Search Results

Search found 16784 results on 672 pages for 'white box'.

Page 78/672 | < Previous Page | 74 75 76 77 78 79 80 81 82 83 84 85  | Next Page >

  • Day 4 - Game Sprites In Action

    - by dapostolov
    Yesterday I drew an image on the screen. Most exciting, but ... I spent more time blogging about it then actual coding. So this next little while I'm going to streamline my game and research and simply post key notes. Quick notes on the last session: The most important thing I wanted to point out were the following methods:           spriteBatch.Begin(SpriteBlendMode.AlphaBlend);           spriteBatch.Draw(sprite, position, Color.White);           spriteBatch.End(); The spriteBatch object is used to draw Textures and a 2D texture is called a Sprite A texture is generally an image, which is called an Asset in XNA The Draw Method in the Game1.cs is looped (until exit) and utilises the spriteBatch object to draw a Scene To begin drawing a Scene you call the Begin Method. To end a Scene you call the End Method. And to place an image on the Scene you call the Draw method. The most simple implementation of the draw method is:           spriteBatch.Draw(sprite, position, Color.White); 1) sprite - the 2D texture you loaded to draw 2) position - the 2d vector, a set of x & y coordinates 3) Color.White - the tint to apply to the texture, in this case, white light = nothing, nada, no tint. Game Sprites In Action! Today, I played around with Draw methods to get comfortable with their "quirks". The following is an example of the above draw method, but with more parameters available for us to use. Let's investigate!             spriteBatch.Draw(sprite, position2, null, Color.White, MathHelper.ToRadians(45.0f), new Vector2(sprite.Width / 2, sprite.Height / 2), 1.0F, SpriteEffects.None, 0.0F); The parameters (in order): 1) sprite  the texture to display 2) position2 the position on the screen / scene this can also be a rectangle 3) null the portion of the image to display within an image null = display full image this is generally used for animation strips / grids (more on this below) 4) Color.White Texture tinting White = no tint 5) MathHelper.ToRadians(45.0f) rotation of the object, in this case 45 degrees rotates from the set plotting point. 6) new Vector(0,0) the plotting point in this case the top left corner the image will rotate from the top left of the texture in the code above, the point is set to the middle of the image. 7) 1.0f Image scaling (1x) 8) SpriteEffects.None you can flip the image horizontally or vertically 9) 0.0f The z index of the image. 0 = closer, 1 behind? And playing around with different combinations I was able to come up with the following whacky display:   Checking off Yesterdays Intention List: learn game development terminology (in progress) - We learned sprite, scene, texture, and asset. how to place and position (rotate) a static image on the screen (completed) - The thing to note was, it's was in radians and I found a cool helper method to convert degrees into radians. Also, the image rotates from it's specified point. how to layer static images on the screen (completed) - I couldn't seem to get the zIndex working, but one things for sure, the order you draw the image in also determines how it is rendered on the screen. understand image scaling (in progress) - I'm not sure I have this fully covered, but for the most part plug a number in the scaling field and the image grows / shrinks accordingly. can we reuse images? (completed) - yes, I loaded one image and plotted the bugger all over the screen. understand how framerate is handled in XNA (in progress) - I hacked together some code to display the framerate each second. A framerate of 60 appears to be the standard. Interesting to note, the GameTime object does provide you with some cool timing capabilities, such as...is the game running slow? Need to investigate this down the road. how to display text , basic shapes, and colors on the screen (in progress) - i got text rendered on the screen, and i understand containing rectangles. However, I didn't display "shapes" & "colors" how to interact with an image (collision of user input?) (todo) how to animate an image and understand basic animation techniques (in progress) - I was able to create a stripe animation of numbers ranging from 1 - 4, each block was 40 x 40 pixles for a total stripe size of 160 x 40. Using the portion (source Rectangle) parameter, i limited this display to each section at varying intervals. It was interesting to note my first implementation animated at rocket speed. I then tried to create a smoother animation by limiting the redraw capacity, which seemed to work. I guess a little more research will have to be put into this for animating characters / scenes. how to detect colliding images or screen edges (todo) - but the rectangle object can detect collisions I believe. how to manipulate the image, lets say colors, stretching (in progress) - I haven't figured out how to modify a specific color to be another color, but the tinting parameter definately could be used. As for stretching, use the rectangle object as the positioning and the image will stretch to fit! how to focus on a segment of an image...like only displaying a frame on a film reel (completed) - as per basic animation techniques what's the best way to manage images (compression, storage, location, prevent artwork theft, etc.) (todo) Tomorrows Intention Tomorrow I am going to take a stab at rendering a game menu and from there I'm going to investigate how I can improve upon the code and techniques. Intention List: Render a menu, fancy or not Show the mouse cursor Hook up click event A basic animation of somesort Investigate image / menu techniques D.

    Read the article

  • Unexpected behaviour with glFramebufferTexture1D

    - by Roshan
    I am using render to texture concept with glFramebufferTexture1D. I am drawing a cube on non-default FBO with all the vertices as -1,1 (maximum) in X Y Z direction. Now i am setting viewport to X while rendering on non default FBO. My background is blue with white color of cube. For default FBO, i have created 1-D texture and attached this texture to above FBO with color attachment. I am setting width of texture equal to width*height of above FBO view-port. Now, when i render this texture to on another cube, i can see continuous white color on start or end of each face of the cube. That means part of the face is white and rest is blue. I am not sure whether this behavior is correct or not. I expect all the texels should be white as i am using -1 and 1 coordinates for cube rendered on non-default FBO. code: #define WIDTH 3 #define HEIGHT 3 GLfloat vertices8[]={ 1.0f,1.0f,1.0f, -1.0f,1.0f,1.0f, -1.0f,-1.0f,1.0f, 1.0f,-1.0f,1.0f,//face 1 1.0f,-1.0f,-1.0f, -1.0f,-1.0f,-1.0f, -1.0f,1.0f,-1.0f, 1.0f,1.0f,-1.0f,//face 2 1.0f,1.0f,1.0f, 1.0f,-1.0f,1.0f, 1.0f,-1.0f,-1.0f, 1.0f,1.0f,-1.0f,//face 3 -1.0f,1.0f,1.0f, -1.0f,1.0f,-1.0f, -1.0f,-1.0f,-1.0f, -1.0f,-1.0f,1.0f,//face 4 1.0f,1.0f,1.0f, 1.0f,1.0f,-1.0f, -1.0f,1.0f,-1.0f, -1.0f,1.0f,1.0f,//face 5 -1.0f,-1.0f,1.0f, -1.0f,-1.0f,-1.0f, 1.0f,-1.0f,-1.0f, 1.0f,-1.0f,1.0f//face 6 }; GLfloat vertices[]= { 0.5f,0.5f,0.5f, -0.5f,0.5f,0.5f, -0.5f,-0.5f,0.5f, 0.5f,-0.5f,0.5f,//face 1 0.5f,-0.5f,-0.5f, -0.5f,-0.5f,-0.5f, -0.5f,0.5f,-0.5f, 0.5f,0.5f,-0.5f,//face 2 0.5f,0.5f,0.5f, 0.5f,-0.5f,0.5f, 0.5f,-0.5f,-0.5f, 0.5f,0.5f,-0.5f,//face 3 -0.5f,0.5f,0.5f, -0.5f,0.5f,-0.5f, -0.5f,-0.5f,-0.5f, -0.5f,-0.5f,0.5f,//face 4 0.5f,0.5f,0.5f, 0.5f,0.5f,-0.5f, -0.5f,0.5f,-0.5f, -0.5f,0.5f,0.5f,//face 5 -0.5f,-0.5f,0.5f, -0.5f,-0.5f,-0.5f, 0.5f,-0.5f,-0.5f, 0.5f,-0.5f,0.5f//face 6 }; GLuint indices[] = { 0, 2, 1, 0, 3, 2, 4, 5, 6, 4, 6, 7, 8, 9, 10, 8, 10, 11, 12, 15, 14, 12, 14, 13, 16, 17, 18, 16, 18, 19, 20, 23, 22, 20, 22, 21 }; GLfloat texcoord[] = { 0.0, 0.0, 1.0, 0.0, 1.0, 1.0, 0.0, 1.0, 0.0, 0.0, 1.0, 0.0, 1.0, 1.0, 0.0, 1.0, 0.0, 0.0, 1.0, 0.0, 1.0, 1.0, 0.0, 1.0, 0.0, 0.0, 1.0, 0.0, 1.0, 1.0, 0.0, 1.0, 0.0, 0.0, 1.0, 0.0, 1.0, 1.0, 0.0, 1.0, 0.0, 0.0, 1.0, 0.0, 1.0, 1.0, 0.0, 1.0 }; glGenTextures(1, &id1); glBindTexture(GL_TEXTURE_1D, id1); glGenFramebuffers(1, &Fboid); glTexParameterf(GL_TEXTURE_1D, GL_TEXTURE_MIN_FILTER, GL_NEAREST); glTexParameterf(GL_TEXTURE_1D, GL_TEXTURE_MAG_FILTER, GL_NEAREST); glTexParameterf(GL_TEXTURE_1D, GL_TEXTURE_WRAP_S, GL_CLAMP_TO_EDGE); glTexImage1D(GL_TEXTURE_1D, 0, GL_RGBA, WIDTH*HEIGHT , 0, GL_RGBA, GL_UNSIGNED_BYTE,0); glBindFramebuffer(GL_FRAMEBUFFER, Fboid); glFramebufferTexture1D(GL_DRAW_FRAMEBUFFER,GL_COLOR_ATTACHMENT0,GL_TEXTURE_1D,id1,0); draw_cube(); glBindFramebuffer(GL_FRAMEBUFFER, 0); draw(); } draw_cube() { glViewport(0, 0, WIDTH, HEIGHT); glClearColor(0.0f, 0.0f, 0.5f, 1.0f); glClear(GL_COLOR_BUFFER_BIT); glEnableVertexAttribArray(glGetAttribLocation(temp.psId,"position")); glVertexAttribPointer(glGetAttribLocation(temp.psId,"position"), 3, GL_FLOAT, GL_FALSE, 0,vertices8); glDrawArrays (GL_TRIANGLE_FAN, 0, 24); } draw() { glClearColor(1.0f, 0.0f, 0.0f, 1.0f); glClearDepth(1.0f); glClear(GL_COLOR_BUFFER_BIT | GL_DEPTH_BUFFER_BIT); glEnableVertexAttribArray(glGetAttribLocation(shader_data.psId,"tk_position")); glVertexAttribPointer(glGetAttribLocation(shader_data.psId,"tk_position"), 3, GL_FLOAT, GL_FALSE, 0,vertices); nResult = GL_ERROR_CHECK((GL_NO_ERROR, "glVertexAttribPointer(position, 3, GL_FLOAT, GL_FALSE, 0,vertices);")); glEnableVertexAttribArray(glGetAttribLocation(shader_data.psId,"inputtexcoord")); glVertexAttribPointer(glGetAttribLocation(shader_data.psId,"inputtexcoord"), 2, GL_FLOAT, GL_FALSE, 0,texcoord); glBindTexture(*target11, id1); glDrawElements ( GL_TRIANGLES, 36,GL_UNSIGNED_INT, indices ); when i change WIDTH=HEIGHT=2, and call a glreadpixels with height, width equal to 4 in draw_cube() i can see first 2 pixels with white color, next two with blue(glclearcolor), next two white and then blue and so on.. Now when i change width parameter in glTeximage1D to 16 then ideally i should see alternate patches of white and blue right? But its not the case here. why so?

    Read the article

  • How to change all selected chars to _ in Vim

    - by Kev
    I try to draw a class diagram using Vim. I fill the editor window with white-spaces. Type :match SpellBad /\s/ to highlight all the white-spaces. Ctrl+Q to select vertical white-spaces. Ctrl+I to insert Bar(|) and then Esc ........................... v+l +... + l to select horizontal white-spaces But I don't know how to change all selected horizontal white-spaces to underscore(_). I have to hit _ serval times. When comes to long horizontal line, it's bad. ___________ ___________ | | | | | BaseClass |/__________| Client | |___________|\ |___________| /_\ | |____________________________________ | | | _____|_____ _____|_____ _____|_____ | | | | | | | SubClass1 | | SubClass2 | | SubClass3 | |___________| |___________| |¦¦¦¦¦¦¦¦¦¦¦| I want a quick method to do this. Select it - Change it - Done! Maybe map F6 to do it. Thanks!

    Read the article

  • What is wrong with my gtkrc file?

    - by PP
    I have written following gtkrc file from some other theme gtkrc file. This theme is normal theme with buttons using pixmap theme engine. I have also given background image to GtkEntry. Problem is that, When i use this theme my buttons doesn't show text one them and my entry box does not show cursor. Plus in engine "pixmap" tag I need to specify image name with it's path as I have already mentioned pixmap_path on the top of rc file but why I still need to specify the path in file = "xxx" # gtkrc file. pixmap_path "./backgrounds:./icons:./buttons:./emotions" gtk-button-images = 1 #Icon Sizes and color definitions gtk-icon-sizes = "gtk-small-toolbar=16,16:gtk-large-toolbar=24,24:gtk-button=16,16" gtk-toolbar-icon-size = GTK_ICON_SIZE_SMALL_TOOLBAR gtk_color_scheme = "fg_color:#000000\nbg_color:#848484\nbase_color:#000000\ntext_color:#000000\nselected_bg_color:#f39638\nselected_fg_color:#000000\ntooltip_bg_color:#634110\ntooltip_fg_color:#ffffff" style "theme-default" { xthickness = 10 ythickness = 10 GtkEntry::honors-transparent-bg-hint = 0 GtkMenuItem::arrow-spacing = 20 GtkMenuItem::horizontal-padding = 50 GtkMenuItem::toggle-spacing = 30 GtkOptionMenu::indicator-size = {11, 5} GtkOptionMenu::indicator-spacing = {6, 5, 4, 4} GtkTreeView::horizontal_separator = 5 GtkTreeView::odd_row_color = "#efefef" GtkTreeView::even_row_color = "#e3e3e3" GtkWidget::link-color = "#0062dc" # blue GtkWidget::visited-link-color = "#8c00dc" #purple GtkButton::default_border = { 0, 0, 0, 0 } GtkButton::child-displacement-x = 0 GtkButton::child-displacement-y = 1 GtkWidget::focus-padding = 0 GtkRange::trough-border = 0 GtkRange::slider-width = 19 GtkRange::stepper-size = 19 GtkScrollbar::min_slider_length = 36 GtkScrollbar::has-secondary-backward-stepper = 1 GtkPaned::handle_size = 8 GtkMenuBar::internal-padding = 0 GtkTreeView::expander_size = 13 #15 GtkExpander::expander_size = 13 #17 GtkScale::slider-length = 35 GtkScale::slider-width = 17 GtkScale::trough-border = 0 GtkWidget::link-color = "#0062dc" GtkWidget::visited-link-color = "#8c00dc" #purple WnckTasklist::fade-overlay-rect = 0 WnckTasklist::fade-loop-time = 5.0 # 5 seconds WnckTasklist::fade-opacity = 0.5 # final opacity #makes menu only overlap border GtkMenu::horizontal-offset = -1 #removes extra padding at top and bottom of menus. Makes menuitem overlap border GtkMenu::vertical-padding = 0 #set to the same as roundness, used for better hotspot selection of tabs GtkNotebook::tab-curvature = 2 GtkNotebook::tab-overlap = 4 GtkMenuItem::arrow-spacing = 10 GtkOptionMenu ::indicator-size = {11, 5} GtkCheckButton ::indicator-size = 16 GtkCheckButton ::indicator-spacing = 1 GtkRadioButton ::indicator-size = 16 GtkTreeView::horizontal_separator = 2 GtkTreeView::odd_row_color = "#efefef" GtkTreeView::even_row_color = "#e3e3e3" NautilusIconContainer::normal_icon_color = "#ff0000" GtkEntry::inner-border = {0, 0, 0, 0} GtkScrolledWindow::scrollbar-spacing = 0 GtkScrolledWindow::scrollbars-within-bevel = 1 fg[NORMAL] = @fg_color fg[ACTIVE] = @fg_color fg[PRELIGHT] = @fg_color fg[SELECTED] = @selected_fg_color fg[INSENSITIVE] = shade (3.0,@fg_color) bg[NORMAL] = @bg_color bg[ACTIVE] = shade (0.95,@bg_color) bg[PRELIGHT] = mix(0.92, shade (1.1,@bg_color), @selected_bg_color) bg[SELECTED] = @selected_bg_color bg[INSENSITIVE] = shade (1.06,@bg_color) base[NORMAL] = @base_color base[ACTIVE] = shade (0.65,@base_color) base[PRELIGHT] = @base_color base[SELECTED] = @selected_bg_color base[INSENSITIVE] = shade (1.025,@bg_color) text[NORMAL] = @text_color text[ACTIVE] = shade (0.95,@base_color) text[PRELIGHT] = @text_color text[SELECTED] = @selected_fg_color text[INSENSITIVE] = mix (0.675,shade (0.95,@bg_color),@fg_color) } style "theme-entry" { xthickness = 10 ythickness = 10 GtkEntry::inner-border = {10, 10, 10, 10} GtkEntry::progress-border = {10, 10, 10, 10} GtkEntry::icon-prelight = 1 GtkEntry::state-hintt = 1 #GtkEntry::honors-transparent-bg-hint = 1 text[NORMAL] = "#000000" text[ACTIVE] = "#787878" text[INSENSITIVE] = "#787878" text[SELECTED] = "#FFFFFF" engine "pixmap" { image { function = FLAT_BOX state = NORMAL recolorable = FALSE file = "./backgrounds/entry_background.png" border = { 0, 0, 0, 0 } stretch = TRUE } image { function = FLAT_BOX state = PRELIGHT recolorable = FALSE file = "./backgrounds/entry_background.png" border = { 0, 0, 0, 0 } stretch = TRUE } image { function = FLAT_BOX state = ACTIVE recolorable = FALSE file = "./backgrounds/entry_background.png" border = { 0, 0, 0, 0 } stretch = TRUE } } } #----------------------------------------------- #Chat Balloon Incoming background. style "theme-event-box-top-in" { xthickness = 1 ythickness = 1 GtkEventBox::inner-border = {0, 0, 0, 0} engine "pixmap" { image { function = FLAT_BOX state = NORMAL recolorable = TRUE file = "./backgrounds/chat_in_top.png" border = { 0, 0, 0, 0 } stretch = TRUE } } } style "theme-event-box-mid-in" { xthickness = 1 ythickness = 1 GtkEventBox::inner-border = {0, 0, 0, 0} engine "pixmap" { image { function = FLAT_BOX state = NORMAL recolorable = TRUE file = "./backgrounds/chat_in_mid.png" border = { 0, 0, 0, 0 } stretch = TRUE } } } style "theme-event-box-bot-in" { xthickness = 1 ythickness = 1 GtkEventBox::inner-border = {0, 0, 0, 0} engine "pixmap" { image { function = FLAT_BOX state = NORMAL recolorable = TRUE file = "./backgrounds/chat_in_bot.png" border = { 0, 0, 0, 0 } stretch = TRUE } } } #----------------------------------------------- #Chat Balloon Outgoing background. style "theme-event-box-top-out" { xthickness = 1 ythickness = 1 GtkEventBox::inner-border = {0, 0, 0, 0} engine "pixmap" { image { function = FLAT_BOX state = NORMAL recolorable = TRUE file = "./backgrounds/chat_out_top.png" border = { 0, 0, 0, 0 } stretch = TRUE } } } style "theme-event-box-mid-out" { xthickness = 1 ythickness = 1 GtkEventBox::inner-border = {0, 0, 0, 0} engine "pixmap" { image { function = FLAT_BOX state = NORMAL recolorable = TRUE file = "./backgrounds/chat_out_mid.png" border = { 0, 0, 0, 0 } stretch = TRUE } } } style "theme-event-box-bot-out" { xthickness = 1 ythickness = 1 GtkEventBox::inner-border = {0, 0, 0, 0} engine "pixmap" { image { function = FLAT_BOX state = NORMAL recolorable = TRUE file = "./backgrounds/chat_out_bot.png" border = { 0, 0, 0, 0 } stretch = TRUE } } } style "theme-wide" = "theme-default" { xthickness = 2 ythickness = 2 } style "theme-wider" = "theme-default" { xthickness = 3 ythickness = 3 } style "theme-button" { GtkButton::inner-border = {0, 0, 0, 0} GtkWidget::focus-line-width = 0 GtkWidget::focus-padding = 0 bg[NORMAL] = "#414143" bg[ACTIVE] = "#c19676" bg[PRELIGHT] = "#7f4426" bg[SELECTED] = "#ff0000" bg[INSENSITIVE] = "#434346" fg[NORMAL] = "#ffffff" fg[INSENSITIVE] = "#000000" fg[PRELIGHT] = "#ffffff" fg[SELECTED] = "#ffffff" fg[ACTIVE] = "#ffffff" text[NORMAL] = "#ff0000" text[INSENSITIVE] = "#ff0000" text[PRELIGHT] = "#ff0000" text[SELECTED] = "#ff0000" text[INSENSITIVE] = "#434346" text[ACTIVE] = "#ff0000" base[NORMAL] = "#ff0000" base[INSENSITIVE] = "#ff0000" base[PRELIGHT] = "#ff0000" base[SELECTED] = "#ff0000" base[INSENSITIVE] = "#ff0000" engine "pixmap" { image { function = BOX state = NORMAL recolorable = TRUE file = "./buttons/LightButtonAct.png" border = { 0, 0, 0, 0 } stretch = TRUE } image { function = BOX state = PRELIGHT recolorable = TRUE file = "./buttons/LightButtonRoll.png" border = { 0, 0, 0, 0 } stretch = TRUE } image { function = BOX state = ACTIVE recolorable = TRUE file = "./buttons/LightButtonClicked.png" border = { 0, 0, 0, 0 } stretch = TRUE } image { function = BOX state = INSENSITIVE recolorable = TRUE file = "./buttons/LightButtonInact.png" border = { 0, 0, 0, 0 } stretch = TRUE } } } style "theme-toolbar" { xthickness = 2 ythickness = 2 bg[NORMAL] = shade (1.078,@bg_color) } style "theme-handlebox" { bg[NORMAL] = shade (0.95,@bg_color) } style "theme-scale" { bg[NORMAL] = shade (1.06, @bg_color) bg[PRELIGHT] = mix(0.85, shade (1.1,@bg_color), @selected_bg_color) bg[SELECTED] = "#4d4d55" } style "theme-range" { bg[NORMAL] = shade (1.12,@bg_color) bg[ACTIVE] = @bg_color bg[PRELIGHT] = mix(0.95, shade (1.10,@bg_color), @selected_bg_color) #Arrows text[NORMAL] = shade (0.275,@selected_fg_color) text[PRELIGHT] = @selected_fg_color text[ACTIVE] = shade (0.10,@selected_fg_color) text[INSENSITIVE] = mix (0.80,shade (0.90,@bg_color),@fg_color) } style "theme-notebook" = "theme-wider" { xthickness = 4 ythickness = 4 GtkNotebook::tab-curvature = 5 GtkNotebook::tab-vborder = 1 GtkNotebook::tab-overlap = 1 GtkNotebook::tab-vborder = 1 bg[NORMAL] = "#d2d2d2" bg[ACTIVE] = "#e3e3e3" bg[PRELIGHT] = "#848484" bg[SELECTED] = "#848484" bg[INSENSITIVE] = "#848484" text[PRELIGHT] = @selected_fg_color text[NORMAL] = "#000000" text[ACTIVE] = "#737373" text[SELECTED] = "#000000" text[INSENSITIVE] = "#737373" fg[PRELIGHT] = @selected_fg_color fg[NORMAL] = "#000000" fg[ACTIVE] = "#737373" fg[SELECTED] = "#000000" fg[INSENSITIVE] = "#737373" } style "theme-paned" { bg[PRELIGHT] = shade (1.1,@bg_color) } style "theme-panel" { # Menu fg[PRELIGHT] = @selected_fg_color font_name = "Bold 9" text[PRELIGHT] = @selected_fg_color } style "theme-menu" { xthickness = 0 ythickness = 0 bg[NORMAL] = shade (1.16,@bg_color) bg[SELECTED] = "#ff9a00" text[PRELIGHT] = @selected_fg_color fg[PRELIGHT] = @selected_fg_color } style "theme-menu-item" = "theme-menu" { xthickness = 3 ythickness = 3 base[SELECTED] = "#ff9a00" base[NORMAL] = "#ff9a00" base[PRELIGHT] = "#ff9a00" base[INSENSITIVE] = "#ff9a00" base[ACTIVE] = "#ff9a00" bg[SELECTED] = "#ff9a00" bg[NORMAL] = shade (1.16,@bg_color) } style "theme-menubar" { #TODO } style "theme-menubar-item" = "theme-menu-item" { #TODO bg[SELECTED] = "#ff9a00" } style "theme-tree" { xthickness = 2 ythickness = 1 font_name = "Bold 9" GtkWidget::focus-padding = 0 bg[NORMAL] = "#5a595a" bg[PRELIGHT] = "#5a595a" bg[ACTIVE] = "#5a5a5a" fg[NORMAL] = "#ffffff" fg[ACTIVE] = "#ffffff" fg[SELECTED] = "#ff9a00" fg[PRELIGHT] = "#ffffff" bg[SELECTED] = "#ff9a00" base[SELECTED] = "#ff9a00" base[NORMAL] = "#ff9a00" base[PRELIGHT] = "#ff9a00" base[INSENSITIVE] = "#ff9a00" base[ACTIVE] = "#ff9a00" text[NORMAL] = "#000000" text[PRELIGHT] = "#ff9a00" text[ACTIVE] = "#ff9a00" text[SELECTED] = "#ff9a00" text[INSENSITIVE] = "#434346" } style "theme-tree-arrow" { bg[NORMAL] = mix(0.70, shade (0.60,@bg_color), shade (0.80,@selected_bg_color)) bg[PRELIGHT] = mix(0.80, @bg_color, @selected_bg_color) } style "theme-progressbar" { font_name = "Bold" bg[SELECTED] = @selected_bg_color fg[PRELIGHT] = @selected_fg_color bg[ACTIVE] = "#fe7e00" bg[NORMAL] = "#ffba00" } style "theme-tooltips" = "theme-wider" { font_name = "Liberation sans 10" bg[NORMAL] = @tooltip_bg_color fg[NORMAL] = @tooltip_fg_color text[NORMAL] = @tooltip_fg_color } style "theme-combo" = "theme-button" { xthickness = 4 ythickness = 4 text[NORMAL] = "#fd7d00" text[INSENSITIVE] = "#8a8a8a" base[NORMAL] = "#e0e0e0" base[INSENSITIVE] = "#aeaeae" } style "theme-combo-box" = "theme-button" { xthickness = 3 ythickness = 2 bg[NORMAL] = "#343539" bg[PRELIGHT] = "#343539" bg[ACTIVE] = "#26272b" bg[INSENSITIVE] = "#404145" } style "theme-entry-combo-box" { xthickness = 6 ythickness = 3 text[NORMAL] = "#000000" text[INSENSITIVE] = "#8a8a8a" base[NORMAL] = "#ffffff" base[INSENSITIVE] = "#aeaeae" } style "theme-combo-arrow" = "theme-button" { xthickness = 1 ythickness = 1 } style "theme-view" { xthickness = 0 ythickness = 0 } style "theme-check-radio-buttons" { GtkWidget::interior-focus = 0 GtkWidget::focus-padding = 1 text[NORMAL] = "#ff0000" base[NORMAL] = "#ff0000" text[SELECTED] = "#ffffff" text[INSENSITIVE] = shade (0.625,@bg_color) base[PRELIGHT] = mix(0.80, @base_color, @selected_bg_color) bg[NORMAL] = "#438FC6" bg[INSENSITIVE] = "#aeaeae" bg[SELECTED] = "#ff8a01" } style "theme-radio-buttons" = "theme-button" { GtkWidget::interior-focus = 0 GtkWidget::focus-padding = 1 text[SELECTED] = @selected_fg_color text[INSENSITIVE] = shade (0.625,@bg_color) base[PRELIGHT] = mix(0.80, @base_color, @selected_bg_color) bg[NORMAL] = "#ffffff" bg[INSENSITIVE] = "#dcdcdc" bg[SELECTED] = @selected_bg_color } style "theme-spin-button" { bg[NORMAL] = "#d2d2d2" bg[ACTIVE] = "#868686" bg[PRELIGHT] = "#7f4426" bg[SELECTED] = shade(1.10,@selected_bg_color) bg[INSENSITIVE] = "#dcdcdc" base[NORMAL] = "#ffffff" base[INSENSITIVE] = "#dcdcdc" text[NORMAL] = "#000000" text[INSENSITIVE] = "#aeaeae" } style "theme-calendar" { xthickness = 0 ythickness = 0 bg[NORMAL] = "#676767" bg[PRELIGHT] = shade(0.92,@bg_color) bg[ACTIVE] = "#ff0000" bg[INSENSITIVE] = "#ff0000" bg[SELECTED] = "#ff0000" text[PRELIGHT] = "#000000" text[NORMAL] = "#000000" text[INSENSITIVE]= "#000000" text[SELECTED] = "#ffffff" text[ACTIVE] = "#000000" fg[NORMAL] = "#ffffff" fg[PRELIGHT] = "#ffffff" fg[INSENSITIVE] = "#ffffff" fg[SELECTED] = "#ffffff" fg[ACTIVE] = "#ffffff" base[NORMAL] = "#ff0000" base[NORMAL] = "#aeaeae" base[INSENSITIVE] = "#00ff00" base[SELECTED] = "#f3720d" base[ACTIVE] = "#f3720d" } style "theme-separator-menu-item" { xthickness = 1 ythickness = 0 GtkSeparatorMenuItem::horizontal-padding = 2 # We are setting the desired height by using wide-separators # There is no other way to get the odd height ... GtkWidget::wide-separators = 1 GtkWidget::separator-width = 1 GtkWidget::separator-height = 5 } style "theme-frame" { xthickness = 10 ythickness = 0 GtkWidget::LABEL-SIDE-PAD = 14 GtkWidget::LABEL-PAD = 23 fg[NORMAL] = "#000000" fg[ACTIVE] = "#000000" fg[PRELIGHT] = "#000000" fg[SELECTED] = "#000000" fg[INSENSITIVE] = "#000000" bg[NORMAL] = "#e2e2e2" bg[ACTIVE] = "#000000" bg[PRELIGHT] = "#000000" bg[SELECTED] = "#000000" bg[INSENSITIVE] = "#000000" base[NORMAL] = "#000000" base[ACTIVE] = "#000000" base[PRELIGHT] = "#000000" base[SELECTED] = "#000000" base[INSENSITIVE]= "#000000" text[NORMAL] = "#000000" text[ACTIVE] = "#000000" text[PRELIGHT] = "#000000" text[SELECTED] = "#000000" text[INSENSITIVE]= "#000000" } style "theme-textview" { text[NORMAL] = "#000000" text[ACTIVE] = "#000000" text[PRELIGHT] = "#000000" text[SELECTED] = "#000000" text[INSENSITIVE] = "#434648" bg[NORMAL] = "#ffffff" bg[ACTIVE] = "#ffffff" bg[PRELIGHT] = "#ffffff" bg[SELECTED] = "#ffffff" bg[INSENSITIVE] = "#ffffff" fg[NORMAL] = "#ffffff" fg[ACTIVE] = "#ffffff" fg[PRELIGHT] = "#ffffff" fg[SELECTED] = "#ffffff" fg[INSENSITIVE] = "#ffffff" base[NORMAL] = "#ffffff" base[ACTIVE] = "#ffffff" base[PRELIGHT] = "#ffffff" base[SELECTED] = "#ff9a00" base[INSENSITIVE] = "#ffffff" } style "theme-clist" { text[NORMAL] = "#000000" text[ACTIVE] = "#000000" text[PRELIGHT] = "#000000" text[SELECTED] = "#000000" text[INSENSITIVE] = "#434648" bg[NORMAL] = "#353438" bg[ACTIVE] = "#ff9a00" bg[PRELIGHT] = "#ff9a00" bg[SELECTED] = "#ff9a00" bg[INSENSITIVE] = "#ffffff" fg[NORMAL] = "#000000" fg[ACTIVE] = "#ff9a00" fg[PRELIGHT] = "#ff9a00" fg[SELECTED] = "#fdff00" fg[INSENSITIVE] = "#757575" base[NORMAL] = "#ffffff" base[ACTIVE] = "#fdff00" base[PRELIGHT] = "#000000" base[SELECTED] = "#fdff00" base[INSENSITIVE] = "#757575" } style "theme-label" { bg[NORMAL] = "#414143" bg[ACTIVE] = "#c19676" bg[PRELIGHT] = "#7f4426" bg[SELECTED] = "#000000" bg[INSENSITIVE] = "#434346" fg[NORMAL] = "#000000" fg[INSENSITIVE] = "#434346" fg[PRELIGHT] = "#000000" fg[SELECTED] = "#000000" fg[ACTIVE] = "#000000" text[NORMAL] = "#ffffff" text[INSENSITIVE] = "#434346" text[PRELIGHT] = "#ffffff" text[SELECTED] = "#ffffff" text[ACTIVE] = "#ffffff" base[NORMAL] = "#000000" base[INSENSITIVE] = "#00ff00" base[PRELIGHT] = "#0000ff" base[ACTIVE] = "#f39638" } style "theme-button-label" { bg[NORMAL] = "#414143" bg[ACTIVE] = "#c19676" bg[PRELIGHT] = "#7f4426" bg[SELECTED] = "#000000" bg[INSENSITIVE] = "#434346" fg[NORMAL] = "#ffffff" fg[INSENSITIVE] = "#434346" fg[PRELIGHT] = "#ffffff" fg[SELECTED] = "#ffffff" fg[ACTIVE] = "#ffffff" text[NORMAL] = "#000000" text[INSENSITIVE] = "#434346" text[PRELIGHT] = "#000000" text[SELECTED] = "#000000" text[ACTIVE] = "#000000" base[NORMAL] = "#000000" base[INSENSITIVE] = "#00ff00" base[PRELIGHT] = "#0000ff" base[SELECTED] = "#ff00ff" base[ACTIVE] = "#ffff00" } style "theme-button-check-radio-label" { bg[NORMAL] = "#414143" bg[ACTIVE] = "#c19676" bg[PRELIGHT] = "#7f4426" bg[SELECTED] = "#000000" bg[INSENSITIVE] = "#434346" fg[NORMAL] = "#000000" fg[INSENSITIVE] = "#434346" fg[PRELIGHT] = "#000000" fg[SELECTED] = "#000000" fg[ACTIVE] = "#000000" text[NORMAL] = "#ffffff" text[INSENSITIVE] = "#434346" text[PRELIGHT] = "#ffffff" text[SELECTED] = "#000000" text[ACTIVE] = "#ffffff" base[NORMAL] = "#000000" base[INSENSITIVE] = "#00ff00" base[PRELIGHT] = "#0000ff" base[SELECTED] = "#ff00ff" base[ACTIVE] = "#ffff00" } style "theme-table" { bg[NORMAL] = "#848484" bg[ACTIVE] = "#c19676" bg[PRELIGHT] = "#7f4426" bg[SELECTED] = "#000000" bg[INSENSITIVE] = "#434346" } style "theme-iconview" { GtkWidget::focus-line-width=1 bg[NORMAL] = "#000000" bg[ACTIVE] = "#c19676" bg[PRELIGHT] = "#c19676" bg[SELECTED] = "#c19676" bg[INSENSITIVE] = "#969696" fg[NORMAL] = "#ffffff" fg[INSENSITIVE] = "#ffffff" fg[PRELIGHT] = "#ffffff" fg[SELECTED] = "#ffffff" fg[ACTIVE] = "#ffffff" text[NORMAL] = "#000000" text[INSENSITIVE] = "#434346" text[PRELIGHT] = "#000000" text[SELECTED] = "#000000" text[ACTIVE] = "#000000" base[NORMAL] = "#ffffff" base[INSENSITIVE] = "#434346" base[PRELIGHT] = "#FAD184" base[SELECTED] = "#FAD184" base[ACTIVE] = "#FAD184" } # Set Widget styles class "GtkWidget" style "theme-default" class "GtkScale" style "theme-scale" class "GtkRange" style "theme-range" class "GtkPaned" style "theme-paned" class "GtkFrame" style "theme-frame" class "GtkMenu" style "theme-menu" class "GtkMenuBar" style "theme-menubar" class "GtkEntry" style "theme-entry" class "GtkProgressBar" style "theme-progressbar" class "GtkToolbar" style "theme-toolbar" class "GtkSeparator" style "theme-wide" class "GtkCalendar" style "theme-calendar" class "GtkTable" style "theme-table" widget_class "*<GtkMenuItem>*" style "theme-menu-item" widget_class "*<GtkMenuBar>.<GtkMenuItem>*" style "theme-menubar-item" widget_class "*<GtkSeparatorMenuItem>*" style "theme-separator-menu-item" widget_class "*<GtkLabel>" style "theme-label" widget_class "*<GtkButton>" style "theme-button" widget_class "*<GtkButton>*<GtkLabel>*" style "theme-button-label" widget_class "*<GtkCheckButton>" style "theme-check-radio-buttons" widget_class "*<GtkToggleButton>.<GtkLabel>*" style "theme-button" widget_class "*<GtkCheckButton>.<GtkLabel>*" style "theme-button-check-radio-label" widget_class "*<GtkRadioButton>.<GtkLabel>*" style "theme-button-check-radio-label" widget_class "*<GtkTextView>" style "theme-textview" widget_class "*<GtkList>" style "theme-textview" widget_class "*<GtkCList>" style "theme-clist" widget_class "*<GtkIconView>" style "theme-iconview" widget_class "*<GtkHandleBox>" style "theme-handlebox" widget_class "*<GtkNotebook>" style "theme-notebook" widget_class "*<GtkNotebook>*<GtkEventBox>" style "theme-notebook" widget_class "*<GtkNotebook>*<GtkDrawingArea>" style "theme-notebook" widget_class "*<GtkNotebook>*<GtkLayout>" style "theme-notebook" widget_class "*<GtkNotebook>*<GtkViewport>" style "theme-notebook" widget_class "*<GtkNotebook>.<GtkLabel>*" style "theme-notebook" #for tabs # Combo Box Stuff widget_class "*<GtkCombo>*" style "theme-combo" widget_class "*<GtkComboBox>*<GtkButton>" style "theme-combo-box" widget_class "*<GtkComboBoxEntry>*" style "theme-entry-combo-box" widget_class "*<GtkSpinButton>*" style "theme-spin-button" widget_class "*<GtkSpinButton>*<GtkArrow>*" style:highest "theme-tree-arrow" # Tool Tips Stuff widget "gtk-tooltip*" style "theme-tooltips" # Tree View Stuff widget_class "*<GtkTreeView>.<GtkButton>*" style "theme-tree" widget_class "*<GtkCTree>.<GtkButton>*" style "theme-tree" widget_class "*<GtkList>.<GtkButton>*" style "theme-tree" widget_class "*<GtkCList>.<GtkButton>*" style "theme-tree" # For arrow bg widget_class "*<GtkTreeView>.<GtkButton>*<GtkArrow>" style "theme-tree-arrow" widget_class "*<GtkCTree>.<GtkButton>*<GtkArrow>" style "theme-tree-arrow" widget_class "*<GtkList>.<GtkButton>*<GtkArrow>" style "theme-tree-arrow" ####################################################### ## GNOME specific ####################################################### widget_class "*.ETree.ECanvas" style "theme-tree" widget_class "*.ETable.ECanvas" style "theme-tree" style "panelbuttons" = "theme-button" { # As buttons are draw lower this helps center text xthickness = 3 ythickness = 3 } widget_class "*Panel*<GtkButton>*" style "panelbuttons" style "murrine-fg-is-text-color-workaround" { text[NORMAL] = "#000000" text[ACTIVE] = "#fdff00" text[SELECTED] = "#fdff00" text[INSENSITIVE] = "#757575" bg[SELECTED] = "#b85e03" bg[ACTIVE] = "#b85e03" bg[SELECTED] = "#b85e03" fg[SELECTED] = "#ffffff" fg[NORMAL] = "#ffffff" fg[ACTIVE] = "#ffffff" fg[INSENSITIVE] = "#434348" fg[PRELIGHT] = "#ffffff" base[SELECTED] = "#ff9a00" base[NORMAL] = "#ffffff" base[ACTIVE] = "#ff9a00" base[INSENSITIVE] = "#434348" base[PRELIGHT] = "#ffffff" } widget_class "*.<GtkTreeView>*" style "murrine-fg-is-text-color-workaround" style "murrine-combobox-text-color-workaround" { text[NORMAL] = "#FFFFF" text[PRELIGHT] = "#FFFFF" text[SELECTED] = "#FFFFF" text[ACTIVE] = "#FFFFF" text[INSENSITIVE] = "#FFFFF" } widget_class "*.<GtkComboBox>.<GtkCellView>" style "murrine-combobox-text-color-workaround" style "murrine-menuitem-text-is-fg-color-workaround" { bg[NORMAL] = "#0000ff" text[NORMAL] = "#ffffff" text[PRELIGHT] = "#ffffff"#"#FD7D00" text[SELECTED] = "#ffffff"#"#ff0000"# @selected_fg_color text[ACTIVE] = "#ffffff"#"#ff0000"# "#FD7D00" text[INSENSITIVE] = "#ffffff"#ff0000"# "#414143" } widget "*.gtk-combobox-popup-menu.*" style "murrine-menuitem-text-is-fg-color-workaround"

    Read the article

  • How to manually patch Blogger template to use Disqus

    - by user317944
    I'm trying to add disqus to my blog and I tried following this guide to do so: http://disqus.com/docs/patch-blogger/ However their instructions are completely off with what I have on my custom template. Here is the template: <b:skin><![CDATA[/*----------------------------------------------- Blogger Template Style Name: Picture Window Designer: Josh Peterson URL: www.noaesthetic.com ----------------------------------------------- */ /* Variable definitions ==================== */ /* Content ----------------------------------------------- */ body { font: $(body.font); color: $(body.text.color); } html body .region-inner { min-width: 0; max-width: 100%; width: auto; } .content-outer { font-size: 90%; } a:link { text-decoration:none; color: $(link.color); } a:visited { text-decoration:none; color: $(link.visited.color); } a:hover { text-decoration:underline; color: $(link.hover.color); } .body-fauxcolumn-outer { background: $(body.background); } .content-outer { background: $(content.background); -moz-border-radius: $(content.border.radius); -webkit-border-radius: $(content.border.radius); -goog-ms-border-radius: $(content.border.radius); border-radius: $(content.border.radius); -moz-box-shadow: 0 0 $(content.shadow.spread) rgba(0, 0, 0, .15); -webkit-box-shadow: 0 0 $(content.shadow.spread) rgba(0, 0, 0, .15); -goog-ms-box-shadow: 0 0 $(content.shadow.spread) rgba(0, 0, 0, .15); box-shadow: 0 0 $(content.shadow.spread) rgba(0, 0, 0, .15); margin: $(content.margin) auto; } .content-inner { padding: $(content.padding); } /* Header ----------------------------------------------- */ .header-outer { background: $(header.background.color) $(header.background.gradient) repeat-x scroll top left; _background-image: none; color: $(header.text.color); -moz-border-radius: $(header.border.radius); -webkit-border-radius: $(header.border.radius); -goog-ms-border-radius: $(header.border.radius); border-radius: $(header.border.radius); } .Header img, .Header #header-inner { -moz-border-radius: $(header.border.radius); -webkit-border-radius: $(header.border.radius); -goog-ms-border-radius: $(header.border.radius); border-radius: $(header.border.radius); } .header-inner .Header .titlewrapper, .header-inner .Header .descriptionwrapper { padding-left: $(header.padding); padding-right: $(header.padding); } .Header h1 { font: $(header.font); text-shadow: 1px 1px 3px rgba(0, 0, 0, 0.3); } .Header h1 a { color: $(header.text.color); } .Header .description { font-size: 130%; } /* Tabs ----------------------------------------------- */ .tabs-inner { margin: .5em $(tabs.margin.sides) $(tabs.margin.bottom); padding: 0; } .tabs-inner .section { margin: 0; } .tabs-inner .widget ul { padding: 0; background: $(tabs.background.color) $(tabs.background.gradient) repeat scroll bottom; -moz-border-radius: $(tabs.border.radius); -webkit-border-radius: $(tabs.border.radius); -goog-ms-border-radius: $(tabs.border.radius); border-radius: $(tabs.border.radius); } .tabs-inner .widget li { border: none; } .tabs-inner .widget li a { display: block; padding: .5em 1em; margin-$endSide: $(tabs.spacing); color: $(tabs.text.color); font: $(tabs.font); -moz-border-radius: $(tab.border.radius) $(tab.border.radius) 0 0; -webkit-border-top-left-radius: $(tab.border.radius); -webkit-border-top-right-radius: $(tab.border.radius); -goog-ms-border-radius: $(tab.border.radius) $(tab.border.radius) 0 0; border-radius: $(tab.border.radius) $(tab.border.radius) 0 0; background: $(tab.background); border-$endSide: 1px solid $(tabs.separator.color); } .tabs-inner .widget li:first-child a { padding-$startSide: 1.25em; -moz-border-radius-top$startSide: $(tab.first.border.radius); -moz-border-radius-bottom$startSide: $(tabs.border.radius); -webkit-border-top-$startSide-radius: $(tab.first.border.radius); -webkit-border-bottom-$startSide-radius: $(tabs.border.radius); -goog-ms-border-top-$startSide-radius: $(tab.first.border.radius); -goog-ms-border-bottom-$startSide-radius: $(tabs.border.radius); border-top-$startSide-radius: $(tab.first.border.radius); border-bottom-$startSide-radius: $(tabs.border.radius); } .tabs-inner .widget li.selected a, .tabs-inner .widget li a:hover { position: relative; z-index: 1; background: $(tabs.selected.background.color) $(tab.selected.background.gradient) repeat scroll bottom; color: $(tabs.selected.text.color); -moz-box-shadow: 0 0 $(region.shadow.spread) rgba(0, 0, 0, .15); -webkit-box-shadow: 0 0 $(region.shadow.spread) rgba(0, 0, 0, .15); -goog-ms-box-shadow: 0 0 $(region.shadow.spread) rgba(0, 0, 0, .15); box-shadow: 0 0 $(region.shadow.spread) rgba(0, 0, 0, .15); } /* Headings ----------------------------------------------- */ h2 { font: $(widget.title.font); text-transform: $(widget.title.text.transform); color: $(widget.title.text.color); margin: .5em 0; } /* Main ----------------------------------------------- */ .main-outer { background: $(main.background); -moz-border-radius: $(main.border.radius.top) $(main.border.radius.top) 0 0; -webkit-border-top-left-radius: $(main.border.radius.top); -webkit-border-top-right-radius: $(main.border.radius.top); -webkit-border-bottom-left-radius: 0; -webkit-border-bottom-right-radius: 0; -goog-ms-border-radius: $(main.border.radius.top) $(main.border.radius.top) 0 0; border-radius: $(main.border.radius.top) $(main.border.radius.top) 0 0; -moz-box-shadow: 0 $(region.shadow.offset) $(region.shadow.spread) rgba(0, 0, 0, .15); -webkit-box-shadow: 0 $(region.shadow.offset) $(region.shadow.spread) rgba(0, 0, 0, .15); -goog-ms-box-shadow: 0 $(region.shadow.offset) $(region.shadow.spread) rgba(0, 0, 0, .15); box-shadow: 0 $(region.shadow.offset) $(region.shadow.spread) rgba(0, 0, 0, .15); } .main-inner { padding: 15px $(main.padding.sides) 20px; } .main-inner .column-center-inner { padding: 0 0; } .main-inner .column-left-inner { padding-left: 0; } .main-inner .column-right-inner { padding-right: 0; } /* Posts ----------------------------------------------- */ h3.post-title { margin: 0; font: $(post.title.font); } .comments h4 { margin: 1em 0 0; font: $(post.title.font); } .post-outer { background-color: $(post.background.color); border: solid 1px $(post.border.color); -moz-border-radius: $(post.border.radius); -webkit-border-radius: $(post.border.radius); border-radius: $(post.border.radius); -goog-ms-border-radius: $(post.border.radius); padding: 15px 20px; margin: 0 $(post.margin.sides) 20px; } .post-body { line-height: 1.4; font-size: 110%; } .post-header { margin: 0 0 1.5em; color: $(post.footer.text.color); line-height: 1.6; } .post-footer { margin: .5em 0 0; color: $(post.footer.text.color); line-height: 1.6; } blog-pager { font-size: 140% } comments .comment-author { padding-top: 1.5em; border-top: dashed 1px #ccc; border-top: dashed 1px rgba(128, 128, 128, .5); background-position: 0 1.5em; } comments .comment-author:first-child { padding-top: 0; border-top: none; } .avatar-image-container { margin: .2em 0 0; } /* Widgets ----------------------------------------------- */ .widget ul, .widget #ArchiveList ul.flat { padding: 0; list-style: none; } .widget ul li, .widget #ArchiveList ul.flat li { border-top: dashed 1px #ccc; border-top: dashed 1px rgba(128, 128, 128, .5); } .widget ul li:first-child, .widget #ArchiveList ul.flat li:first-child { border-top: none; } .widget .post-body ul { list-style: disc; } .widget .post-body ul li { border: none; } /* Footer ----------------------------------------------- */ .footer-outer { color:$(footer.text.color); background: $(footer.background); -moz-border-radius: $(footer.border.radius.top) $(footer.border.radius.top) $(footer.border.radius.bottom) $(footer.border.radius.bottom); -webkit-border-top-left-radius: $(footer.border.radius.top); -webkit-border-top-right-radius: $(footer.border.radius.top); -webkit-border-bottom-left-radius: $(footer.border.radius.bottom); -webkit-border-bottom-right-radius: $(footer.border.radius.bottom); -goog-ms-border-radius: $(footer.border.radius.top) $(footer.border.radius.top) $(footer.border.radius.bottom) $(footer.border.radius.bottom); border-radius: $(footer.border.radius.top) $(footer.border.radius.top) $(footer.border.radius.bottom) $(footer.border.radius.bottom); -moz-box-shadow: 0 $(region.shadow.offset) $(region.shadow.spread) rgba(0, 0, 0, .15); -webkit-box-shadow: 0 $(region.shadow.offset) $(region.shadow.spread) rgba(0, 0, 0, .15); -goog-ms-box-shadow: 0 $(region.shadow.offset) $(region.shadow.spread) rgba(0, 0, 0, .15); box-shadow: 0 $(region.shadow.offset) $(region.shadow.spread) rgba(0, 0, 0, .15); } .footer-inner { padding: 10px $(main.padding.sides) 20px; } .footer-outer a { color: $(footer.link.color); } .footer-outer a:visited { color: $(footer.link.visited.color); } .footer-outer a:hover { color: $(footer.link.hover.color); } .footer-outer .widget h2 { color: $(footer.widget.title.text.color); } ]] <b:template-skin> <b:variable default='930px' name='content.width' type='length' value='930px'/> <b:variable default='0' name='main.column.left.width' type='length' value='180px'/> <b:variable default='360px' name='main.column.right.width' type='length' value='180px'/> <![CDATA[ body { min-width: $(content.width); } .content-outer, .region-inner { min-width: $(content.width); max-width: $(content.width); _width: $(content.width); } .main-inner .columns { padding-left: $(main.column.left.width); padding-right: $(main.column.right.width); } .main-inner .fauxcolumn-center-outer { left: $(main.column.left.width); right: $(main.column.right.width); /* IE6 does not respect left and right together */ _width: expression(this.parentNode.offsetWidth - parseInt("$(main.column.left.width)") - parseInt("$(main.column.right.width)") + 'px'); } .main-inner .fauxcolumn-left-outer { width: $(main.column.left.width); } .main-inner .fauxcolumn-right-outer { width: $(main.column.right.width); } .main-inner .column-left-outer { width: $(main.column.left.width); right: $(main.column.left.width); margin-right: -$(main.column.left.width); } .main-inner .column-right-outer { width: $(main.column.right.width); margin-right: -$(main.column.right.width); } #layout { min-width: 0; } #layout .content-outer { min-width: 0; width: 800px; } #layout .region-inner { min-width: 0; width: auto; } ]]> </b:template-skin> <div class='main-cap-bottom cap-bottom'> <div class='cap-left'/> <div class='cap-right'/> </div> </div> <footer> <div class='footer-outer'> <div class='footer-cap-top cap-top'> <div class='cap-left'/> <div class='cap-right'/> </div> <div class='fauxborder-left footer-fauxborder-left'> <div class='fauxborder-right footer-fauxborder-right'/> <div class='region-inner footer-inner'> <macro:include id='footer-sections' name='sections'> <macro:param default='2' name='num' value='3'/> <macro:param default='footer' name='idPrefix'/> <macro:param default='foot' name='class'/> <macro:param default='false' name='includeBottom'/> </macro:include> <!-- outside of the include in order to lock Attribution widget --> <b:section class='foot' id='footer-3' showaddelement='no'> document.body.className = document.body.className.replace('loading', ''); <macro:if cond='data:col.num &gt;= 2'> <table border='0' cellpadding='0' cellspacing='0' mexpr:class='&quot;section-columns columns-&quot; + data:col.num'> <tbody> <tr> <td class='first columns-cell'> <b:section mexpr:class='data:col.class' mexpr:id='data:col.idPrefix + &quot;-2-1&quot;'/> </td> <td class='columns-cell'> <b:section mexpr:class='data:col.class' mexpr:id='data:col.idPrefix + &quot;-2-2&quot;'/> </td> <macro:if cond='data:col.num &gt;= 3'> <td class='columns-cell'> <b:section mexpr:class='data:col.class' mexpr:id='data:col.idPrefix + &quot;-2-3&quot;'/> </td> </macro:if> <macro:if cond='data:col.num &gt;= 4'> <td class='columns-cell'> <b:section mexpr:class='data:col.class' mexpr:id='data:col.idPrefix + &quot;-2-4&quot;'/> </td> </macro:if> </tr> </tbody> </table> <macro:if cond='data:col.includeBottom'> <b:section mexpr:class='data:col.class' mexpr:id='data:col.idPrefix + &quot;-3&quot;' showaddelement='no'/> </macro:if> </macro

    Read the article

  • Improving Partitioned Table Join Performance

    - by Paul White
    The query optimizer does not always choose an optimal strategy when joining partitioned tables. This post looks at an example, showing how a manual rewrite of the query can almost double performance, while reducing the memory grant to almost nothing. Test Data The two tables in this example use a common partitioning partition scheme. The partition function uses 41 equal-size partitions: CREATE PARTITION FUNCTION PFT (integer) AS RANGE RIGHT FOR VALUES ( 125000, 250000, 375000, 500000, 625000, 750000, 875000, 1000000, 1125000, 1250000, 1375000, 1500000, 1625000, 1750000, 1875000, 2000000, 2125000, 2250000, 2375000, 2500000, 2625000, 2750000, 2875000, 3000000, 3125000, 3250000, 3375000, 3500000, 3625000, 3750000, 3875000, 4000000, 4125000, 4250000, 4375000, 4500000, 4625000, 4750000, 4875000, 5000000 ); GO CREATE PARTITION SCHEME PST AS PARTITION PFT ALL TO ([PRIMARY]); There two tables are: CREATE TABLE dbo.T1 ( TID integer NOT NULL IDENTITY(0,1), Column1 integer NOT NULL, Padding binary(100) NOT NULL DEFAULT 0x,   CONSTRAINT PK_T1 PRIMARY KEY CLUSTERED (TID) ON PST (TID) );   CREATE TABLE dbo.T2 ( TID integer NOT NULL, Column1 integer NOT NULL, Padding binary(100) NOT NULL DEFAULT 0x,   CONSTRAINT PK_T2 PRIMARY KEY CLUSTERED (TID, Column1) ON PST (TID) ); The next script loads 5 million rows into T1 with a pseudo-random value between 1 and 5 for Column1. The table is partitioned on the IDENTITY column TID: INSERT dbo.T1 WITH (TABLOCKX) (Column1) SELECT (ABS(CHECKSUM(NEWID())) % 5) + 1 FROM dbo.Numbers AS N WHERE n BETWEEN 1 AND 5000000; In case you don’t already have an auxiliary table of numbers lying around, here’s a script to create one with 10 million rows: CREATE TABLE dbo.Numbers (n bigint PRIMARY KEY);   WITH L0 AS(SELECT 1 AS c UNION ALL SELECT 1), L1 AS(SELECT 1 AS c FROM L0 AS A CROSS JOIN L0 AS B), L2 AS(SELECT 1 AS c FROM L1 AS A CROSS JOIN L1 AS B), L3 AS(SELECT 1 AS c FROM L2 AS A CROSS JOIN L2 AS B), L4 AS(SELECT 1 AS c FROM L3 AS A CROSS JOIN L3 AS B), L5 AS(SELECT 1 AS c FROM L4 AS A CROSS JOIN L4 AS B), Nums AS(SELECT ROW_NUMBER() OVER (ORDER BY (SELECT NULL)) AS n FROM L5) INSERT dbo.Numbers WITH (TABLOCKX) SELECT TOP (10000000) n FROM Nums ORDER BY n OPTION (MAXDOP 1); Table T1 contains data like this: Next we load data into table T2. The relationship between the two tables is that table 2 contains ‘n’ rows for each row in table 1, where ‘n’ is determined by the value in Column1 of table T1. There is nothing particularly special about the data or distribution, by the way. INSERT dbo.T2 WITH (TABLOCKX) (TID, Column1) SELECT T.TID, N.n FROM dbo.T1 AS T JOIN dbo.Numbers AS N ON N.n >= 1 AND N.n <= T.Column1; Table T2 ends up containing about 15 million rows: The primary key for table T2 is a combination of TID and Column1. The data is partitioned according to the value in column TID alone. Partition Distribution The following query shows the number of rows in each partition of table T1: SELECT PartitionID = CA1.P, NumRows = COUNT_BIG(*) FROM dbo.T1 AS T CROSS APPLY (VALUES ($PARTITION.PFT(TID))) AS CA1 (P) GROUP BY CA1.P ORDER BY CA1.P; There are 40 partitions containing 125,000 rows (40 * 125k = 5m rows). The rightmost partition remains empty. The next query shows the distribution for table 2: SELECT PartitionID = CA1.P, NumRows = COUNT_BIG(*) FROM dbo.T2 AS T CROSS APPLY (VALUES ($PARTITION.PFT(TID))) AS CA1 (P) GROUP BY CA1.P ORDER BY CA1.P; There are roughly 375,000 rows in each partition (the rightmost partition is also empty): Ok, that’s the test data done. Test Query and Execution Plan The task is to count the rows resulting from joining tables 1 and 2 on the TID column: SET STATISTICS IO ON; DECLARE @s datetime2 = SYSUTCDATETIME();   SELECT COUNT_BIG(*) FROM dbo.T1 AS T1 JOIN dbo.T2 AS T2 ON T2.TID = T1.TID;   SELECT DATEDIFF(Millisecond, @s, SYSUTCDATETIME()); SET STATISTICS IO OFF; The optimizer chooses a plan using parallel hash join, and partial aggregation: The Plan Explorer plan tree view shows accurate cardinality estimates and an even distribution of rows across threads (click to enlarge the image): With a warm data cache, the STATISTICS IO output shows that no physical I/O was needed, and all 41 partitions were touched: Running the query without actual execution plan or STATISTICS IO information for maximum performance, the query returns in around 2600ms. Execution Plan Analysis The first step toward improving on the execution plan produced by the query optimizer is to understand how it works, at least in outline. The two parallel Clustered Index Scans use multiple threads to read rows from tables T1 and T2. Parallel scan uses a demand-based scheme where threads are given page(s) to scan from the table as needed. This arrangement has certain important advantages, but does result in an unpredictable distribution of rows amongst threads. The point is that multiple threads cooperate to scan the whole table, but it is impossible to predict which rows end up on which threads. For correct results from the parallel hash join, the execution plan has to ensure that rows from T1 and T2 that might join are processed on the same thread. For example, if a row from T1 with join key value ‘1234’ is placed in thread 5’s hash table, the execution plan must guarantee that any rows from T2 that also have join key value ‘1234’ probe thread 5’s hash table for matches. The way this guarantee is enforced in this parallel hash join plan is by repartitioning rows to threads after each parallel scan. The two repartitioning exchanges route rows to threads using a hash function over the hash join keys. The two repartitioning exchanges use the same hash function so rows from T1 and T2 with the same join key must end up on the same hash join thread. Expensive Exchanges This business of repartitioning rows between threads can be very expensive, especially if a large number of rows is involved. The execution plan selected by the optimizer moves 5 million rows through one repartitioning exchange and around 15 million across the other. As a first step toward removing these exchanges, consider the execution plan selected by the optimizer if we join just one partition from each table, disallowing parallelism: SELECT COUNT_BIG(*) FROM dbo.T1 AS T1 JOIN dbo.T2 AS T2 ON T2.TID = T1.TID WHERE $PARTITION.PFT(T1.TID) = 1 AND $PARTITION.PFT(T2.TID) = 1 OPTION (MAXDOP 1); The optimizer has chosen a (one-to-many) merge join instead of a hash join. The single-partition query completes in around 100ms. If everything scaled linearly, we would expect that extending this strategy to all 40 populated partitions would result in an execution time around 4000ms. Using parallelism could reduce that further, perhaps to be competitive with the parallel hash join chosen by the optimizer. This raises a question. If the most efficient way to join one partition from each of the tables is to use a merge join, why does the optimizer not choose a merge join for the full query? Forcing a Merge Join Let’s force the optimizer to use a merge join on the test query using a hint: SELECT COUNT_BIG(*) FROM dbo.T1 AS T1 JOIN dbo.T2 AS T2 ON T2.TID = T1.TID OPTION (MERGE JOIN); This is the execution plan selected by the optimizer: This plan results in the same number of logical reads reported previously, but instead of 2600ms the query takes 5000ms. The natural explanation for this drop in performance is that the merge join plan is only using a single thread, whereas the parallel hash join plan could use multiple threads. Parallel Merge Join We can get a parallel merge join plan using the same query hint as before, and adding trace flag 8649: SELECT COUNT_BIG(*) FROM dbo.T1 AS T1 JOIN dbo.T2 AS T2 ON T2.TID = T1.TID OPTION (MERGE JOIN, QUERYTRACEON 8649); The execution plan is: This looks promising. It uses a similar strategy to distribute work across threads as seen for the parallel hash join. In practice though, performance is disappointing. On a typical run, the parallel merge plan runs for around 8400ms; slower than the single-threaded merge join plan (5000ms) and much worse than the 2600ms for the parallel hash join. We seem to be going backwards! The logical reads for the parallel merge are still exactly the same as before, with no physical IOs. The cardinality estimates and thread distribution are also still very good (click to enlarge): A big clue to the reason for the poor performance is shown in the wait statistics (captured by Plan Explorer Pro): CXPACKET waits require careful interpretation, and are most often benign, but in this case excessive waiting occurs at the repartitioning exchanges. Unlike the parallel hash join, the repartitioning exchanges in this plan are order-preserving ‘merging’ exchanges (because merge join requires ordered inputs): Parallelism works best when threads can just grab any available unit of work and get on with processing it. Preserving order introduces inter-thread dependencies that can easily lead to significant waits occurring. In extreme cases, these dependencies can result in an intra-query deadlock, though the details of that will have to wait for another time to explore in detail. The potential for waits and deadlocks leads the query optimizer to cost parallel merge join relatively highly, especially as the degree of parallelism (DOP) increases. This high costing resulted in the optimizer choosing a serial merge join rather than parallel in this case. The test results certainly confirm its reasoning. Collocated Joins In SQL Server 2008 and later, the optimizer has another available strategy when joining tables that share a common partition scheme. This strategy is a collocated join, also known as as a per-partition join. It can be applied in both serial and parallel execution plans, though it is limited to 2-way joins in the current optimizer. Whether the optimizer chooses a collocated join or not depends on cost estimation. The primary benefits of a collocated join are that it eliminates an exchange and requires less memory, as we will see next. Costing and Plan Selection The query optimizer did consider a collocated join for our original query, but it was rejected on cost grounds. The parallel hash join with repartitioning exchanges appeared to be a cheaper option. There is no query hint to force a collocated join, so we have to mess with the costing framework to produce one for our test query. Pretending that IOs cost 50 times more than usual is enough to convince the optimizer to use collocated join with our test query: -- Pretend IOs are 50x cost temporarily DBCC SETIOWEIGHT(50);   -- Co-located hash join SELECT COUNT_BIG(*) FROM dbo.T1 AS T1 JOIN dbo.T2 AS T2 ON T2.TID = T1.TID OPTION (RECOMPILE);   -- Reset IO costing DBCC SETIOWEIGHT(1); Collocated Join Plan The estimated execution plan for the collocated join is: The Constant Scan contains one row for each partition of the shared partitioning scheme, from 1 to 41. The hash repartitioning exchanges seen previously are replaced by a single Distribute Streams exchange using Demand partitioning. Demand partitioning means that the next partition id is given to the next parallel thread that asks for one. My test machine has eight logical processors, and all are available for SQL Server to use. As a result, there are eight threads in the single parallel branch in this plan, each processing one partition from each table at a time. Once a thread finishes processing a partition, it grabs a new partition number from the Distribute Streams exchange…and so on until all partitions have been processed. It is important to understand that the parallel scans in this plan are different from the parallel hash join plan. Although the scans have the same parallelism icon, tables T1 and T2 are not being co-operatively scanned by multiple threads in the same way. Each thread reads a single partition of T1 and performs a hash match join with the same partition from table T2. The properties of the two Clustered Index Scans show a Seek Predicate (unusual for a scan!) limiting the rows to a single partition: The crucial point is that the join between T1 and T2 is on TID, and TID is the partitioning column for both tables. A thread that processes partition ‘n’ is guaranteed to see all rows that can possibly join on TID for that partition. In addition, no other thread will see rows from that partition, so this removes the need for repartitioning exchanges. CPU and Memory Efficiency Improvements The collocated join has removed two expensive repartitioning exchanges and added a single exchange processing 41 rows (one for each partition id). Remember, the parallel hash join plan exchanges had to process 5 million and 15 million rows. The amount of processor time spent on exchanges will be much lower in the collocated join plan. In addition, the collocated join plan has a maximum of 8 threads processing single partitions at any one time. The 41 partitions will all be processed eventually, but a new partition is not started until a thread asks for it. Threads can reuse hash table memory for the new partition. The parallel hash join plan also had 8 hash tables, but with all 5,000,000 build rows loaded at the same time. The collocated plan needs memory for only 8 * 125,000 = 1,000,000 rows at any one time. Collocated Hash Join Performance The collated join plan has disappointing performance in this case. The query runs for around 25,300ms despite the same IO statistics as usual. This is much the worst result so far, so what went wrong? It turns out that cardinality estimation for the single partition scans of table T1 is slightly low. The properties of the Clustered Index Scan of T1 (graphic immediately above) show the estimation was for 121,951 rows. This is a small shortfall compared with the 125,000 rows actually encountered, but it was enough to cause the hash join to spill to physical tempdb: A level 1 spill doesn’t sound too bad, until you realize that the spill to tempdb probably occurs for each of the 41 partitions. As a side note, the cardinality estimation error is a little surprising because the system tables accurately show there are 125,000 rows in every partition of T1. Unfortunately, the optimizer uses regular column and index statistics to derive cardinality estimates here rather than system table information (e.g. sys.partitions). Collocated Merge Join We will never know how well the collocated parallel hash join plan might have worked without the cardinality estimation error (and the resulting 41 spills to tempdb) but we do know: Merge join does not require a memory grant; and Merge join was the optimizer’s preferred join option for a single partition join Putting this all together, what we would really like to see is the same collocated join strategy, but using merge join instead of hash join. Unfortunately, the current query optimizer cannot produce a collocated merge join; it only knows how to do collocated hash join. So where does this leave us? CROSS APPLY sys.partitions We can try to write our own collocated join query. We can use sys.partitions to find the partition numbers, and CROSS APPLY to get a count per partition, with a final step to sum the partial counts. The following query implements this idea: SELECT row_count = SUM(Subtotals.cnt) FROM ( -- Partition numbers SELECT p.partition_number FROM sys.partitions AS p WHERE p.[object_id] = OBJECT_ID(N'T1', N'U') AND p.index_id = 1 ) AS P CROSS APPLY ( -- Count per collocated join SELECT cnt = COUNT_BIG(*) FROM dbo.T1 AS T1 JOIN dbo.T2 AS T2 ON T2.TID = T1.TID WHERE $PARTITION.PFT(T1.TID) = p.partition_number AND $PARTITION.PFT(T2.TID) = p.partition_number ) AS SubTotals; The estimated plan is: The cardinality estimates aren’t all that good here, especially the estimate for the scan of the system table underlying the sys.partitions view. Nevertheless, the plan shape is heading toward where we would like to be. Each partition number from the system table results in a per-partition scan of T1 and T2, a one-to-many Merge Join, and a Stream Aggregate to compute the partial counts. The final Stream Aggregate just sums the partial counts. Execution time for this query is around 3,500ms, with the same IO statistics as always. This compares favourably with 5,000ms for the serial plan produced by the optimizer with the OPTION (MERGE JOIN) hint. This is another case of the sum of the parts being less than the whole – summing 41 partial counts from 41 single-partition merge joins is faster than a single merge join and count over all partitions. Even so, this single-threaded collocated merge join is not as quick as the original parallel hash join plan, which executed in 2,600ms. On the positive side, our collocated merge join uses only one logical processor and requires no memory grant. The parallel hash join plan used 16 threads and reserved 569 MB of memory:   Using a Temporary Table Our collocated merge join plan should benefit from parallelism. The reason parallelism is not being used is that the query references a system table. We can work around that by writing the partition numbers to a temporary table (or table variable): SET STATISTICS IO ON; DECLARE @s datetime2 = SYSUTCDATETIME();   CREATE TABLE #P ( partition_number integer PRIMARY KEY);   INSERT #P (partition_number) SELECT p.partition_number FROM sys.partitions AS p WHERE p.[object_id] = OBJECT_ID(N'T1', N'U') AND p.index_id = 1;   SELECT row_count = SUM(Subtotals.cnt) FROM #P AS p CROSS APPLY ( SELECT cnt = COUNT_BIG(*) FROM dbo.T1 AS T1 JOIN dbo.T2 AS T2 ON T2.TID = T1.TID WHERE $PARTITION.PFT(T1.TID) = p.partition_number AND $PARTITION.PFT(T2.TID) = p.partition_number ) AS SubTotals;   DROP TABLE #P;   SELECT DATEDIFF(Millisecond, @s, SYSUTCDATETIME()); SET STATISTICS IO OFF; Using the temporary table adds a few logical reads, but the overall execution time is still around 3500ms, indistinguishable from the same query without the temporary table. The problem is that the query optimizer still doesn’t choose a parallel plan for this query, though the removal of the system table reference means that it could if it chose to: In fact the optimizer did enter the parallel plan phase of query optimization (running search 1 for a second time): Unfortunately, the parallel plan found seemed to be more expensive than the serial plan. This is a crazy result, caused by the optimizer’s cost model not reducing operator CPU costs on the inner side of a nested loops join. Don’t get me started on that, we’ll be here all night. In this plan, everything expensive happens on the inner side of a nested loops join. Without a CPU cost reduction to compensate for the added cost of exchange operators, candidate parallel plans always look more expensive to the optimizer than the equivalent serial plan. Parallel Collocated Merge Join We can produce the desired parallel plan using trace flag 8649 again: SELECT row_count = SUM(Subtotals.cnt) FROM #P AS p CROSS APPLY ( SELECT cnt = COUNT_BIG(*) FROM dbo.T1 AS T1 JOIN dbo.T2 AS T2 ON T2.TID = T1.TID WHERE $PARTITION.PFT(T1.TID) = p.partition_number AND $PARTITION.PFT(T2.TID) = p.partition_number ) AS SubTotals OPTION (QUERYTRACEON 8649); The actual execution plan is: One difference between this plan and the collocated hash join plan is that a Repartition Streams exchange operator is used instead of Distribute Streams. The effect is similar, though not quite identical. The Repartition uses round-robin partitioning, meaning the next partition id is pushed to the next thread in sequence. The Distribute Streams exchange seen earlier used Demand partitioning, meaning the next partition id is pulled across the exchange by the next thread that is ready for more work. There are subtle performance implications for each partitioning option, but going into that would again take us too far off the main point of this post. Performance The important thing is the performance of this parallel collocated merge join – just 1350ms on a typical run. The list below shows all the alternatives from this post (all timings include creation, population, and deletion of the temporary table where appropriate) from quickest to slowest: Collocated parallel merge join: 1350ms Parallel hash join: 2600ms Collocated serial merge join: 3500ms Serial merge join: 5000ms Parallel merge join: 8400ms Collated parallel hash join: 25,300ms (hash spill per partition) The parallel collocated merge join requires no memory grant (aside from a paltry 1.2MB used for exchange buffers). This plan uses 16 threads at DOP 8; but 8 of those are (rather pointlessly) allocated to the parallel scan of the temporary table. These are minor concerns, but it turns out there is a way to address them if it bothers you. Parallel Collocated Merge Join with Demand Partitioning This final tweak replaces the temporary table with a hard-coded list of partition ids (dynamic SQL could be used to generate this query from sys.partitions): SELECT row_count = SUM(Subtotals.cnt) FROM ( VALUES (1),(2),(3),(4),(5),(6),(7),(8),(9),(10), (11),(12),(13),(14),(15),(16),(17),(18),(19),(20), (21),(22),(23),(24),(25),(26),(27),(28),(29),(30), (31),(32),(33),(34),(35),(36),(37),(38),(39),(40),(41) ) AS P (partition_number) CROSS APPLY ( SELECT cnt = COUNT_BIG(*) FROM dbo.T1 AS T1 JOIN dbo.T2 AS T2 ON T2.TID = T1.TID WHERE $PARTITION.PFT(T1.TID) = p.partition_number AND $PARTITION.PFT(T2.TID) = p.partition_number ) AS SubTotals OPTION (QUERYTRACEON 8649); The actual execution plan is: The parallel collocated hash join plan is reproduced below for comparison: The manual rewrite has another advantage that has not been mentioned so far: the partial counts (per partition) can be computed earlier than the partial counts (per thread) in the optimizer’s collocated join plan. The earlier aggregation is performed by the extra Stream Aggregate under the nested loops join. The performance of the parallel collocated merge join is unchanged at around 1350ms. Final Words It is a shame that the current query optimizer does not consider a collocated merge join (Connect item closed as Won’t Fix). The example used in this post showed an improvement in execution time from 2600ms to 1350ms using a modestly-sized data set and limited parallelism. In addition, the memory requirement for the query was almost completely eliminated  – down from 569MB to 1.2MB. The problem with the parallel hash join selected by the optimizer is that it attempts to process the full data set all at once (albeit using eight threads). It requires a large memory grant to hold all 5 million rows from table T1 across the eight hash tables, and does not take advantage of the divide-and-conquer opportunity offered by the common partitioning. The great thing about the collocated join strategies is that each parallel thread works on a single partition from both tables, reading rows, performing the join, and computing a per-partition subtotal, before moving on to a new partition. From a thread’s point of view… If you have trouble visualizing what is happening from just looking at the parallel collocated merge join execution plan, let’s look at it again, but from the point of view of just one thread operating between the two Parallelism (exchange) operators. Our thread picks up a single partition id from the Distribute Streams exchange, and starts a merge join using ordered rows from partition 1 of table T1 and partition 1 of table T2. By definition, this is all happening on a single thread. As rows join, they are added to a (per-partition) count in the Stream Aggregate immediately above the Merge Join. Eventually, either T1 (partition 1) or T2 (partition 1) runs out of rows and the merge join stops. The per-partition count from the aggregate passes on through the Nested Loops join to another Stream Aggregate, which is maintaining a per-thread subtotal. Our same thread now picks up a new partition id from the exchange (say it gets id 9 this time). The count in the per-partition aggregate is reset to zero, and the processing of partition 9 of both tables proceeds just as it did for partition 1, and on the same thread. Each thread picks up a single partition id and processes all the data for that partition, completely independently from other threads working on other partitions. One thread might eventually process partitions (1, 9, 17, 25, 33, 41) while another is concurrently processing partitions (2, 10, 18, 26, 34) and so on for the other six threads at DOP 8. The point is that all 8 threads can execute independently and concurrently, continuing to process new partitions until the wider job (of which the thread has no knowledge!) is done. This divide-and-conquer technique can be much more efficient than simply splitting the entire workload across eight threads all at once. Related Reading Understanding and Using Parallelism in SQL Server Parallel Execution Plans Suck © 2013 Paul White – All Rights Reserved Twitter: @SQL_Kiwi

    Read the article

  • Advanced TSQL Tuning: Why Internals Knowledge Matters

    - by Paul White
    There is much more to query tuning than reducing logical reads and adding covering nonclustered indexes.  Query tuning is not complete as soon as the query returns results quickly in the development or test environments.  In production, your query will compete for memory, CPU, locks, I/O and other resources on the server.  Today’s entry looks at some tuning considerations that are often overlooked, and shows how deep internals knowledge can help you write better TSQL. As always, we’ll need some example data.  In fact, we are going to use three tables today, each of which is structured like this: Each table has 50,000 rows made up of an INTEGER id column and a padding column containing 3,999 characters in every row.  The only difference between the three tables is in the type of the padding column: the first table uses CHAR(3999), the second uses VARCHAR(MAX), and the third uses the deprecated TEXT type.  A script to create a database with the three tables and load the sample data follows: USE master; GO IF DB_ID('SortTest') IS NOT NULL DROP DATABASE SortTest; GO CREATE DATABASE SortTest COLLATE LATIN1_GENERAL_BIN; GO ALTER DATABASE SortTest MODIFY FILE ( NAME = 'SortTest', SIZE = 3GB, MAXSIZE = 3GB ); GO ALTER DATABASE SortTest MODIFY FILE ( NAME = 'SortTest_log', SIZE = 256MB, MAXSIZE = 1GB, FILEGROWTH = 128MB ); GO ALTER DATABASE SortTest SET ALLOW_SNAPSHOT_ISOLATION OFF ; ALTER DATABASE SortTest SET AUTO_CLOSE OFF ; ALTER DATABASE SortTest SET AUTO_CREATE_STATISTICS ON ; ALTER DATABASE SortTest SET AUTO_SHRINK OFF ; ALTER DATABASE SortTest SET AUTO_UPDATE_STATISTICS ON ; ALTER DATABASE SortTest SET AUTO_UPDATE_STATISTICS_ASYNC ON ; ALTER DATABASE SortTest SET PARAMETERIZATION SIMPLE ; ALTER DATABASE SortTest SET READ_COMMITTED_SNAPSHOT OFF ; ALTER DATABASE SortTest SET MULTI_USER ; ALTER DATABASE SortTest SET RECOVERY SIMPLE ; USE SortTest; GO CREATE TABLE dbo.TestCHAR ( id INTEGER IDENTITY (1,1) NOT NULL, padding CHAR(3999) NOT NULL,   CONSTRAINT [PK dbo.TestCHAR (id)] PRIMARY KEY CLUSTERED (id), ) ; CREATE TABLE dbo.TestMAX ( id INTEGER IDENTITY (1,1) NOT NULL, padding VARCHAR(MAX) NOT NULL,   CONSTRAINT [PK dbo.TestMAX (id)] PRIMARY KEY CLUSTERED (id), ) ; CREATE TABLE dbo.TestTEXT ( id INTEGER IDENTITY (1,1) NOT NULL, padding TEXT NOT NULL,   CONSTRAINT [PK dbo.TestTEXT (id)] PRIMARY KEY CLUSTERED (id), ) ; -- ============= -- Load TestCHAR (about 3s) -- ============= INSERT INTO dbo.TestCHAR WITH (TABLOCKX) ( padding ) SELECT padding = REPLICATE(CHAR(65 + (Data.n % 26)), 3999) FROM ( SELECT TOP (50000) n = ROW_NUMBER() OVER (ORDER BY (SELECT 0)) - 1 FROM master.sys.columns C1, master.sys.columns C2, master.sys.columns C3 ORDER BY n ASC ) AS Data ORDER BY Data.n ASC ; -- ============ -- Load TestMAX (about 3s) -- ============ INSERT INTO dbo.TestMAX WITH (TABLOCKX) ( padding ) SELECT CONVERT(VARCHAR(MAX), padding) FROM dbo.TestCHAR ORDER BY id ; -- ============= -- Load TestTEXT (about 5s) -- ============= INSERT INTO dbo.TestTEXT WITH (TABLOCKX) ( padding ) SELECT CONVERT(TEXT, padding) FROM dbo.TestCHAR ORDER BY id ; -- ========== -- Space used -- ========== -- EXECUTE sys.sp_spaceused @objname = 'dbo.TestCHAR'; EXECUTE sys.sp_spaceused @objname = 'dbo.TestMAX'; EXECUTE sys.sp_spaceused @objname = 'dbo.TestTEXT'; ; CHECKPOINT ; That takes around 15 seconds to run, and shows the space allocated to each table in its output: To illustrate the points I want to make today, the example task we are going to set ourselves is to return a random set of 150 rows from each table.  The basic shape of the test query is the same for each of the three test tables: SELECT TOP (150) T.id, T.padding FROM dbo.Test AS T ORDER BY NEWID() OPTION (MAXDOP 1) ; Test 1 – CHAR(3999) Running the template query shown above using the TestCHAR table as the target, we find that the query takes around 5 seconds to return its results.  This seems slow, considering that the table only has 50,000 rows.  Working on the assumption that generating a GUID for each row is a CPU-intensive operation, we might try enabling parallelism to see if that speeds up the response time.  Running the query again (but without the MAXDOP 1 hint) on a machine with eight logical processors, the query now takes 10 seconds to execute – twice as long as when run serially. Rather than attempting further guesses at the cause of the slowness, let’s go back to serial execution and add some monitoring.  The script below monitors STATISTICS IO output and the amount of tempdb used by the test query.  We will also run a Profiler trace to capture any warnings generated during query execution. DECLARE @read BIGINT, @write BIGINT ; SELECT @read = SUM(num_of_bytes_read), @write = SUM(num_of_bytes_written) FROM tempdb.sys.database_files AS DBF JOIN sys.dm_io_virtual_file_stats(2, NULL) AS FS ON FS.file_id = DBF.file_id WHERE DBF.type_desc = 'ROWS' ; SET STATISTICS IO ON ; SELECT TOP (150) TC.id, TC.padding FROM dbo.TestCHAR AS TC ORDER BY NEWID() OPTION (MAXDOP 1) ; SET STATISTICS IO OFF ; SELECT tempdb_read_MB = (SUM(num_of_bytes_read) - @read) / 1024. / 1024., tempdb_write_MB = (SUM(num_of_bytes_written) - @write) / 1024. / 1024., internal_use_MB = ( SELECT internal_objects_alloc_page_count / 128.0 FROM sys.dm_db_task_space_usage WHERE session_id = @@SPID ) FROM tempdb.sys.database_files AS DBF JOIN sys.dm_io_virtual_file_stats(2, NULL) AS FS ON FS.file_id = DBF.file_id WHERE DBF.type_desc = 'ROWS' ; Let’s take a closer look at the statistics and query plan generated from this: Following the flow of the data from right to left, we see the expected 50,000 rows emerging from the Clustered Index Scan, with a total estimated size of around 191MB.  The Compute Scalar adds a column containing a random GUID (generated from the NEWID() function call) for each row.  With this extra column in place, the size of the data arriving at the Sort operator is estimated to be 192MB. Sort is a blocking operator – it has to examine all of the rows on its input before it can produce its first row of output (the last row received might sort first).  This characteristic means that Sort requires a memory grant – memory allocated for the query’s use by SQL Server just before execution starts.  In this case, the Sort is the only memory-consuming operator in the plan, so it has access to the full 243MB (248,696KB) of memory reserved by SQL Server for this query execution. Notice that the memory grant is significantly larger than the expected size of the data to be sorted.  SQL Server uses a number of techniques to speed up sorting, some of which sacrifice size for comparison speed.  Sorts typically require a very large number of comparisons, so this is usually a very effective optimization.  One of the drawbacks is that it is not possible to exactly predict the sort space needed, as it depends on the data itself.  SQL Server takes an educated guess based on data types, sizes, and the number of rows expected, but the algorithm is not perfect. In spite of the large memory grant, the Profiler trace shows a Sort Warning event (indicating that the sort ran out of memory), and the tempdb usage monitor shows that 195MB of tempdb space was used – all of that for system use.  The 195MB represents physical write activity on tempdb, because SQL Server strictly enforces memory grants – a query cannot ‘cheat’ and effectively gain extra memory by spilling to tempdb pages that reside in memory.  Anyway, the key point here is that it takes a while to write 195MB to disk, and this is the main reason that the query takes 5 seconds overall. If you are wondering why using parallelism made the problem worse, consider that eight threads of execution result in eight concurrent partial sorts, each receiving one eighth of the memory grant.  The eight sorts all spilled to tempdb, resulting in inefficiencies as the spilled sorts competed for disk resources.  More importantly, there are specific problems at the point where the eight partial results are combined, but I’ll cover that in a future post. CHAR(3999) Performance Summary: 5 seconds elapsed time 243MB memory grant 195MB tempdb usage 192MB estimated sort set 25,043 logical reads Sort Warning Test 2 – VARCHAR(MAX) We’ll now run exactly the same test (with the additional monitoring) on the table using a VARCHAR(MAX) padding column: DECLARE @read BIGINT, @write BIGINT ; SELECT @read = SUM(num_of_bytes_read), @write = SUM(num_of_bytes_written) FROM tempdb.sys.database_files AS DBF JOIN sys.dm_io_virtual_file_stats(2, NULL) AS FS ON FS.file_id = DBF.file_id WHERE DBF.type_desc = 'ROWS' ; SET STATISTICS IO ON ; SELECT TOP (150) TM.id, TM.padding FROM dbo.TestMAX AS TM ORDER BY NEWID() OPTION (MAXDOP 1) ; SET STATISTICS IO OFF ; SELECT tempdb_read_MB = (SUM(num_of_bytes_read) - @read) / 1024. / 1024., tempdb_write_MB = (SUM(num_of_bytes_written) - @write) / 1024. / 1024., internal_use_MB = ( SELECT internal_objects_alloc_page_count / 128.0 FROM sys.dm_db_task_space_usage WHERE session_id = @@SPID ) FROM tempdb.sys.database_files AS DBF JOIN sys.dm_io_virtual_file_stats(2, NULL) AS FS ON FS.file_id = DBF.file_id WHERE DBF.type_desc = 'ROWS' ; This time the query takes around 8 seconds to complete (3 seconds longer than Test 1).  Notice that the estimated row and data sizes are very slightly larger, and the overall memory grant has also increased very slightly to 245MB.  The most marked difference is in the amount of tempdb space used – this query wrote almost 391MB of sort run data to the physical tempdb file.  Don’t draw any general conclusions about VARCHAR(MAX) versus CHAR from this – I chose the length of the data specifically to expose this edge case.  In most cases, VARCHAR(MAX) performs very similarly to CHAR – I just wanted to make test 2 a bit more exciting. MAX Performance Summary: 8 seconds elapsed time 245MB memory grant 391MB tempdb usage 193MB estimated sort set 25,043 logical reads Sort warning Test 3 – TEXT The same test again, but using the deprecated TEXT data type for the padding column: DECLARE @read BIGINT, @write BIGINT ; SELECT @read = SUM(num_of_bytes_read), @write = SUM(num_of_bytes_written) FROM tempdb.sys.database_files AS DBF JOIN sys.dm_io_virtual_file_stats(2, NULL) AS FS ON FS.file_id = DBF.file_id WHERE DBF.type_desc = 'ROWS' ; SET STATISTICS IO ON ; SELECT TOP (150) TT.id, TT.padding FROM dbo.TestTEXT AS TT ORDER BY NEWID() OPTION (MAXDOP 1, RECOMPILE) ; SET STATISTICS IO OFF ; SELECT tempdb_read_MB = (SUM(num_of_bytes_read) - @read) / 1024. / 1024., tempdb_write_MB = (SUM(num_of_bytes_written) - @write) / 1024. / 1024., internal_use_MB = ( SELECT internal_objects_alloc_page_count / 128.0 FROM sys.dm_db_task_space_usage WHERE session_id = @@SPID ) FROM tempdb.sys.database_files AS DBF JOIN sys.dm_io_virtual_file_stats(2, NULL) AS FS ON FS.file_id = DBF.file_id WHERE DBF.type_desc = 'ROWS' ; This time the query runs in 500ms.  If you look at the metrics we have been checking so far, it’s not hard to understand why: TEXT Performance Summary: 0.5 seconds elapsed time 9MB memory grant 5MB tempdb usage 5MB estimated sort set 207 logical reads 596 LOB logical reads Sort warning SQL Server’s memory grant algorithm still underestimates the memory needed to perform the sorting operation, but the size of the data to sort is so much smaller (5MB versus 193MB previously) that the spilled sort doesn’t matter very much.  Why is the data size so much smaller?  The query still produces the correct results – including the large amount of data held in the padding column – so what magic is being performed here? TEXT versus MAX Storage The answer lies in how columns of the TEXT data type are stored.  By default, TEXT data is stored off-row in separate LOB pages – which explains why this is the first query we have seen that records LOB logical reads in its STATISTICS IO output.  You may recall from my last post that LOB data leaves an in-row pointer to the separate storage structure holding the LOB data. SQL Server can see that the full LOB value is not required by the query plan until results are returned, so instead of passing the full LOB value down the plan from the Clustered Index Scan, it passes the small in-row structure instead.  SQL Server estimates that each row coming from the scan will be 79 bytes long – 11 bytes for row overhead, 4 bytes for the integer id column, and 64 bytes for the LOB pointer (in fact the pointer is rather smaller – usually 16 bytes – but the details of that don’t really matter right now). OK, so this query is much more efficient because it is sorting a very much smaller data set – SQL Server delays retrieving the LOB data itself until after the Sort starts producing its 150 rows.  The question that normally arises at this point is: Why doesn’t SQL Server use the same trick when the padding column is defined as VARCHAR(MAX)? The answer is connected with the fact that if the actual size of the VARCHAR(MAX) data is 8000 bytes or less, it is usually stored in-row in exactly the same way as for a VARCHAR(8000) column – MAX data only moves off-row into LOB storage when it exceeds 8000 bytes.  The default behaviour of the TEXT type is to be stored off-row by default, unless the ‘text in row’ table option is set suitably and there is room on the page.  There is an analogous (but opposite) setting to control the storage of MAX data – the ‘large value types out of row’ table option.  By enabling this option for a table, MAX data will be stored off-row (in a LOB structure) instead of in-row.  SQL Server Books Online has good coverage of both options in the topic In Row Data. The MAXOOR Table The essential difference, then, is that MAX defaults to in-row storage, and TEXT defaults to off-row (LOB) storage.  You might be thinking that we could get the same benefits seen for the TEXT data type by storing the VARCHAR(MAX) values off row – so let’s look at that option now.  This script creates a fourth table, with the VARCHAR(MAX) data stored off-row in LOB pages: CREATE TABLE dbo.TestMAXOOR ( id INTEGER IDENTITY (1,1) NOT NULL, padding VARCHAR(MAX) NOT NULL,   CONSTRAINT [PK dbo.TestMAXOOR (id)] PRIMARY KEY CLUSTERED (id), ) ; EXECUTE sys.sp_tableoption @TableNamePattern = N'dbo.TestMAXOOR', @OptionName = 'large value types out of row', @OptionValue = 'true' ; SELECT large_value_types_out_of_row FROM sys.tables WHERE [schema_id] = SCHEMA_ID(N'dbo') AND name = N'TestMAXOOR' ; INSERT INTO dbo.TestMAXOOR WITH (TABLOCKX) ( padding ) SELECT SPACE(0) FROM dbo.TestCHAR ORDER BY id ; UPDATE TM WITH (TABLOCK) SET padding.WRITE (TC.padding, NULL, NULL) FROM dbo.TestMAXOOR AS TM JOIN dbo.TestCHAR AS TC ON TC.id = TM.id ; EXECUTE sys.sp_spaceused @objname = 'dbo.TestMAXOOR' ; CHECKPOINT ; Test 4 – MAXOOR We can now re-run our test on the MAXOOR (MAX out of row) table: DECLARE @read BIGINT, @write BIGINT ; SELECT @read = SUM(num_of_bytes_read), @write = SUM(num_of_bytes_written) FROM tempdb.sys.database_files AS DBF JOIN sys.dm_io_virtual_file_stats(2, NULL) AS FS ON FS.file_id = DBF.file_id WHERE DBF.type_desc = 'ROWS' ; SET STATISTICS IO ON ; SELECT TOP (150) MO.id, MO.padding FROM dbo.TestMAXOOR AS MO ORDER BY NEWID() OPTION (MAXDOP 1, RECOMPILE) ; SET STATISTICS IO OFF ; SELECT tempdb_read_MB = (SUM(num_of_bytes_read) - @read) / 1024. / 1024., tempdb_write_MB = (SUM(num_of_bytes_written) - @write) / 1024. / 1024., internal_use_MB = ( SELECT internal_objects_alloc_page_count / 128.0 FROM sys.dm_db_task_space_usage WHERE session_id = @@SPID ) FROM tempdb.sys.database_files AS DBF JOIN sys.dm_io_virtual_file_stats(2, NULL) AS FS ON FS.file_id = DBF.file_id WHERE DBF.type_desc = 'ROWS' ; TEXT Performance Summary: 0.3 seconds elapsed time 245MB memory grant 0MB tempdb usage 193MB estimated sort set 207 logical reads 446 LOB logical reads No sort warning The query runs very quickly – slightly faster than Test 3, and without spilling the sort to tempdb (there is no sort warning in the trace, and the monitoring query shows zero tempdb usage by this query).  SQL Server is passing the in-row pointer structure down the plan and only looking up the LOB value on the output side of the sort. The Hidden Problem There is still a huge problem with this query though – it requires a 245MB memory grant.  No wonder the sort doesn’t spill to tempdb now – 245MB is about 20 times more memory than this query actually requires to sort 50,000 records containing LOB data pointers.  Notice that the estimated row and data sizes in the plan are the same as in test 2 (where the MAX data was stored in-row). The optimizer assumes that MAX data is stored in-row, regardless of the sp_tableoption setting ‘large value types out of row’.  Why?  Because this option is dynamic – changing it does not immediately force all MAX data in the table in-row or off-row, only when data is added or actually changed.  SQL Server does not keep statistics to show how much MAX or TEXT data is currently in-row, and how much is stored in LOB pages.  This is an annoying limitation, and one which I hope will be addressed in a future version of the product. So why should we worry about this?  Excessive memory grants reduce concurrency and may result in queries waiting on the RESOURCE_SEMAPHORE wait type while they wait for memory they do not need.  245MB is an awful lot of memory, especially on 32-bit versions where memory grants cannot use AWE-mapped memory.  Even on a 64-bit server with plenty of memory, do you really want a single query to consume 0.25GB of memory unnecessarily?  That’s 32,000 8KB pages that might be put to much better use. The Solution The answer is not to use the TEXT data type for the padding column.  That solution happens to have better performance characteristics for this specific query, but it still results in a spilled sort, and it is hard to recommend the use of a data type which is scheduled for removal.  I hope it is clear to you that the fundamental problem here is that SQL Server sorts the whole set arriving at a Sort operator.  Clearly, it is not efficient to sort the whole table in memory just to return 150 rows in a random order. The TEXT example was more efficient because it dramatically reduced the size of the set that needed to be sorted.  We can do the same thing by selecting 150 unique keys from the table at random (sorting by NEWID() for example) and only then retrieving the large padding column values for just the 150 rows we need.  The following script implements that idea for all four tables: SET STATISTICS IO ON ; WITH TestTable AS ( SELECT * FROM dbo.TestCHAR ), TopKeys AS ( SELECT TOP (150) id FROM TestTable ORDER BY NEWID() ) SELECT T1.id, T1.padding FROM TestTable AS T1 WHERE T1.id = ANY (SELECT id FROM TopKeys) OPTION (MAXDOP 1) ; WITH TestTable AS ( SELECT * FROM dbo.TestMAX ), TopKeys AS ( SELECT TOP (150) id FROM TestTable ORDER BY NEWID() ) SELECT T1.id, T1.padding FROM TestTable AS T1 WHERE T1.id IN (SELECT id FROM TopKeys) OPTION (MAXDOP 1) ; WITH TestTable AS ( SELECT * FROM dbo.TestTEXT ), TopKeys AS ( SELECT TOP (150) id FROM TestTable ORDER BY NEWID() ) SELECT T1.id, T1.padding FROM TestTable AS T1 WHERE T1.id IN (SELECT id FROM TopKeys) OPTION (MAXDOP 1) ; WITH TestTable AS ( SELECT * FROM dbo.TestMAXOOR ), TopKeys AS ( SELECT TOP (150) id FROM TestTable ORDER BY NEWID() ) SELECT T1.id, T1.padding FROM TestTable AS T1 WHERE T1.id IN (SELECT id FROM TopKeys) OPTION (MAXDOP 1) ; SET STATISTICS IO OFF ; All four queries now return results in much less than a second, with memory grants between 6 and 12MB, and without spilling to tempdb.  The small remaining inefficiency is in reading the id column values from the clustered primary key index.  As a clustered index, it contains all the in-row data at its leaf.  The CHAR and VARCHAR(MAX) tables store the padding column in-row, so id values are separated by a 3999-character column, plus row overhead.  The TEXT and MAXOOR tables store the padding values off-row, so id values in the clustered index leaf are separated by the much-smaller off-row pointer structure.  This difference is reflected in the number of logical page reads performed by the four queries: Table 'TestCHAR' logical reads 25511 lob logical reads 000 Table 'TestMAX'. logical reads 25511 lob logical reads 000 Table 'TestTEXT' logical reads 00412 lob logical reads 597 Table 'TestMAXOOR' logical reads 00413 lob logical reads 446 We can increase the density of the id values by creating a separate nonclustered index on the id column only.  This is the same key as the clustered index, of course, but the nonclustered index will not include the rest of the in-row column data. CREATE UNIQUE NONCLUSTERED INDEX uq1 ON dbo.TestCHAR (id); CREATE UNIQUE NONCLUSTERED INDEX uq1 ON dbo.TestMAX (id); CREATE UNIQUE NONCLUSTERED INDEX uq1 ON dbo.TestTEXT (id); CREATE UNIQUE NONCLUSTERED INDEX uq1 ON dbo.TestMAXOOR (id); The four queries can now use the very dense nonclustered index to quickly scan the id values, sort them by NEWID(), select the 150 ids we want, and then look up the padding data.  The logical reads with the new indexes in place are: Table 'TestCHAR' logical reads 835 lob logical reads 0 Table 'TestMAX' logical reads 835 lob logical reads 0 Table 'TestTEXT' logical reads 686 lob logical reads 597 Table 'TestMAXOOR' logical reads 686 lob logical reads 448 With the new index, all four queries use the same query plan (click to enlarge): Performance Summary: 0.3 seconds elapsed time 6MB memory grant 0MB tempdb usage 1MB sort set 835 logical reads (CHAR, MAX) 686 logical reads (TEXT, MAXOOR) 597 LOB logical reads (TEXT) 448 LOB logical reads (MAXOOR) No sort warning I’ll leave it as an exercise for the reader to work out why trying to eliminate the Key Lookup by adding the padding column to the new nonclustered indexes would be a daft idea Conclusion This post is not about tuning queries that access columns containing big strings.  It isn’t about the internal differences between TEXT and MAX data types either.  It isn’t even about the cool use of UPDATE .WRITE used in the MAXOOR table load.  No, this post is about something else: Many developers might not have tuned our starting example query at all – 5 seconds isn’t that bad, and the original query plan looks reasonable at first glance.  Perhaps the NEWID() function would have been blamed for ‘just being slow’ – who knows.  5 seconds isn’t awful – unless your users expect sub-second responses – but using 250MB of memory and writing 200MB to tempdb certainly is!  If ten sessions ran that query at the same time in production that’s 2.5GB of memory usage and 2GB hitting tempdb.  Of course, not all queries can be rewritten to avoid large memory grants and sort spills using the key-lookup technique in this post, but that’s not the point either. The point of this post is that a basic understanding of execution plans is not enough.  Tuning for logical reads and adding covering indexes is not enough.  If you want to produce high-quality, scalable TSQL that won’t get you paged as soon as it hits production, you need a deep understanding of execution plans, and as much accurate, deep knowledge about SQL Server as you can lay your hands on.  The advanced database developer has a wide range of tools to use in writing queries that perform well in a range of circumstances. By the way, the examples in this post were written for SQL Server 2008.  They will run on 2005 and demonstrate the same principles, but you won’t get the same figures I did because 2005 had a rather nasty bug in the Top N Sort operator.  Fair warning: if you do decide to run the scripts on a 2005 instance (particularly the parallel query) do it before you head out for lunch… This post is dedicated to the people of Christchurch, New Zealand. © 2011 Paul White email: @[email protected] twitter: @SQL_Kiwi

    Read the article

  • Windows Phone 7 ActiveSync error 86000C09 (My First Post!)

    - by Chris Heacock
    Hello fellow geeks! I'm kicking off this new blog with an issue that was a real nuisance, but was relatively easy to fix. During a recent Exchange 2003 to 2010 migration, one of the users was getting an error on his Windows Phone 7 device. The error code that popped up on the phone on every sync attempt was 86000C09 We tested the following: Different user on the same device: WORKED Problem user on a different device: FAILED   Seemed to point (conclusively) at the user's account as the crux of the issue. This error can come up if a user has too many devices syncing, but he had no other phones. We verified that using the following command: Get-ActiveSyncDeviceStatistics -Identity USERID Turns out, it was the old familiar inheritable permissions issue in Active Directory. :-/ This user was not an admin, nor had he ever been one. HOWEVER, his account was cloned from an ex-admin user, so the unchecked box stayed unchecked. We checked the box and voila, data started flowing to his device(s). Here's a refresher on enabling Inheritable permissions: Open ADUC, and enable Advanced Features: Then open properties and go to the Security tab for the user in question: Click on Advanced, and the following screen should pop up: Verify that "Include inheritable permissions from this object's parent" is *checked*.   You will notice that for certain users, this box keeps getting unchecked. This is normal behavior due to the inbuilt security of Active Directory. People that are in the following groups will have this flag altered by AD: Account Operators Administrators Backup Operators Domain Admins Domain Controllers Enterprise Admins Print Operators Read-Only Domain Controllers Replicator Schema Admins Server Operators Once the box is cheked, permissions will flow and the user will be set correctly. Even if the box is unchecked, they will function normally as they now has the proper permissions configured. You need to perform this same excercise when enabling users for Lync, but that's another blog. :-)   -Chris

    Read the article

  • Consolas Font In Vista And Win7

    - by Sean M
    I have downloaded the Consolas font from Microsoft and installed it on my Windows Vista box. Consolas is also present on my Windows 7 box. When I use PuTTY, being sure to use the same settings on both machines, the Windows 7 box can render Unicode line/box drawing characters in Consolas, but the Windows Vista box cannot. What is the relevant difference between them? If Consolas has the characters, why would they only appear on one system, and not on the other? I am logging into the same remote host each time, and I have been very carefully checking PuTTY's settings to make sure that they're the same on both machines. How can I make Consolas render Unicode line-drawing characters on Vista?

    Read the article

  • SharePoint 2007 Hosting :: How to Move a Document from One Lbrary to Another

    - by mbridge
    Moving a document using a SharePoint Designer workflow involves copying the document to the SharePoint document library you want to move the document to, and then deleting the document from the current document library it is in. You can use the Copy List Item action to copy the document and the Delete item action to delete the document. To create a SharePoint Designer workflow that can move a document from one document library to another: 1. In SharePoint Designer 2007, open the SharePoint site on which the document library that contains the documents to move is located. 2. On the Define your new workflow screen of the Workflow Designer, enter a name for the workflow, select the document library you want to attach the workflow to (this would be a document library containing documents to move), select Allow this workflow to be manually started from an item, and click Next. 3. On the Step 1 screen of the Workflow Designer, click Actions, and then click More Actions from the drop-down menu. 4. On the Workflow Actions dialog box, select List Actions from the category drop-down list box, select Copy List Item from the actions list, and click Add. The following text is added to the Workflow Designer: Copy item in this list to this list 5. On the Step 1 screen of the Workflow Designer, click the first this list (representing the document library to copy the document from) in the text of the Copy List Item action. 6. On the Choose List Item dialog box, leave Current Item selected, and click OK. 7. On the Step 1 screen of the Workflow Designer, click the second this list (representing the document library to copy the document to) in the text of the Copy List Item action, and select the document library (this is the document library to where you want to move the document) from the drop-down list box that appears. 8. On the Step 1 screen of the Workflow Designer, click Actions, and then click More Actions from the drop-down menu. 9. On the Workflow Actions dialog box, select List Actions from the category drop-down list box, select Delete Item from the actions list, and click Add. The following text is added to the Workflow Designer: then Delete item in this list 10. On the Step 1 screen of the Workflow Designer, click this list in the text of the Delete Item action. 11. On the Choose List Item dialog box, leave Current Item selected and click OK. The final text for the workflow should now look like: Copy item in DocLib1 to DocLib2   then Delete item in DocLib1 where DocLib1 is the SharePoint document library containing the document to move and DocLib2 the document library to move the document to. 12. On the Step 1 screen of the Workflow Designer, click Finish. How to Test the Workflow? 1. Go to the SharePoint document library to which you attached the workflow, click on a document, and select Workflows from the drop-down menu. 2. On the Workflows page, click the name of your SharePoint Designer workflow. 3. On the workflow initiation page, click Start.

    Read the article

  • Creating Custom HTML Helpers in ASP.NET MVC

    - by Shravan
    ASP.NET MVC provides many built-in HTML Helpers.  With help of HTML Helpers we can reduce the amount of typing of HTML tags for creating a HTML page. For example we use Html.TextBox() helper method it generates html input textbox. Write the following code snippet in MVC View: <%=Html.TextBox("txtName",20)%> It generates the following html in output page: <input id="txtName" name="txtName" type="text" value="20" /> List of built-in HTML Helpers provided by ASP.NET MVC. ActionLink() - Links to an action method. BeginForm() - Marks the start of a form and links to the action method that renders the form. CheckBox() - Renders a check box. DropDownList() - Renders a drop-down list. Hidden() - Embeds information in the form that is not rendered for the user to see. ListBox() - Renders a list box. Password() - Renders a text box for entering a password. RadioButton() - Renders a radio button.TextArea() - Renders a text area (multi-line text box). TextBox () - Renders a text box. How to develop our own Custom HTML Helpers? For developing custom HTML helpers the simplest way is to write an extension method for the HtmlHelper class. See the below code, it builds a custom Image HTML Helper for generating image tag. Read The Remaing Blog Post @ http://theshravan.net/blog/creating-custom-html-helpers-in-asp-net-mvc/

    Read the article

  • MacGyver Moments

    - by Geoff N. Hiten
    Denny Cherry tagged me to write about my best MacGyver Moment.  Usually I ignore blogosphere fluff and just use this space to write what I think is important.  However, #MVP10 just ended and I have a stronger sense of community.  Besides, where else would I mention my second best Macgyver moment was making a BIOS jumper out of a soda can.  Aluminum is conductive and I didn't have any real jumpers lying around. My best moment is probably my entire home computer network.  Every system but one is hand-built, usually cobbled together out of spare parts and 'adapted' from its original purpose. My Primary Domain Controller is a Dell 2300.   The Service Tag indicates it was shipped to the original owner in 1999.  Box has a PERC/1 RAID controller.  I acquired this from a previous employer for $50.  It runs Windows Server 2003 Enterprise Edition.  Does DNS, DHCP, and RADIUS services as a bonus.  RADIUS authentication is used for VPN and Wireless access.  It is nice to sign in once and be done with it. The Secondary Domain Controller is an old desktop.  Dual P-III 933 with some extra drives. My VPN box is a P-II 250 with 384MB of RAM and a 21 GB hard drive.  I did a P-to-V to my Hyper-V box a year or so ago and retired the hardware again.  Dynamic DNS lets me connect no matter how often Comcast shuffles my IP. The Hyper-V box is a desktop system with 8GB RAM and an AMD Athlon 5000+ processor.  Cost me less than $500 to put together nearly two years ago.  I reasoned that if Vista and Windows 2008 were the same code then Vista 64-bit certified meant the drivers for Vista would load into Windows 2008.  Turns out I was right. Later I added three 1TB drives but wasn't too happy with how that turned out.  I recovered two of the drives yesterday and am building an iSCSI storage unit. (Much thanks to Starwind.  Great product).  I am using an old AMD 1.1GhZ box with 1.5 GB RAM (cobbled together from three old PCs) as my storave server.  The Hyper-V box is slated for an OS rebuild to 2008 R2 once I get the storage system worked out.  maybe in a week or two. A couple of DLink Gigabit switches ties everything together. Add in the Vonage box, the three PCs, the Wireless-N Access Point, the two notebooks and the XBox and you have gone from MacGyver to darn near Rube Goldberg. The only thing I really spend money on is power supplies and fans.  I buy top-of-the-line for both. I even pull and crimp my own cables. Oh, and if my kids hose up a PC, I have all of their data on a server elsewhere.  Every PC and laptop is pretty much interchangable for email and basic workstation tasks.  That helps a lot too. Of course I will tag SQLVariant.

    Read the article

  • How to Assign a Default Signature in Outlook 2013

    - by Lori Kaufman
    If you sign most of your emails the same way, you can easily specify a default signature to automatically insert into new email messages and replies and forwards. This can be done directly in the Signature editor in Outlook 2013. We recently showed you how to create a new signature. You can also create multiple signatures for each email account and define a different default signature for each account. When you change your sending account when composing a new email message, the signature would change automatically as well. NOTE: To have a signature added automatically to new email messages and replies and forwards, you must have a default signature assigned in each email account. If you don’t want a signature in every account, you can create a signature with just a space, a full stop, dashes, or other generic characters. To assign a default signature, open Outlook and click the File tab. Click Options in the menu list on the left side of the Account Information screen. On the Outlook Options dialog box, click Mail in the list of options on the left side of the dialog box. On the Mail screen, click Signatures in the Compose messages section. To change the default signature for an email account, select the account from the E-mail account drop-down list on the top, right side of the dialog box under Choose default signature. Then, select the signature you want to use by default for New messages and for Replies/forwards from the other two drop-down lists. Click OK to accept your changes and close the dialog box. Click OK on the Outlook Options dialog box to close it. You can also access the Signatures and Stationery dialog box from the Message window for new emails and drafts. Click New Email on the Home tab or double-click an email in the Drafts folder to access the Message window. Click Signature in the Include section of the New Mail Message window and select Signatures from the drop-down menu. In the next few days, we will be covering how to use the features of the signature editor next, and then how to insert and change signatures manually, backup and restore your signatures, and modify a signature for use in plain text emails.     

    Read the article

  • XenServer 6.2 crashes everytime I try to install a new Windows 7 x64 vm

    - by Erik
    I'm running XenServer 6.2 with all patches and updates on two boxes. One box is an intel core i7 2600K.. works great. My latest box is an AMD A10 APU.. and everytime I try to install Windows 7 OS as a guest.. I get to the screen where Windows files start to expand, and suddenly the entire box freezes up. Is there a log I can check, and or a way to migrate a working Windows 7 image I created on Box 1 over to Box 2?

    Read the article

  • Ubuntu boots to terminal on start up

    - by Jules
    For a long time I've been unable to get updates due to a "repositories not found" error. Yesterday someone fixed this for me but after installing 94 days worth of updates my system wanted to restart. It looks like it is booting normally but then it opens a terminal and asks for my login and password. I had tried Ctrl+ Alt +F7 and startx to no avail. Here is everything that appears on screen when I turn the computer on. Ubuntu 10.04.4 LTS box-o-doom tty1 box-o-doom login:julian password: last login: Sun Jul 8 10:28:02 BST tty1 Linux box-o-doom 2.6.32-41-generic-pae #91-Ubuntu SMP Wed Jun 13 12:00:09 UTC 20 12 i686 GNU/Linux Ubuntu 10.04.4 LTS Welcome to Ubuntu! *Documentation: http://help.ubuntu.com julian@box-o-doom:~$_ i then tried dmesg which produced hundreds of lines all very similar to the first line reproduced here [ 9.453119] type=1505 audit1341742405.022:10): operation="profile_replace" pid=743 name="/usr/lib/connman/scripts/dhclient-script" follwed by this at the end [ 9.475880] alloc irq_desc for 27 on node-1 [ 9.475883] alloc kstat_irqs on node-1 [ 9.475890]forcedeth 0000:00:07.0: irq27 for MSI/MSI-X [ 9.760031] hda_code:ALC662 rev1: BIOS auto-probing. [ 10.048095] input:HDA Digital PCBeep as /devices/pci 0000:00:05.o/inp ut/input6 [ 10.862278] ppdev: user-space parallel port driver [ 20.268018] eth0: no IPv6 routers present julian@box-o-doom:~$_ results of startx lots of text scrolls off the screen and i have no way of reading it. but everything i can see is reproduced below current version of pixman: 0.16.4 Before reporting problems, check http://wiki.x.org to make sure that you have the latest version Markers: (--) probed, (**) from config file, (==) defult setting, (++) from command line, (!!) notice, (II) informational. (WW) Warning, (EE) error, (NI) not implemented, (??) unknown. (==) log file: "/var/log/Xorg.0.log", Time: SUn Jul 8 12:02:23 2012 (==) using config file: "/etc/X11/xorg.conf" (==)using config directory: "/usr/lib/X11/xorg.conf.d" FATAL: Module nvidia not found. (EE) NVIDIA: Failed to load the NVIDIA kernal module please check your (EE) NVIDIA: systems kernal log for aditional error messages. (EE) Failed to load module "nvidia" (module specific error, 0) (EE) No drivers available. Fatal server error: no screens found please consult the X.org foundation support at http://wiki.x.org for help please also check the log files at "/var/log/X.org.0.log" for aditional informati on ddxSigGiveUp: Closing log giving up xinit: No such file or directory (errno 2): unable to connect to X server xinit: No suck process (errno 3): server error julian@box-o-doom:~$_

    Read the article

  • trying to connect to non-standard port over esxi guest network

    - by user52874
    I've got an exsi 5.5 box that has a redhat 6.5 guest and a win7 guest. The guest nics are connected on a vsphere standard switch. There is no connection from the vswitch to an outside physical nic. I can ping between the two boxes, each way. I can successfully psping redhat:22 from the win7 box. I can successfully tcping win7:139 from the redhat box. All firewalls are down on both boxes. I cannot connect from the win7 box to redhat:8003, either via psping redhat:8003, nor telnet redhat 8003, nor by the application client itself. sudo netstat -patn | grep 8003 on the redhat box shows that it's listening on 0.0.0.0. Any thoughts? suggestions?

    Read the article

  • Getting BootMgr not found errors repeatedly on Win7 x64

    - by abszero
    So here is the basic configuration of the box: Primary RAID 1 (Mirror, Bootable): 2x 300GB WD SATA drives AMD Phenom Quad Core x64 @2.2 ASUS M3N78 Pro Board 4GB RAM Win 7 Ultimate Additionally, this box is a Host OS for several CentOS Boxes via VirtualBox. The box runs like a champ but, for whatever reason, everytime I restart the machine I get a BootMgr not found error when the box tries to boot. I pop in my Win DVD, select 'Repair Windows' then 'Fix Start Up Problems' and everything works fine...once. When I restart the box again I have to go back through this process. Any ideas on what is going on?

    Read the article

  • Use Your Web Browser As A Calculator

    - by Gopinath
    Quite often most of us require Calculator application to evaluate percentage calculations, divisions,etc. Whenever I needed a calculator application I launch Windows Calculator application as it’s built into each and every version of Windows I use.  But the moment I learn that almost all the web browsers have a built in calculator, I stopped using Windows Calculator.  Google Search Box – Every Browser’s Built In Calculator Google Search Box is the built in calculator of every web browser. The search box is capable of evaluation simple expressions like 20/50+10 as well as complex arithmetic formulas that include functions like sin, cos, tan,etc. Almost every web browser has Google Search box by default, if not you can install it very quickly. In Google Chrome browser, Google Search box is built in right inside the address bar. In Firefox & Internet Explorer you can locate it on the top right corner.    To perform calculations, why to launch Calculator when we have a web browser open on our desktop most of the time? Join us on Facebook to read all our stories right inside your Facebook news feed.

    Read the article

  • How to re-do the hard disks in a WD Word Book Edition II ?

    - by jfmessier
    I recently purchased a WD World Book II, a 2 TB one. I call it the "White Box". It has those 2 1TB drives, and they were in this RAID 1 config, only giving me about 1 TB. I could not delete the raid array, and I took the drives in a Linux box. But I also deleted the entire partitions of the disks, and I cannot even et the existing RAID array on this WD White Box. The drives are fine, but I cannot get them to work on the WD White Box. My goal was to get back to a real 2 TB storage space. If I cannot get those drives back in the White Box, I can re-use them elsewhere, but this would mean a waste of the firmware and network connection. After the fact, I read that, anyway, the network performance is rather poor. Thanks :-)

    Read the article

  • 10 gigabit or 1 gigabit switch

    - by Guntis
    We are planning to move mysql to dedicated box. At this moment we have web servers and mysql is running on each. Question is: cheaper is to buy 10G switch and put 10G network card into mysql server. Or buy normal gigabit switch and connect mysql box to switch with multiple network cables. In 1G scenario then we give each web server different mysql IP address. I don't think, that mysql box with one 1G link is enough to to satisfy multiple web box mysql traffic. At this moment we have 3 servers witch are running mysql/web. Plan is to add fourth server for mysql only. Thanks. Edit: if we buy 1G switch with mini-GBIC ports. Can we put in mini-GBIC 10G connectors and then connect mysql box to that port?

    Read the article

  • How to rotate a sprite using multi-touch with AndEngine?

    - by 786
    I am new to Android game development. I am using AndEngine GLES-2. I have created a box with a sprite. This box is now draggable by using the code below. It works fine. But I want multi-touch on this: I want to rotate the sprite with two fingers on that box, and to keep it draggable. I've no idea how do do that, which way should I go? final float centerX = (CAMERA_WIDTH - this.mBox.getWidth()) / 2; final float centerY = (CAMERA_HEIGHT - this.mBox.getHeight()) / 2; Box = new Sprite(centerX, centerY, this.mBox, this.getVertexBufferObjectManager()) { public boolean onAreaTouched(TouchEvent pSceneTouchEvent, float pTouchAreaLocalX, float pTouchAreaLocalY) { this.setPosition(pSceneTouchEvent.getX() - this.getWidth()/ 2, pSceneTouchEvent.getY() - this.getHeight() / 2); float pValueX = pSceneTouchEvent.getX(); float pValueY = CAMERA_HEIGHT-pSceneTouchEvent.getY(); float dx = pValueX - gun.getX(); float dy = pValueY - gun.getY(); double Radius = Math.atan2(dy,dx); double Angle = Radius * 360 ; Box.setRotation((float)Math.toDegrees(Angle)); return true; }

    Read the article

  • Startup a Ubuntu FTP and SSH server without logging in

    - by Jenko
    I have a Ubuntu Server 11.10 box, which I would like to run "headless", and without a keyboard, mouse or display. Is it possible that immediately after startup the machine logs into an account, or allows me to control the machine from my Windows 7 machine nearby via SSH and SFTP? I've got OpenSSH server installed, and even when the box is not logged in I can start an SSH session. I've tried installing VSFTPD but its very troublesome (hardly connects) and only starts when I login locally into the box.

    Read the article

  • Win 7: Share internet connection via Ethernet and WiFi

    - by Anvaka
    I have the following configuration: Box 1. Running Win 7, connected to the internet via Eth0. Has one wireless network adapter and one more ethernet adapter (say, Eth1). Box 2. Running Win XP, has one ethernet adapter. I'd like to share Internet connection of the Box 1 with Box 2 via cable and have box 1 also sharing the Internet with other wireless devices. I don't want to buy any additional hardware. Is it possible? PS: Sorry if I'm unclear. I merely know nothing about NAT and network administration

    Read the article

  • Win 7: Share internet connection via Ethernet and WiFi

    - by Anvaka
    I have the following configuration: Box 1. Running Win 7, connected to the Internet via Eth0. Has one wireless network adapter and one another ethernet adapter (say, Eth1). Box 2. Running Win XP, has one ethernet adapter. I'd like to share Internet connection of the Box 1 with Box 2 via cable and have box 1 also sharing the Internet with other wireless devices. I don't want to buy any additional hardware. Is it possible? PS: Sorry if I'm unclear. I merely know nothing about NAT and network administration.

    Read the article

< Previous Page | 74 75 76 77 78 79 80 81 82 83 84 85  | Next Page >