Search Results

Search found 14762 results on 591 pages for 'window placement'.

Page 1/591 | 1 2 3 4 5 6 7 8 9 10 11 12  | Next Page >

  • expand window to free space on screen in kde

    - by Pascal Rosin
    I am using Kubuntu and I want to expand the current window to the free space on the screen or to say it more precisely: I want to make the current window as big as possible without overlapping new windows (windows already overlapped should be ignored). Is there a keyboard shortcut or an extension to the KDE Window management, that realizes such a shortcut or a window button? I would also appreciate a hint, how to write a script that could do this window thing on keyboard shortcut invocation. I am a programmer but don't know what the best way is to control KDE Windows via script.

    Read the article

  • Start/Stop Window Service from ASP.NET page

    - by kaushalparik27
    Last week, I needed to complete one task on which I am going to blog about in this entry. The task is "Create a control panel like webpage to control (Start/Stop) Window Services which are part of my solution installed on computer where the main application is hosted". Here are the important points to accomplish:[1] You need to add System.ServiceProcess reference in your application. This namespace holds ServiceController Class to access the window service.[2] You need to check the status of the window services before you explicitly start or stop it.[3] By default, IIS application runs under ASP.NET account which doesn't have access rights permission to window service. So, Very Important part of the solution is: Impersonation. You need to impersonate the application/part of the code with the User Credentials which is having proper rights and permission to access the window service. If you try to access window service it will generate "access denied" error.The alternatives are: You can either impersonate whole application by adding Identity tag in web.cofig as:        <identity impersonate="true" userName="" password=""/>This tag will be under System.Web section. the "userName" and "password" will be the credentials of the user which is having rights to access the window service. But, this would not be a wise and good solution; because you may not impersonate whole website like this just to have access window service (which is going to be a small part of code).Second alternative is: Only impersonate part of code where you need to access the window service to start or stop it. I opted this one. But, to be fair; I am really unaware of the code part for impersonation. So, I just googled it and injected the code in my solution in a separate class file named as "Impersonate" with required static methods. In Impersonate class; impersonateValidUser() is the method to impersonate a part of code and undoImpersonation() is the method to undo the impersonation. Below is one example:  You need to provide domain name (which is "." if you are working on your home computer), username and password of appropriate user to impersonate.[4] Here, it is very important to note that: You need to have to store the Access Credentials (username and password) which you are going to user for impersonation; to some secured and encrypted format. I have used Machinekey Encryption to store the value encrypted value inside database.[5] So now; The real part is to start or stop a window service. You are almost done; because ServiceController class has simple Start() and Stop() methods to start or stop a window service. A ServiceController class has parametrized constructor that takes name of the service as parameter.Code to Start the window service: Code to Stop the window service: Isn't that too easy! ServiceController made it easy :) I have attached a working example with this post here to start/stop "SQLBrowser" service where you need to provide proper credentials who have permission to access to window service.  hope it would helps./.

    Read the article

  • Window borders missing - gtk-window-decorator segmentation fault

    - by Balakrshnan Ramakrishnan
    I have been using Ubuntu for about 1 year now and I got a problem just two days ago. Suddenly I started experiencing a problem with the window borders (title bar with close, maximize, and minimize buttons). The problem : The window borders disappear I run "gtk-window-decorator --replace" For like 20 seconds everything is back to normal But it again returns to the problem. I searched over the Internet and found that my problem is similar to what is specified in this bug report: https://bugs.launchpad.net/ubuntu/+source/compiz/+bug/814091 This bug report says that the "fix released". I updated everything using the Update Manager, but still the problem remains. Can anyone let me know whether the problem is fixed? If yes, can you please let me know how to do it? I have already tried normal replace/reset commands like unity --reset unity --replace compiz --replace The "window borders" plugin is on in CCSM (CompizConfig Settings Manager) and it points to "gtk-window-decorator". I use Ubuntu 11.10 on an Intel Core2Duo T6500 with AMD Mobility Radeon HD 4300 graphics card. If you need more information, please let me know.

    Read the article

  • User upstart session in different window manager wIndow manager

    - by Joelmob
    I am using Ubuntu 14.04 and i3 as window manager. After I have logged in to i3, upstart won't find my user jobs under ~/.config/upstart/. How can I make upstart find these config files without having to execute something like gnome-session? Thanks. edit, one of the jobs starts redshift, here is the config ~/.config/upstar/redshift.conf: respawn exec redshift -l 59:18 When i try to start this job with initctl start redshift: initctl: Unknown job: redshift

    Read the article

  • window decoration command change

    - by Pedro
    I'm quite new to Ubuntu, and I don't know how to fix this: I have been experiencing the problem with window decorations that go away. I have been told that I could correct that with compiz fusion icon, just restarting the graphics whenever it happened. This solution only fixed the problem for like 30 seconds. Then I found out that the command inside the window decorator had changed, so I only had to reset it to default (/usr/bin/compiz-decorator). The only problem is that it keeps changing the command by itself once or twice a day. I can live with that, but I would like to find a more permanent solution, if there is one. I am running Ubuntu 11.10

    Read the article

  • IE won't load PDF in a window created with window.open

    - by Dean
    Here's the problem, which only occurs in Internet Explorer (IE). I have a page that has links to several different types of files. Links from these files execute a Javascript function that opens a new window and loads the specific file. This works great, unless the file I need to open in the new window is a PDF in which case the window is blank, even though the URL is in the address field. Refreshing that window using F5 doesn't help. However, if I put the cursor in the address field and press <enter> the PDF loads right up. This problem only occurs in IE. I have seen it in IE 7 and 8 and am using Adobe Acrobat Reader 9. In Firefox (PC and Mac) everything works perfectly. In Chrome (Mac), the PDF is downloaded. In Safari (Mac) it works. In Opera (Mac) it prompts me to open or save. Basically, everything probably works fine, except for IE. I have searched for similar problems and have seen some posts where it was suggested to adjust some of the Internet Options on IE. I have tried this but it doesn't help, and the problem wasn't exactly the same anyway. Here's the Javascript function I use to open the new window. function newwin(url,w,h) { win = window.open(url,"temp","width="+w+",height="+h+",menubar=yes,toolbar=yes,location=yes,status=yes,scrollbars=auto,resizable=yes"); win.focus(); } You can see that I pass in the URL as well as the height, h, and width, w, of the window. I've used a function like this for years and as far as I know have never had a problem. I call the newwin() function using this. <a href="javascript:newwin('/path/document.pdf',400,300)">document.pdf</a> (Yes, I know there are other, better ways than using inline JS, and I've even tried some of them because I've run out of things to try, but nothing works.) So, if anyone has an idea as to what might be causing this problem, I'd love to hear it.

    Read the article

  • javascript window.open lost session on closing popup window

    - by user341799
    Hi I'm using window.open function to popup the window as I close it, the link in parent window are not working here is my code `var rcount =0; var radio_val = document.form1.ac_voucher_type_id.length; for (counter = 0; counter < radio_val; counter++){ if (document.form1.ac_voucher_type_id[counter].checked){ radio_choice = document.form1.ac_voucher_type_id[counter].value; //alert(document.form1.ac_voucher_type_id[counter].value); url1 = '?compid=&date='+document.getElementById('datepicker2').value+'&vouchertype='+radio_choice+'&voucherid='; url= typ+"entry.php"+url1; window.open(url,'',"width=500,height=500,scrollbars=yes"); }else{ rcount++; } } // alert(rcount); if(rcount==radio_val){ alert("Please select voucher type"); }` Please suggest the solution for this

    Read the article

  • When clicking below window title bar in KDE, window dragging is initiated. Can I change this behavior?

    - by hdpq
    this one is regarding KDE in Ubuntu 12.04. The thing is that whenever I click below the window title bar in KDE, the window manager always assumes I want to drag the window, not just activate it. Clicking the title bar activates the window without dragging it just fine, however I sometimes find it hard to click because it's rather narrow :) Is there any way to change this behavior (besides the obvious "make your title bar larger" thing)? I also assume this is happening to Gtk applications only (Firefox, Thunderbird, Synaptic), as I have tried it with several KDE apps (KWrite, KTorrent, Gwenview, Konsole) and noticed no such behavior. Thanks in advance!

    Read the article

  • How to prevent launcher from influencing window placement?

    - by Jeromy Anglim
    As I understand it, in Ubuntu 11.04 setting the AutoHide setting on Compiz Config - Unity (see here) not only hid the launcher but prevented the launcher from influencing window placement. However, since updating to Ubuntu 11.10 when I open a new window it gets indented the space of the launcher. I just want new windows to open flush left. This is annoying, because I like to have two windows on my large screen, one taking up the left half of the screen and the other taking up the right half of the screen. Indenting causes them to overlap. How can I stop Ubuntu 11.10 thinking that the launcher takes up desktop space when placing windows? The following screen shot shows what I'm talking about. I've just opened chromium and the window is indented by the width of the launcher.

    Read the article

  • Find window previously opened by window.open

    - by spender
    We've got the following situation, running from a single domain: Page A uses window.open() to open a named window (a popup player). window.open() gives page A a reference to the window. User now reloads page A. The reference to the named window is lost. Using window.open() to "find" the window has the unfortunate side effect of reloading it (undesirable). Is there any other way to get a reference to this window?

    Read the article

  • In a drag and drop Is here a way to make a destination folder/app window stay on top when the source folder window is underneath? (Windows/Linux)

    - by Rob
    Scenario: You have a file Window, where you want to drag a file from, and on top of this Window is another window, the destination: an app or another window. So you click on the file you want on the window underneath, and what happens? This is brought to the front and so the destination window is out of view, what a pain. How can you make the destination window stay on top? Windows and Linux please.

    Read the article

  • Can't specify window's class in CCSM's Place Windows

    - by Anto
    Following this method, I wanted to "place" some of my windows on specific viewports. For instance, I want Xchat to always start on my top right viewport (x=2, y=1). But after I have set everything up and clicked the "Close" button to validate it (I don't have any other action button, like "OK" or "Apply"), the only thing that get saved up are the viewport X and Y coordinates. The window class (in my case class=Xchat) remains empty. This bug only occurs in the "Windows with fixed viewport" section; it works well in the "Windows with fixed placement mode", for instance. How can I solve or circumvent this issue ?

    Read the article

  • Compiz not drawing window controls in ubuntu 11.10

    - by Siva Prasad Varma
    I have recently installed driver for my ATI graphic card in my Dell Studio laptop. I have also read this somewhere on the web that Ubuntu enables compiz window manager by default if your hardware can run it. Is it true ? In my case before Installing graphic card driver the window manager was Metacity, but now I have compiz as my Window manager. I found this out uisng Displex Indicator applet also confirmed by wmctrl -m. From the time I have installed graphic card drivers, the window manager(Compiz) is not drawing window control buttons for some of the windows. For example if I open a terminal I have to close it using key board shortcuts or use the File - Quit option in app-menu. Also I am not able to move the window because of this. From then when-ever I find a window without window control buttons I am restarting the window manager using Displex Indicator applet. But this is very annoying and also consumes a lot of time(when I am doing my work). Can any one suggest any solution for this. What are up's and down's of using Compiz Vs Metacity.

    Read the article

  • window.event !== window.event in IE

    - by iacnats
    Code: <html> <head> <script type="text/javascript"> onload = function(){ document.getElementById('btn1').onclick = function(){ if (window === window) alert('window === window') else alert('window !== window'); if (window.event === window.event) alert('window.event === window.event') else alert('window.event !== window.event' ); } } </script> </head> <body> <button id="btn1" >click</button> </body> </html> Result: IE(i have tested IE6 - IE8) says: window === window window.event !== window.event All other browsers say: window === window window.event === window.event What's the reason for IE's response? Thanks.

    Read the article

  • Can't center dialog window above main window when using Glade (Quickly)

    - by Niklas
    I have a dialog opening when I press a button in my application and I want it to center above the application's main window, but I can't find a way to do this. I'm using Quickly (Python) and Glade. In Glade I've tried the following settings: Window Type = Popup Modal = Yes Window Position = Center on Parent But I simply can't get the behaviour I'm after. When using the "Window Type" option "Top level" the dialog opens in the center of the screen. When I use "Window Type" = "Popup" it opens in the very top left corner, with the window controls unaccessible. I know this must be a very basic setting but I've just started developing with GTK and Python so I obviously managed to miss this. Thanks for any advice! :)

    Read the article

  • How to drag from a background window to the front window

    - by Luis Alvarado
    Is the following I will explain possible with a key combination? Here is the image: As you can see, the terminal is the focus window (Front window) and Nautilus is in the background (Back Window). How can I grab a folder or file from Nautilus without loosing focus on the terminal (Without making the terminal go to the background and Nautilus to the front) and drop it in the terminal?. What I want is not to have to ALT+TAB again just to do this. Options like resizing the windows to fit the screen are not what I am looking for. Like in the image, we have a fullscreen window that we want it to stay like that. We can drag the terminal window around but anytime I access the background nautilus window, I should not loose focus on the terminal (It should not go to the background every time I access Nautilus). Maybe like a key combination that freezes the current focus windows positions and I can drag from background windows to background windows or background windows to the front focused one.

    Read the article

  • Window tiling in Gnome without Compiz?

    - by Melmacian
    I really do like the Grid plugin in Compiz, but at work I cannot use Compiz on my workstation (Fedora 12). Is there anything similar for plain X or Metacity? The Grid-plugin gives you a few keyboard shortcuts for tiling active window. For example Ctrl+Alt+Numpad_4 would tile window to half screen.

    Read the article

  • NUMA-aware placement of communication variables

    - by Dave
    For classic NUMA-aware programming I'm typically most concerned about simple cold, capacity and compulsory misses and whether we can satisfy the miss by locally connected memory or whether we have to pull the line from its home node over the coherent interconnect -- we'd like to minimize channel contention and conserve interconnect bandwidth. That is, for this style of programming we're quite aware of where memory is homed relative to the threads that will be accessing it. Ideally, a page is collocated on the node with the thread that's expected to most frequently access the page, as simple misses on the page can be satisfied without resorting to transferring the line over the interconnect. The default "first touch" NUMA page placement policy tends to work reasonable well in this regard. When a virtual page is first accessed, the operating system will attempt to provision and map that virtual page to a physical page allocated from the node where the accessing thread is running. It's worth noting that the node-level memory interleaving granularity is usually a multiple of the page size, so we can say that a given page P resides on some node N. That is, the memory underlying a page resides on just one node. But when thinking about accesses to heavily-written communication variables we normally consider what caches the lines underlying such variables might be resident in, and in what states. We want to minimize coherence misses and cache probe activity and interconnect traffic in general. I don't usually give much thought to the location of the home NUMA node underlying such highly shared variables. On a SPARC T5440, for instance, which consists of 4 T2+ processors connected by a central coherence hub, the home node and placement of heavily accessed communication variables has very little impact on performance. The variables are frequently accessed so likely in M-state in some cache, and the location of the home node is of little consequence because a requester can use cache-to-cache transfers to get the line. Or at least that's what I thought. Recently, though, I was exploring a simple shared memory point-to-point communication model where a client writes a request into a request mailbox and then busy-waits on a response variable. It's a simple example of delegation based on message passing. The server polls the request mailbox, and having fetched a new request value, performs some operation and then writes a reply value into the response variable. As noted above, on a T5440 performance is insensitive to the placement of the communication variables -- the request and response mailbox words. But on a Sun/Oracle X4800 I noticed that was not the case and that NUMA placement of the communication variables was actually quite important. For background an X4800 system consists of 8 Intel X7560 Xeons . Each package (socket) has 8 cores with 2 contexts per core, so the system is 8x8x2. Each package is also a NUMA node and has locally attached memory. Every package has 3 point-to-point QPI links for cache coherence, and the system is configured with a twisted ladder "mobius" topology. The cache coherence fabric is glueless -- there's not central arbiter or coherence hub. The maximum distance between any two nodes is just 2 hops over the QPI links. For any given node, 3 other nodes are 1 hop distant and the remaining 4 nodes are 2 hops distant. Using a single request (client) thread and a single response (server) thread, a benchmark harness explored all permutations of NUMA placement for the two threads and the two communication variables, measuring the average round-trip-time and throughput rate between the client and server. In this benchmark the server simply acts as a simple transponder, writing the request value plus 1 back into the reply field, so there's no particular computation phase and we're only measuring communication overheads. In addition to varying the placement of communication variables over pairs of nodes, we also explored variations where both variables were placed on one page (and thus on one node) -- either on the same cache line or different cache lines -- while varying the node where the variables reside along with the placement of the threads. The key observation was that if the client and server threads were on different nodes, then the best placement of variables was to have the request variable (written by the client and read by the server) reside on the same node as the client thread, and to place the response variable (written by the server and read by the client) on the same node as the server. That is, if you have a variable that's to be written by one thread and read by another, it should be homed with the writer thread. For our simple client-server model that means using split request and response communication variables with unidirectional message flow on a given page. This can yield up to twice the throughput of less favorable placement strategies. Our X4800 uses the QPI 1.0 protocol with source-based snooping. Briefly, when node A needs to probe a cache line it fires off snoop requests to all the nodes in the system. Those recipients then forward their response not to the original requester, but to the home node H of the cache line. H waits for and collects the responses, adjudicates and resolves conflicts and ensures memory-model ordering, and then sends a definitive reply back to the original requester A. If some node B needed to transfer the line to A, it will do so by cache-to-cache transfer and let H know about the disposition of the cache line. A needs to wait for the authoritative response from H. So if a thread on node A wants to write a value to be read by a thread on node B, the latency is dependent on the distances between A, B, and H. We observe the best performance when the written-to variable is co-homed with the writer A. That is, we want H and A to be the same node, as the writer doesn't need the home to respond over the QPI link, as the writer and the home reside on the very same node. With architecturally informed placement of communication variables we eliminate at least one QPI hop from the critical path. Newer Intel processors use the QPI 1.1 coherence protocol with home-based snooping. As noted above, under source-snooping a requester broadcasts snoop requests to all nodes. Those nodes send their response to the home node of the location, which provides memory ordering, reconciles conflicts, etc., and then posts a definitive reply to the requester. In home-based snooping the snoop probe goes directly to the home node and are not broadcast. The home node can consult snoop filters -- if present -- and send out requests to retrieve the line if necessary. The 3rd party owner of the line, if any, can respond either to the home or the original requester (or even to both) according to the protocol policies. There are myriad variations that have been implemented, and unfortunately vendor terminology doesn't always agree between vendors or with the academic taxonomy papers. The key is that home-snooping enables the use of a snoop filter to reduce interconnect traffic. And while home-snooping might have a longer critical path (latency) than source-based snooping, it also may require fewer messages and less overall bandwidth. It'll be interesting to reprise these experiments on a platform with home-based snooping. While collecting data I also noticed that there are placement concerns even in the seemingly trivial case when both threads and both variables reside on a single node. Internally, the cores on each X7560 package are connected by an internal ring. (Actually there are multiple contra-rotating rings). And the last-level on-chip cache (LLC) is partitioned in banks or slices, which with each slice being associated with a core on the ring topology. A hardware hash function associates each physical address with a specific home bank. Thus we face distance and topology concerns even for intra-package communications, although the latencies are not nearly the magnitude we see inter-package. I've not seen such communication distance artifacts on the T2+, where the cache banks are connected to the cores via a high-speed crossbar instead of a ring -- communication latencies seem more regular.

    Read the article

  • Get window handle from window class name

    - by Einar Egilsson
    I'm trying to get a window handle on a child window in my process and the only information I have is the window class name. Are there any win32 functions I can use for that? I'm doing this from C#. A bit more detail: This is a Visual Studio plugin, written in C#. So my process is visual studio, which has lots of windows. One of them has a window class "VsTipWindow". I don't know the immediate parent window of that window, all I have is the class name. Is there any way for me to get the window handle from just that?

    Read the article

  • Open a popup window on top of a popup window in ASP.NET

    - by demogeek
    I have a scenario where I have a popup window open and that will have links to open up another window that's going to pop open on top of the already open popup window. I've tried all sort of tricks (javascript window.open, target="_blank" etc.) but nothing seem to work. It always was opening the page on the already opened popup window. Any suggestions? Appreciated your help.

    Read the article

  • Traversing/Manipulating new window from parent window with jquery

    - by Anthony
    I found a question already with a possible solution, but I had already tried the method suggested before reading it, and it even says it's inconsistent. So I guess I'm wondering why this doesn't work: $("img").click(function() { var imgwindow = window.open('imgwin.html','','width=460,height=345'); alert(imgwindow.find("img").attr("src")); }); The goal is to set the new window's img to the same src as the image that the user clicked to open the new window. But in the above scenario, I'm just trying to tap into the new window to get what it's already-set src. I have imgwin.html already written with a default src for the image, so it should alert that url. Instead I just get undefined. I also tried the following (which should be identical on the back end), but with the same results: $("img").click(function() { var imgwindow = window.open('imgwin.html','','width=460,height=345'); alert($("img",imgwindow).attr("src")); }); I even tried variations of wrapping the imgwindow variable in $() in case somehow jquery wasn't picking up on the variable as a DOM element. The only thing I can guess is that since window.open() is a method, jquery doesn't treat it like a DOM element, even though all documentation from basic javascript tutorials treat the variable as both a method to open the window and the pointer to the window's inner DOM.

    Read the article

  • javascript to launch only ONE window for a Java applet with a given URL

    - by Jonathan Dugan
    I need a javascript solution to launch only one window, with a Java Applet in it, for a given URL. I found a solution posted here on Stack Overflow - here: http://stackoverflow.com/questions/528671/javascript-window-open-only-if-the-window-does-not-already-exist But it doesn't seem to work .. I get Error: launchApplication.winrefs is undefined Line: 29 I can't seem to post the code in this little box and make it look right below, so the code (my working code, plus the solution from above) is here: http://pastie.org/833879 Where is the error? As I understand it, the hash or array or whatever I use to store the called references to the windows opened this way will be lost if the calling window is closed. Is there a way to make this work even if the calling window is closed and reopened? To basically ask the browser: "Do you have a window open with the following URL?" and if so, "What is the reference to that window?" (so I can raise it).

    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

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