Search Results

Search found 7188 results on 288 pages for 'frame buffer'.

Page 146/288 | < Previous Page | 142 143 144 145 146 147 148 149 150 151 152 153  | Next Page >

  • Working out of a vertex array for destrucible objects

    - by bobobobo
    I have diamond-shaped polygonal bullets. There are lots of them on the screen. I did not want to create a vertex array for each, so I packed them into a single vertex array and they're all drawn at once. | bullet1.xyz | bullet1.rgb | bullet2.xyz | bullet2.rgb This is great for performance.. there is struct Bullet { vector<Vector3f*> verts ; // pointers into the vertex buffer } ; This works fine, the bullets can move and do collision detection, all while having their data in one place. Except when a bullet "dies" Then you have to clear a slot, and pack all the bullets towards the beginning of the array. Is this a good approach to handling lots of low poly objects? How else would you do it?

    Read the article

  • InputLayout handling

    - by Kikaimaru
    Where are you supposed to store InputLayout? Suppose i have some basic structure like. class Mesh { List<MeshPart> MeshParts } class MeshPart { Effect Effect; VertexBufferBinding VertexBuffer; ... } Where should I store input layout? It's a connection between vertex buffer and specific pass, i can live with just 1 pass but I still have diffferent techniques so i need at least an array with some connection to effecttechniques, but I would appriciate something not crazy like dictionary. I could also create wrapper for Effect and EffectTechnique, but there must be some normal solution.

    Read the article

  • Practicing SEO

    As with any other city on the planet these, SEO companies in Toronto are equally conscious of businesses and companies serve two markets, and we are not talking about demographics. We mean the walk in customers and also virtual customers. Online retail income is not a line of revenue any business owner can afford to dismiss ever again. With a lot of the world's human population connected to the internet, people eat, sleep, listen to music, watch TV, buffer sitcoms and movies, chat with their friends, Google and Wikipedia any and almost everything beneath the sun they randomly encounter.

    Read the article

  • Creating a DrawableGameComponent

    - by Christian Frantz
    If I'm going to draw cubes effectively, I need to get rid of the numerous amounts of draw calls I have and what has been suggested is that I create a "mesh" of my cubes. I already have them being stored in a single vertex buffer, but the issue lies in my draw method where I am still looping through every cube in order to draw them. I thought this was necessary as each cube will have a set position, but it lowers the frame rate incredibly. What's the easiest way to go about this? I have a class CubeChunk that inherits Microsoft.Stuff.DrawableGameComponent, but I don't know what comes next. I suppose I could just use the chunk of cubes created in my cube class, but that would just keep me going in circles and drawing each cube individually. The goal here is to create a draw method that draws my chunk as a whole, and to not draw individual cubes as I've been doing.

    Read the article

  • Collaborative Whiteboard using WebSocket in GlassFish 4 - Text/JSON and Binary/ArrayBuffer Data Transfer (TOTD #189)

    - by arungupta
    This blog has published a few blogs on using JSR 356 Reference Implementation (Tyrus) as its integrated in GlassFish 4 promoted builds. TOTD #183: Getting Started with WebSocket in GlassFish TOTD #184: Logging WebSocket Frames using Chrome Developer Tools, Net-internals and Wireshark TOTD #185: Processing Text and Binary (Blob, ArrayBuffer, ArrayBufferView) Payload in WebSocket TOTD #186: Custom Text and Binary Payloads using WebSocket One of the typical usecase for WebSocket is online collaborative games. This Tip Of The Day (TOTD) explains a sample that can be used to build such games easily. The application is a collaborative whiteboard where different shapes can be drawn in multiple colors. The shapes drawn on one browser are automatically drawn on all other peer browsers that are connected to the same endpoint. The shape, color, and coordinates of the image are transfered using a JSON structure. A browser may opt-out of sharing the figures. Alternatively any browser can send a snapshot of their existing whiteboard to all other browsers. Take a look at this video to understand how the application work and the underlying code. The complete sample code can be downloaded here. The code behind the application is also explained below. The web page (index.jsp) has a HTML5 Canvas as shown: <canvas id="myCanvas" width="150" height="150" style="border:1px solid #000000;"></canvas> And some radio buttons to choose the color and shape. By default, the shape, color, and coordinates of any figure drawn on the canvas are put in a JSON structure and sent as a message to the WebSocket endpoint. The JSON structure looks like: { "shape": "square", "color": "#FF0000", "coords": { "x": 31.59999942779541, "y": 49.91999053955078 }} The endpoint definition looks like: @WebSocketEndpoint(value = "websocket",encoders = {FigureDecoderEncoder.class},decoders = {FigureDecoderEncoder.class})public class Whiteboard { As you can see, the endpoint has decoder and encoder registered that decodes JSON to a Figure (a POJO class) and vice versa respectively. The decode method looks like: public Figure decode(String string) throws DecodeException { try { JSONObject jsonObject = new JSONObject(string); return new Figure(jsonObject); } catch (JSONException ex) { throw new DecodeException("Error parsing JSON", ex.getMessage(), ex.fillInStackTrace()); }} And the encode method looks like: public String encode(Figure figure) throws EncodeException { return figure.getJson().toString();} FigureDecoderEncoder implements both decoder and encoder functionality but thats purely for convenience. But the recommended design pattern is to keep them in separate classes. In certain cases, you may even need only one of them. On the client-side, the Canvas is initialized as: var canvas = document.getElementById("myCanvas");var context = canvas.getContext("2d");canvas.addEventListener("click", defineImage, false); The defineImage method constructs the JSON structure as shown above and sends it to the endpoint using websocket.send(). An instant snapshot of the canvas is sent using binary transfer with WebSocket. The WebSocket is initialized as: var wsUri = "ws://localhost:8080/whiteboard/websocket";var websocket = new WebSocket(wsUri);websocket.binaryType = "arraybuffer"; The important part is to set the binaryType property of WebSocket to arraybuffer. This ensures that any binary transfers using WebSocket are done using ArrayBuffer as the default type seem to be blob. The actual binary data transfer is done using the following: var image = context.getImageData(0, 0, canvas.width, canvas.height);var buffer = new ArrayBuffer(image.data.length);var bytes = new Uint8Array(buffer);for (var i=0; i<bytes.length; i++) { bytes[i] = image.data[i];}websocket.send(bytes); This comprehensive sample shows the following features of JSR 356 API: Annotation-driven endpoints Send/receive text and binary payload in WebSocket Encoders/decoders for custom text payload In addition, it also shows how images can be captured and drawn using HTML5 Canvas in a JSP. How could this be turned in to an online game ? Imagine drawing a Tic-tac-toe board on the canvas with two players playing and others watching. Then you can build access rights and controls within the application itself. Instead of sending a snapshot of the canvas on demand, a new peer joining the game could be automatically transferred the current state as well. Do you want to build this game ? I built a similar game a few years ago. Do somebody want to rewrite the game using WebSocket APIs ? :-) Many thanks to Jitu and Akshay for helping through the WebSocket internals! Here are some references for you: JSR 356: Java API for WebSocket - Specification (Early Draft) and Implementation (already integrated in GlassFish 4 promoted builds) Subsequent blogs will discuss the following topics (not necessary in that order) ... Error handling Interface-driven WebSocket endpoint Java client API Client and Server configuration Security Subprotocols Extensions Other topics from the API

    Read the article

  • What is a right datatype in C++ for OpenGL scene representation with use of GLSL

    - by Rarach
    I am programming in C++ OpenGl with GLSL. Until now I have been using a data structure that is composed of std::vector filled with structures of vertexes and with their parameters (position , normal, color ...) as a global variable for all the code. My question is, as I am using VBOs for drawing - is this a good approach to this problem? I am asking because I happen to have a lot of memory related trouble with this structure. I am trying to generate a terrain with a lot of vertices - more than 1 million. This seems to work, but as I refill the buffer I get memory related issues (crushes that occur, more or less randomly). So again the question is, is this a good data structure to use / and look for the faults in my code, or should I change to something else? Or what data structure would be advisable?

    Read the article

  • SQLServerCentral Webinar Series #21 - Forgotten Rings & Other Monitoring Stories

    Most common monitoring metrics are important and useful, especially over time, but they can fall short. How do you gather information to determine, for example, if you have buffer cache pressure? Register now for the free webinar. Wednesday, October 17 2012 4:00pm - 5:00pm BST Are you sure you can restore your backups? Run full restore + DBCC CHECKDB quickly and easily with SQL Backup Pro's new automated verification. Check for corruption and prepare for when disaster strikes. Try it now.

    Read the article

  • How to get the Dash and HUD to appear. (and stop Unity spewing error messages.)

    - by Ubuntiac
    I just installed Ubuntu 12.04 on my wifes Dell Inspiron 1501, which uses an R300 ATI graphics chip. Neither the Dash or HUD appear when pushing the appropriate key. When I try unity --reset & in the terminal, I see that over and over it's spitting out: r300: CS space validation failed. (not enough memory?) Skipping rendering. This is just after starting Ubuntu with no apps open, so I find it hard to believe that just rendering the Dash / HUD is completely blowing out the VRAM. Any suggestions on getting this working? /usr/lib/nux/unity_support_test -p shows OpenGL vendor string: X.Org R300 Project OpenGL renderer string: Gallium 0.4 on ATI RS480 OpenGL version string: 2.1 Mesa 8.0.2 Not software rendered: yes Not blacklisted: yes GLX fbconfig: yes GLX texture from pixmap: yes GL npot or rect textures: yes GL vertex program: yes GL fragment program: yes GL vertex buffer object: yes GL framebuffer object: yes GL version is 1.4+: yes Unity 3D supported: yes All sections say "YES"

    Read the article

  • Why fgetc too slow?

    - by user14284
    I've written a program that reads an whole file via fgetc: while ((c = fgetc(f)) != EOF) { ... } But the program is too slow. When I changed fgetc to fread, static unsigned char buf[4096]; while ((n = fread(buf, 1, sizeof(buf), f)) > 0) { ... } the program works about 10 times faster. Why? As I know, fgetc is a buffered function, so it should work as fast as the second version with explicit buffer, isn't it?

    Read the article

  • Slow writing HDD speed, Ubuntu 12.04 64-bit, Thinkpad T520i

    - by pyc
    It seems that (but I'm not completely sure), that when I'm copying files from gigabit network to HDD, I can't use full potential of the network which in my case is about 60 MB/s, because HDD writing is so slow like lower than 10 MB/s, and also it's slowing down the whole system which becomes pretty much unresponsive, almost impossible to work with. Copying files to samba share residing at Ubuntu machine, connected to share from Windows 7, I'm completely sure my network equipment is OK, and there's no CPU intensive process on Ubuntu except smbd getting about 10-20% from time to time which I think is OK. Something here is burried deep I think, maybe even in kernel. Already tried to switch from AHCI to compatibility mode, and turning acpi on and off - nothing helped. So it's like HDD buffer is full and emptying slowly while machine is sluggish, load is about 3 to 4. Somebody experienced the similar problems? Some help on troubleshooting process and identifying the cause would be helpful too :) Thanks!

    Read the article

  • Make public webcam. Which protocol, which codec. (Using VLC)

    - by gsedej
    Hi! I want to use my old (1GHz) PC as webcam video stream server (like you can see those road cameras). I thought of using VLC and already tried using http output but it was not really good. Too cpu hungry, too big stream (kBps), not stable... I been reading VLC how-to's but thre is still a question. Which output should I use? Http, RTSP, UDP? I want to make for more than one computer at the same time (multicast). Which codec should be good? PC is not so fast so it shouldn't be too cpu hungry codec. Mpeg2, mpeg4, xvid? how much video buffer should I use (vb=?)? What about setting IP and ports? So I need some help with ideas, but if someone can make a VLC command line it's even better :) Oh, computer has direct internet connection and own IP.

    Read the article

  • XNA VertexBuffer.SetData performance suggestions

    - by CodeSpeaker
    I have a 3d world in a grid layout where each grid cell contains its separate vertex and index buffer for the mesh/terrain of that cell. When the player moves outside the boundaries of his cell, i dynamically load more cells in his walking direction based on his viewing distance. This triggers x number of vertex and indexbuffer initializations depending on how many cells that needs to be generated and causes the framerate to drop annoyingly during this time. The generation of terrain data is handled in a separate thread and runs smoothly. The vertex and index buffers are added during the update cycle of the game loop. I´ve tried batching the number of cells to be processed to avoid sending too much data at once into the buffers, which worked ok at a shorter viewing distance (about 9 cells to process), but not as well at greater distances with around 30 cells to process. Any idea how i can optimize this?

    Read the article

  • What file formats and conventions should I support to make my game engine artist-friendly?

    - by Avi
    I'm writing a game engine, and I want to know what I should do to make it more artist-friendly. I don't want to be too limiting in terms of what file formats I support, etc. Some specific questions: Are there specific formats artists like to model in? Does it not matter because the 3D modeler abstracts the data storage away? Is it okay if I don't support per-vertex coloration in my game engine? If I have to store a diffuse, specular, ambient, and emissive color value for each vertex, it doubles the size of vertices in the buffer. Is it reasonable to ask artists to do all these things in textures / maps? Any other tips you have about making it so that artists have to adapt their style to my specific engine as little as possible would be nice.

    Read the article

  • Is it possible to save output from a command to a file subsequent the command already has been executed?

    - by NES
    Does an elegant way exists to save the output of a command subsequent to a file as long as the terminal window is open? I mean once the command has been executed in the terminal. The output is still there in terminal. Now i could copy & paste all the lines and save it to a file. But perhaps does a method exist to somehow write the output buffer of a terminal window to a file or even better the output of an already executed command?

    Read the article

  • Looking for parallel programming problem

    - by Chris Lieb
    I am trying to come up with a problem that is easily solvable in a parallel manner and that requires communication between threads for a test. I also am trying to avoid problems that require require random waits, which rules out dining philosophers and producer-consumer (bounded buffer), two of the classics. My goal is for the student to be able to write the program in less than 20-30 minutes in front of a computer not knowing of the problem beforehand. (This is to prevent preparation more than to come up with something novel.) I am trying to stress the communication aspect of the program, though the multi-threaded nature is also important. Does anyone have some ideas? Edit: I'm using Google Go for the language and testing comprehension of the goroutines/channels combo vs an actors library that I authored.

    Read the article

  • Depth Map resolution shifting

    - by user3669538
    the problem is with shadow mapping as you can see, actually it works fine but in a certain condition that the Depth Map size must be equal to the size of rendering buffer, I use an infinite directional light so if the window is 800x600 the depth map must be 800x600, and when i change the size of the shadow map to be 900x600 it starts to be shifted and when it's size be 1024x1024 it also shifts till it disappears the GLSL shadow function float calcShadow(sampler2D Dmap, vec4 coor){ vec4 sh = vec4((coor.xyz/coor.w),1); sh.z *= 0.9; return step(sh.z,texture2D(Dmap,sh.xy).r); } here's the result when it's the same size as the window Colored result & Depth Map and here's the shifted result, as you can notice the depth map is exactly as the previous one with the addition of white space to the right. Colored result http://goo.gl/5lYIFV Depth Map http://goo.gl/7320Dd

    Read the article

  • How to update a mesh position base on a pressed key?

    - by steven166
    I have a mesh loaded from a file, like a tiger mesh. At the first time it locates at A position, then if I press a left key, it will moves to B position but the problem is if I press a left key one more time, it will move from B position to C position. It means that the amount I want to move the mesh will base on the current position instead of the first time rendering position. I can do it if I have a array vertices then I just update the vertex buffer, but a mesh loaded from a file does not have an array vertices, so how to do it? Anybody help me, please?

    Read the article

  • How do I connect the seams between my terrain?

    - by gnomgrol
    I'm using c++ and D3D11 and I'm trying to create a (pretty) large terrain, lets say 4096x4096, maybe larger. I've got the basics of terrain creation and already split it up into chunks. But, when I'm rendering them (every chunk has its own vertex and index buffer, as well as its own heightmap), there are still little pieces missing between them. I read a lot about LOD(Level Of Detail) and GMM(Geometry Mipmap), but I can't really implement the theory I read. At the moment, it looks like this: I could really use some help, everything is welcome. If you have some good tutorials on any of this, please share them.

    Read the article

  • Is there a single book that covers the breadth of computer science fundamentals? [closed]

    - by superFoo
    When I did my undergraduate studies in elecrical engineering, there was this book called "Basic Electricity" by Van Valkenburgh. If you read that book cover to cover, your fundamentals in electrical engineering would be bulletproof. I would recommend it all my juniors and I absolutely loved it. Is there such a book in the field of computer science? I am not so concerned about the algorithms. I am looking more into something that tells me how does everything work beneath the covers. TCPIP, memory management, DNS, routing, SSL, buffer, queuing etc.

    Read the article

  • SLOB: ?????????????

    - by katsumii
    Oracle DB????????????????????????????Introducing SLOB – The Silly Little Oracle Benchmark « Kevin Closson's Blog: Platforms, Databases and StorageSLOB supports testing Oracle logical read (SGA buffer gets) scalingSLOB supports testing physical random single-block reads (db file sequential read)SLOB supports testing random single block writes (DBWR flushing capacity)SLOB supports testing extreme REDO logging I/O????????????????Oracle?????????Swingbench ??????????IPC Semaphore?????C???????????????????Windows???????????Cygwin??????????????????????????????SwingbenchSwingbench can be used to demonstrate and test technologies such as Real Application Clusters, Online table rebuilds, Standby databases, Online backup and recovery etc.???????I/O?????????????????Oracle ORION DownloadsORION (Oracle I/O Calibration Tool) is a standalone tool for calibrating the I/O performance for storage systemsSLOB ??????????????????????????? 

    Read the article

  • adding UIImageView to UIScrollView throws exception cocoa touch for iPad

    - by Brodie4598
    I am a noob at OBJ-C :) I am trying to add a UIImageView to a UIScrollView to display a large image in my iPad app. I have followed the tutorial here exactly: http://howtomakeiphoneapps.com/2009/12/how-to-use-uiscrollview-in-your-iphone-app/ The only difference is that in my App the View is in a seperate tab and I am using a different image. here is my code: - (void)viewDidLoad { [super viewDidLoad]; UIImageView *tempImageView = [[UIImageView alloc] initWithImage:[UIImage imageNamed:@"Cheyenne81"]]; self.imageView = tempImageView; [tempImageView release]; scrollView.contentSize = CGSizeMake(imageView.frame.size.width, imageView.frame.size.height); scrollView.maximumZoomScale = 4.0; scrollView.minimumZoomScale = 0.75; scrollView.clipsToBounds = YES; scrollView.delegate = self; [scrollView addSubview:imageView]; } - (UIView *)viewForZoomingInScrollView:(UIScrollView *)scrollView{ return imageView; } and: @interface UseScrollViewViewController : UIViewController<UIScrollViewDelegate>{ IBOutlet UIScrollView *scrollView; UIImageView *imageView; } @property (nonatomic, retain) UIScrollView *scrollView; @property (nonatomic, retain) UIImageView *imageView; @end I then create a UIScrollView in Interface Builder and link it to the scrollView outlet. Thats when I get the problem. When I run the program it crashes instantly. If I run it without linking the scrollView to the outlet, it will run (allbeit with a blnk screen). The following is the error I get in the console: 2010-03-27 20:18:13.467 UseScrollViewViewController[7421:207] * Terminating app due to uncaught exception 'NSUnknownKeyException', reason: '[ setValue:forUndefinedKey:]: this class is not key value coding-compliant for the key scrollView.'

    Read the article

  • iPhone: image in table cell - wrong position & not scaled correctly in iPhone 3.0

    - by Karsten Silz
    Hi, I have a custom table cell in an iPhone 3.x app to show basketball player data, including a picture. Here's my code from the table controller: UIImageView *theImage = (UIImageView *) [cell viewWithTag:0]; theImage.image = [manager getSmallImageForPlayer: data.number]; theImage.frame = CGRectMake(14, 0, 30, 44); theImage.clipsToBounds = YES; theImage.contentMode = UIViewContentModeScaleAspectFit; theImage.autoresizingMask = UIViewAutoresizingFlexibleHeight && UIViewAutoresizingFlexibleWidth; theImage.autoresizesSubviews = YES; Here's what it looks like in the iPhone SDK 3.0 simulator: Here's what it looks like in the iPhone SDK 3.1.3 simulator: So I have two problems: In both versions, the image view starts at position (0,0) in the table cell, even though I set the frame to start at (14,0). In the 3.0 simulator (and 3.0 device), the image is not shrunk down correctly, even though I set the image view content mode to "UIViewContentModeScaleAspectFit". But the same code works fine in the 3.1.3 simulator / device. Who can help me with either of these two problems?

    Read the article

  • Mixing Objective-C and C++

    - by helixed
    Hello, I'm trying to mix together some Objective-C code with C++. I've always heard it was possible, but I've never actually tried it before. When I try to compile the code, I get a bunch of errors. Here's a simple example I've created which illustrates my problems: AView.h #import <Cocoa/Cocoa.h> #include "B.h" @interface AView : NSView { B *b; } -(void) setB: (B *) theB; @end AView.m #import "AView.h" @implementation AView - (id)initWithFrame:(NSRect)frame { self = [super initWithFrame:frame]; if (self) { // Initialization code here. } return self; } - (void)drawRect:(NSRect)dirtyRect { // Drawing code here. } -(void) setB: (B *) theB { b = theB; } @end B.h #include <iostream> class B { B() { std::cout << "Hello from C++"; } }; Here's the list of errors I get when I try to compile this: /Users/helixed/Desktop/Example/B.h:1:0 /Users/helixed/Desktop/Example/B.h:1:20: error: iostream: No such file or directory /Users/helixed/Desktop/Example/B.h:3:0 /Users/helixed/Desktop/Example/B.h:3: error: expected '=', ',', ';', 'asm' or '__attribute__' before 'B' /Users/helixed/Desktop/Example/AView.h:5:0 /Users/helixed/Desktop/Example/AView.h:5: error: expected specifier-qualifier-list before 'B' /Users/helixed/Desktop/Example/AView.h:8:0 /Users/helixed/Desktop/Example/AView.h:8: error: expected ')' before 'B' /Users/helixed/Desktop/Example/AView.m:26:0 /Users/helixed/Desktop/Example/AView.m:26: error: expected ')' before 'B' /Users/helixed/Desktop/Example/AView.m:27:0 /Users/helixed/Desktop/Example/AView.m:27: error: 'b' undeclared (first use in this function) All I'm doing to compile this right now is using the default compiler built into Xcode. I didn't edit the Cocoa Application template in any way other than adding the two files I created and chaing the NSView to AView in the xib file. Could somebody please tell me what I'm doing wrong? Thanks, helixed

    Read the article

  • Mixing Objective-C and C++

    - by helixed
    I'm trying to mix together some Objective-C code with C++. I've always heard it was possible, but I've never actually tried it before. When I try to compile the code, I get a bunch of errors. Here's a simple example I've created which illustrates my problems: AView.h #import <Cocoa/Cocoa.h> #include "B.h" @interface AView : NSView { B *b; } -(void) setB: (B *) theB; @end AView.m #import "AView.h" @implementation AView - (id)initWithFrame:(NSRect)frame { self = [super initWithFrame:frame]; if (self) { // Initialization code here. } return self; } - (void)drawRect:(NSRect)dirtyRect { // Drawing code here. } -(void) setB: (B *) theB { b = theB; } @end B.h #include <iostream> class B { B() { std::cout << "Hello from C++"; } }; Here's the list of errors I get when I try to compile this: /Users/helixed/Desktop/Example/B.h:1:0 /Users/helixed/Desktop/Example/B.h:1:20: error: iostream: No such file or directory /Users/helixed/Desktop/Example/B.h:3:0 /Users/helixed/Desktop/Example/B.h:3: error: expected '=', ',', ';', 'asm' or '__attribute__' before 'B' /Users/helixed/Desktop/Example/AView.h:5:0 /Users/helixed/Desktop/Example/AView.h:5: error: expected specifier-qualifier-list before 'B' /Users/helixed/Desktop/Example/AView.h:8:0 /Users/helixed/Desktop/Example/AView.h:8: error: expected ')' before 'B' /Users/helixed/Desktop/Example/AView.m:26:0 /Users/helixed/Desktop/Example/AView.m:26: error: expected ')' before 'B' /Users/helixed/Desktop/Example/AView.m:27:0 /Users/helixed/Desktop/Example/AView.m:27: error: 'b' undeclared (first use in this function) All I'm doing to compile this right now is using the default compiler built into Xcode. I didn't edit the Cocoa Application template in any way other than adding the two files I created and chaing the NSView to AView in the xib file. Could somebody please tell me what I'm doing wrong? Thanks, helixed

    Read the article

  • Delphi label and asm weirdness?

    - by egon
    I written an asm function in Delphi 7 but it transforms my code to something else: function f(x: Cardinal): Cardinal; register; label err; asm not eax mov edx,eax shr edx, 1 and eax, edx bsf ecx, eax jz err mov eax, 1 shl eax, cl mov edx, eax add edx, edx or eax, edx ret err: xor eax, eax end; // compiled version f: push ebx // !!! not eax mov edx,eax shr edx, 1 and eax, edx bsf ecx, eax jz +$0e mov eax, 1 shl eax, cl mov edx, eax add edx, edx or eax, edx ret err: xor eax, eax mov eax, ebx // !!! pop ebx // !!! ret // the almost equivalent without asm function f(x: Cardinal): Cardinal; var c: Cardinal; begin x := not x; x := x and x shr 1; if x <> 0 then begin c := bsf(x); // bitscanforward x := 1 shl c; Result := x or (x shl 1) end else Result := 0; end; Why does it generate push ebx and pop ebx? And why does it do mov eax, ebx? It seems that it generates the partial stack frame because of the mov eax, ebx. This simple test generates mov eax, edx but doesn't generate that stack frame: function asmtest(x: Cardinal): Cardinal; register; label err; asm not eax and eax, 1 jz err ret err: xor eax, eax end; // compiled asmtest: not eax and eax, $01 jz +$01 ret xor eax, eax mov eax, edx // !!! ret It seems that it has something to do with the label err. If I remove that I don't get the mov eax, * part. Why does this happen? Made a bug report on Quality Central.

    Read the article

< Previous Page | 142 143 144 145 146 147 148 149 150 151 152 153  | Next Page >