on for internal and external customers,. Organizations have to move beyond simple login and passwords to address trusted relationship questions such as: Is this a trusted customer, client, or citizen? Is this a trusted employee, vendor, or partner? Is this a trusted device? Without a solid technological foundation, organizational performance, collaboration, constituent services, or any other organizational processes will languish. A Single server location presents not only network concerns for distributed user base, but identity challenges. The network risks are centered on latency of the long trip that the traffic has to take. Other risks are a performance around availability and if the single identity server is lost, all access is lost. As you can see, there are many reasons why performance tuning IAM will have a substantial impact on the success of your organization.  In our next installment in the series we roll up our sleeves and get into detailed tuning techniques used everyday by thought leaders in the field implementing Oracle Identity & Access Management Solutions.

    Read the article

  • Thread placement policies on NUMA systems - update

    - by Dave
    In a prior blog entry I noted that Solaris used a "maximum dispersal" placement policy to assign nascent threads to their initial processors. The general idea is that threads should be placed as far away from each other as possible in the resource topology in order to reduce resource contention between concurrently running threads. This policy assumes that resource contention -- pipelines, memory channel contention, destructive interference in the shared caches, etc -- will likely outweigh (a) any potential communication benefits we might achieve by packing our threads more densely onto a subset of the NUMA nodes, and (b) benefits of NUMA affinity between memory allocated by one thread and accessed by other threads. We want our threads spread widely over the system and not packed together. Conceptually, when placing a new thread, the kernel picks the least loaded node NUMA node (the node with lowest aggregate load average), and then the least loaded core on that node, etc. Furthermore, the kernel places threads onto resources -- sockets, cores, pipelines, etc -- without regard to the thread's process membership. That is, initial placement is process-agnostic. Keep reading, though. This description is incorrect. On Solaris 10 on a SPARC T5440 with 4 x T2+ NUMA nodes, if the system is otherwise unloaded and we launch a process that creates 20 compute-bound concurrent threads, then typically we'll see a perfect balance with 5 threads on each node. We see similar behavior on an 8-node x86 x4800 system, where each node has 8 cores and each core is 2-way hyperthreaded. So far so good; this behavior seems in agreement with the policy I described in the 1st paragraph. I recently tried the same experiment on a 4-node T4-4 running Solaris 11. Both the T5440 and T4-4 are 4-node systems that expose 256 logical thread contexts. To my surprise, all 20 threads were placed onto just one NUMA node while the other 3 nodes remained completely idle. I checked the usual suspects such as processor sets inadvertently left around by colleagues, processors left offline, and power management policies, but the system was configured normally. I then launched multiple concurrent instances of the process, and, interestingly, all the threads from the 1st process landed on one node, all the threads from the 2nd process landed on another node, and so on. This happened even if I interleaved thread creating between the processes, so I was relatively sure the effect didn't related to thread creation time, but rather that placement was a function of process membership. I this point I consulted the Solaris sources and talked with folks in the Solaris group. The new Solaris 11 behavior is intentional. The kernel is no longer using a simple maximum dispersal policy, and thread placement is process membership-aware. Now, even if other nodes are completely unloaded, the kernel will still try to pack new threads onto the home lgroup (socket) of the primordial thread until the load average of that node reaches 50%, after which it will pick the next least loaded node as the process's new favorite node for placement. On the T4-4 we have 64 logical thread contexts (strands) per socket (lgroup), so if we launch 48 concurrent threads we will find 32 placed on one node and 16 on some other node. If we launch 64 threads we'll find 32 and 32. That means we can end up with our threads clustered on a small subset of the nodes in a way that's quite different that what we've seen on Solaris 10. So we have a policy that allows process-aware packing but reverts to spreading threads onto other nodes if a node becomes too saturated. It turns out this policy was enabled in Solaris 10, but certain bugs suppressed the mixed packing/spreading behavior. There are configuration variables in /etc/system that allow us to dial the affinity between nascent threads and their primordial thread up and down: see lgrp_expand_proc_thresh, specifically. In the OpenSolaris source code the key routine is mpo_update_tunables(). This method reads the /etc/system variables and sets up some global variables that will subsequently be used by the dispatcher, which calls lgrp_choose() in lgrp.c to place nascent threads. Lgrp_expand_proc_thresh controls how loaded an lgroup must be before we'll consider homing a process's threads to another lgroup. Tune this value lower to have it spread your process's threads out more. To recap, the 'new' policy is as follows. Threads from the same process are packed onto a subset of the strands of a socket (50% for T-series). Once that socket reaches the 50% threshold the kernel then picks another preferred socket for that process. Threads from unrelated processes are spread across sockets. More precisely, different processes may have different preferred sockets (lgroups). Beware that I've simplified and elided details for the purposes of explication. The truth is in the code. Remarks: It's worth noting that initial thread placement is just that. If there's a gross imbalance between the load on different nodes then the kernel will migrate threads to achieve a better and more even distribution over the set of available nodes. Once a thread runs and gains some affinity for a node, however, it becomes "stickier" under the assumption that the thread has residual cache residency on that node, and that memory allocated by that thread resides on that node given the default "first-touch" page-level NUMA allocation policy. Exactly how the various policies interact and which have precedence under what circumstances could the topic of a future blog entry. The scheduler is work-conserving. The x4800 mentioned above is an interesting system. Each of the 8 sockets houses an Intel 7500-series processor. Each processor has 3 coherent QPI links and the system is arranged as a glueless 8-socket twisted ladder "mobius" topology. Nodes are either 1 or 2 hops distant over the QPI links. As an aside the mapping of logical CPUIDs to physical resources is rather interesting on Solaris/x4800. On SPARC/Solaris the CPUID layout is strictly geographic, with the highest order bits identifying the socket, the next lower bits identifying the core within that socket, following by the pipeline (if present) and finally the logical thread context ("strand") on the core. But on Solaris on the x4800 the CPUID layout is as follows. [6:6] identifies the hyperthread on a core; bits [5:3] identify the socket, or package in Intel terminology; bits [2:0] identify the core within a socket. Such low-level details should be of interest only if you're binding threads -- a bad idea, the kernel typically handles placement best -- or if you're writing NUMA-aware code that's aware of the ambient placement and makes decisions accordingly. Solaris introduced the so-called critical-threads mechanism, which is expressed by putting a thread into the FX scheduling class at priority 60. The critical-threads mechanism applies to placement on cores, not on sockets, however. That is, it's an intra-socket policy, not an inter-socket policy. Solaris 11 introduces the Power Aware Dispatcher (PAD) which packs threads instead of spreading them out in an attempt to be able to keep sockets or cores at lower power levels. Maximum dispersal may be good for performance but is anathema to power management. PAD is off by default, but power management polices constitute yet another confounding factor with respect to scheduling and dispatching. If your threads communicate heavily -- one thread reads cache lines last written by some other thread -- then the new dense packing policy may improve performance by reducing traffic on the coherent interconnect. On the other hand if your threads in your process communicate rarely, then it's possible the new packing policy might result on contention on shared computing resources. Unfortunately there's no simple litmus test that says whether packing or spreading is optimal in a given situation. The answer varies by system load, application, number of threads, and platform hardware characteristics. Currently we don't have the necessary tools and sensoria to decide at runtime, so we're reduced to an empirical approach where we run trials and try to decide on a placement policy. The situation is quite frustrating. Relatedly, it's often hard to determine just the right level of concurrency to optimize throughput. (Understanding constructive vs destructive interference in the shared caches would be a good start. We could augment the lines with a small tag field indicating which strand last installed or accessed a line. Given that, we could augment the CPU with performance counters for misses where a thread evicts a line it installed vs misses where a thread displaces a line installed by some other thread.)

    Read the article

< Previous Page | 654 655 656 657 658 659 660 661 662 663 664 665  | Next Page >