Search Results

Search found 12445 results on 498 pages for 'memory fragmentation'.

Page 250/498 | < Previous Page | 246 247 248 249 250 251 252 253 254 255 256 257  | Next Page >

• Visual Studio crashes consistently on web-related projects

  - by Traveling Tech Guy

  Hi, I have a brand new VS2010 installed on a Win2008R2 machine. I started getting this error when debugging a WCF service project: "Attempted to read or write protected memory. This is often an indication that other memory is corrupt." When I started developing a web site a week later, this became consistent - I can't debug it. The stack dump reads: at Microsoft.VisualStudio.WebHost.Host.ProcessRequest(Connection conn) at Microsoft.VisualStudio.WebHost.Server.OnSocketAccept(Object acceptedSocket) at System.Threading.QueueUserWorkItemCallback.WaitCallback_Context(Object state) at System.Threading.ExecutionContext.Run(ExecutionContext executionContext, ContextCallback callback, Object state, Boolean ignoreSyncCtx) at System.Threading.QueueUserWorkItemCallback.System.Threading.IThreadPoolWorkItem.ExecuteWorkItem() at System.Threading.ThreadPoolWorkQueue.Dispatch() at System.Threading._ThreadPoolWaitCallback.PerformWaitCallback() I tried searching online, and some recommend turning off the "Suppress JIT Optimizations" in the Debugging options - this dos not seem to make a difference. Clearly the problem is with the built in web server. But am I doing something wrong? Is there something I can do? Or is this a known bug? Thanks for your time, Guy Update 12/31: Today I tried using CassiniDev as a replacement to the original VS2010 WebServer - exact same result. My suspicion is that there's some internal conflict between VS2010, Windows Server 2008R2 and maybe the fact that it's a 64 bit OS. I switched to using IIS as my debug server - and that seems to work, with some annoying side effects. My conclusion: do not use a 64 bit server system as your dev machine. Develop on 32bit - deploy to 64bit. Side conclusion: there are some scenarios Microsoft's QA doesn't test.

  Read the article

• Alignment in assembly

  - by jena

  Hi, I'm spending some time on assembly programming (Gas, in particular) and recently I learned about the align directive. I think I've understood the very basics, but I would like to gain a deeper understanding of its nature and when to use alignment. For instance, I wondered about the assembly code of a simple C++ switch statement. I know that under certain circumstances switch statements are based on jump tables, as in the following few lines of code: .section .rodata .align 4 .align 4 .L8: .long .L2 .long .L3 .long .L4 .long .L5 ... .align 4 aligns the following data on the next 4-byte boundary which ensures that fetching these memory locations is efficient, right? I think this is done because there might be things happening before the switch statement which caused misalignment. But why are there actually two calls to .align? Are there any rules of thumb when to call .align or should it simply be done whenever a new block of data is stored in memory and something prior to this could have caused misalignment? In case of arrays, it seems that alignment is done on 32-byte boundaries as soon as the array occupies at least 32 byte. Is it more efficient to do it this way or is there another reason for the 32-byte boundary? I'd appreciate any explanation or hint on literature.

  Read the article

• What happens when value types are created?

  - by Bob

  I'm developing a game using XNA and C# and was attempting to avoid calling new struct() type code each frame as I thought it would freak the GC out. "But wait," I said to myself, "struct is a value type. The GC shouldn't get called then, right?" Well, that's why I'm asking here. I only have a very vague idea of what happens to value types. If I create a new struct within a function call, is the struct being created on the stack? Will it simply get pushed and popped and performance not take a hit? Further, would there be some memory limit or performance implications if, say, I need to create many instances in a single call? Take, for instance, this code: spriteBatch.Draw(tex, new Rectangle(x, y, width, height), Color.White); Rectangle in this case is a struct. What happens when that new Rectangle is created? What are the implications of having to repeat that line many times (say, thousands of times)? Is this Rectangle created, a copy sent to the Draw method, and then discarded (meaning no memory getting eaten up the more Draw is called in that manner in the same function)? P.S. I know this may be pre-mature optimization, but I'm mostly curious and wish to have a better understanding of what is happening.

  Read the article

• How to create an instance of object with RTTI in Delphi 2010?

  - by Paul

  As we all known, when we call a constructor of a class like this: instance := TSomeClass.Create; The Delphi compiler actually do the following things: Call the static NewInstance method to allocate memory and initialize the memory layout. Call the constructor method to perform the initialization of the class Call the AfterConstruction method It's simple and easy to understand. but I'm not very sure how the compiler handle exceptions in the second and the third step. It seems there are no explicit way to create an instance using a RTTI constructor method in D2010. so I wrote a simple function in the Spring Framework for Delphi to reproduce the process of the creation. class function TActivator.CreateInstance(instanceType: TRttiInstanceType; constructorMethod: TRttiMethod; const arguments: array of TValue): TObject; var classType: TClass; begin TArgument.CheckNotNull(instanceType, 'instanceType'); TArgument.CheckNotNull(constructorMethod, 'constructorMethod'); classType := instanceType.MetaclassType; Result := classType.NewInstance; try constructorMethod.Invoke(Result, arguments); except on Exception do begin if Result is TInterfacedObject then begin Dec(TInterfacedObjectHack(Result).FRefCount); end; Result.Free; raise; end; end; try Result.AfterConstruction; except on Exception do begin Result.Free; raise; end; end; end; I feel it maybe not 100% right. so please show me the way. Thanks!

  Read the article

• Atomic swap in GNU C++

  - by Steve

  I want to verify that my understanding is correct. This kind of thing is tricky so I'm almost sure I am missing something. I have a program consisting of a real-time thread and a non-real-time thread. I want the non-RT thread to be able to swap a pointer to memory that is used by the RT thread. From the docs, my understanding is that this can be accomplished in g++ with: // global Data *rt_data; Data *swap_data(Data *new_data) { #ifdef __GNUC__ // Atomic pointer swap. Data *old_d = __sync_lock_test_and_set(&rt_data, new_data); #else // Non-atomic, cross your fingers. Data *old_d = rt_data; rt_data = new_data; #endif return old_d; } This is the only place in the program (other than initial setup) where rt_data is modified. When rt_data is used in the real-time context, it is copied to a local pointer. For old_d, later on when it is sure that the old memory is not used, it will be freed in the non-RT thread. Is this correct? Do I need volatile anywhere? Are there other synchronization primitives I should be calling? By the way I am doing this in C++, although I'm interested in whether the answer differs for C. Thanks ahead of time.

  Read the article

• vectorizing loops in Matlab - performance issues

  - by Gacek

  This question is related to these two: http://stackoverflow.com/questions/2867901/introduction-to-vectorizing-in-matlab-any-good-tutorials http://stackoverflow.com/questions/2561617/filter-that-uses-elements-from-two-arrays-at-the-same-time Basing on the tutorials I read, I was trying to vectorize some procedure that takes really a lot of time. I've rewritten this: function B = bfltGray(A,w,sigma_r) dim = size(A); B = zeros(dim); for i = 1:dim(1) for j = 1:dim(2) % Extract local region. iMin = max(i-w,1); iMax = min(i+w,dim(1)); jMin = max(j-w,1); jMax = min(j+w,dim(2)); I = A(iMin:iMax,jMin:jMax); % Compute Gaussian intensity weights. F = exp(-0.5*(abs(I-A(i,j))/sigma_r).^2); B(i,j) = sum(F(:).*I(:))/sum(F(:)); end end into this: function B = rngVect(A, w, sigma) W = 2*w+1; I = padarray(A, [w,w],'symmetric'); I = im2col(I, [W,W]); H = exp(-0.5*(abs(I-repmat(A(:)', size(I,1),1))/sigma).^2); B = reshape(sum(H.*I,1)./sum(H,1), size(A, 1), []); But this version seems to be as slow as the first one, but in addition it uses a lot of memory and sometimes causes memory problems. I suppose I've made something wrong. Probably some logic mistake regarding vectorizing. Well, in fact I'm not surprised - this method creates really big matrices and probably the computations are proportionally longer. I have also tried to write it using nlfilter (similar to the second solution given by Jonas) but it seems to be hard since I use Matlab 6.5 (R13) (there are no sophisticated function handles available). So once again, I'm asking not for ready solution, but for some ideas that would help me to solve this in reasonable time. Maybe you will point me what I did wrong.

  Read the article

• App hosting Report Viewer crashes on exit after export

  - by Paul Sasik

  We have a .NET Winforms application that hosts the Crystal Reports Viewer control (Version XI). It works well for the most part but when an export of data from the viewer is performed the application will crash on exit and in unmanaged code. The error message is not very useful and just says that an incorrect memory location was accessed. No other info such a specific DLL etc. is provided. This only happens after the viewer is used to export a report to CSV, XML etc. My guess is that at some point in the export process Crystal creates a resource that attempts an action on shut down to a parent window (perhaps) that no longer exists. I've seen a number of memory leak and shut down issues with Crystal but this one's new. Has anyone seen it and come up with a workaround or has ideas for workarounds? So far we've tried explicitly disposing of all crystal-related objects, setting to null and even setting a Thread.Sleep cycle on shut down to "give Crystal time to clean up." Update: The crash happens only on shut down (so not immediate) All export formats work All export files are created properly CR is installed on the same machine as the hosting .NET app not sure about exporting from the IDE... is that even possible?

  Read the article

• C++ Beginner Delete Question

  - by Pooch

  Hi all, This is my first year learning C++ so bear with me. I am attempting to dynamically allocate memory to the heap and then delete the allocated memory. Below is the code that is giving me a hard time: // String.cpp #include "String.h" String::String() {} String::String(char* source) { this->Size = this->GetSize(source); this->CharArray = new char[this->Size + 1]; int i = 0; for (; i < this->Size; i++) this->CharArray[i] = source[i]; this->CharArray[i] = '\0'; } int String::GetSize(const char * source) { int i = 0; for (; source[i] != '\0'; i++); return i; } String::~String() { delete[] this->CharArray; } Here is the error I get when the compiler tries to delete the CharArray: 0xC0000005: Access violation reading location 0xccccccc0. And here is the last call on the stack: msvcr100d.dll!operator delete(void * pUserData) Line 52 + 0x3 bytes C++ I am fairly certain the error exists within this piece of code but will provide you with any other information needed. Oh yeah, using VS 2010 for XP. Thanks for any and all help!

  Read the article

• Simplest distributed persistent key/value store that supports primary key range queries

  - by StaxMan

  I am looking for a properly distributed (i.e. not just sharded) and persisted (not bounded by available memory on single node, or cluster of nodes) key/value ("nosql") store that does support range queries by primary key. So far closest such system is Cassandra, which does above. However, it adds support for other features that are not essential for me. So while I like it (and will consider using it of course), I am trying to figure out if there might be other mature projects that implement what I need. Specifically, for me the only aspect of value I need is to access it as a blob. For key, however, I need range queries (as in, access values ordered, limited by start and/or end values). While values can have structures, there is no need to use that structure for anything on server side (can do client-side data binding, flexible value/content types etc). For added bonus, Cassandra style storage (journaled, all sequential writes) seems quite optimal for my use case. To help filter out answers, I have investigated some alternatives within general domain like: Voldemort (key/value, but no ordering) and CouchDB (just sharded, more batch-oriented); and am aware of systems that are not quite distributed while otherwise qualifying (bdb variants, tokyo cabinet itself (not sure if Tyrant might qualify), redis (in-memory store only)).

  Read the article

• Question about cloning in Java

  - by devoured elysium

  In Effective Java, the author states that: If a class implements Cloneable, Object's clone method returns a field-by-field copy of the object; otherwise it throws CloneNotSupportedException. What I'd like to know is what he means with field-by-field copy. Does it mean that if the class has X bytes in memory, it will just copy that piece of memory? If yes, then can I assume all value types of the original class will be copied to the new object? class Point { private int x; private int y; @Override public Point clone() { return (Point)super.clone(); } } If what Object.clone() does is a field by field copy of the Point class, I'd say that I wouldn't need to explicitly copy fields x and y, being that the code shown above will be more than enough to make a clone of the Point class. That is, the following bit of code is redundant: @Override public Point clone() { Point newObj = (Point)super.clone(); newObj.x = this.x; //redundant newObj.y = this.y; //redundant } Am I right? I know references of the cloned object will point automatically to where the original object's references pointed to, I'm just not sure what happens specifically with value types. If anyone could state clearly what Object.clone()'s algorithm specification is (in easy language) that'd be great. Thanks

  Read the article

• Why is Core Data not persisting these changes to disk?

  - by scott

  I added a new entity to my model and it loads fine but no changes made in memory get persisted to disk. My values set on the car object work fine in memory but aren't getting persisted to disk on hitting the home button (in simulator). I am using almost exactly the same code on another entity in my application and its values persist to disk fine (core data - sqlite3); Does anyone have a clue what I'm overlooking here? Car is the managed object, cars in an NSMutableArray of car objects and Car is the entity and Visible is the attribute on the entity which I am trying to set. Thanks for you assistance. Scott - (void)viewDidLoad { myAppDelegate* appDelegate = (myAppDelegate*)[[UIApplication sharedApplication] delegate]; NSManagedObjectContext* managedObjectContex = appDelegate.managedObjectContext; NSFetchRequest* request = [[NSFetchRequest alloc] init]; NSEntityDescription* entity = [NSEntityDescription entityForName:@"Car" inManagedObjectContext:managedObjectContex]; [request setEntity:entity]; NSSortDescriptor* sortDescriptor = [[NSSortDescriptor alloc] initWithKey:@"Name" ascending:YES]; NSArray* sortDescriptors = [[NSArray alloc] initWithObjects:sortDescriptor, nil]; [request setSortDescriptors:sortDescriptors]; [sortDescriptors release]; [sortDescriptor release]; NSError* error = nil; cars = [[managedObjectContex executeFetchRequest:request error:&error] mutableCopy]; if (cars == nil) { NSLog(@"Can't load the Cars data! Error: %@, %@", error, [error userInfo]); } [request release]; } - (void)tableView:(UITableView *)tableView didSelectRowAtIndexPath:(NSIndexPath*)indexPath { Car* car = [cars objectAtIndex:indexPath.row]; if (car.Visible == [NSNumber numberWithBool:YES]) { car.Visible = [NSNumber numberWithBool:NO]; [tableView cellForRowAtIndexPath:indexPath].accessoryType = UITableViewCellAccessoryNone; } else { car.Visible = [NSNumber numberWithBool:YES]; [tableView cellForRowAtIndexPath:indexPath].accessoryType = UITableViewCellAccessoryCheckmark; } }

  Read the article

• Using Javascript to generate and save a file

  - by Toji

  I've been fiddling with WebGL lately, and have gotten a Collada reader working. Problem is it's pretty slow (Collada is a very verbose format), so I'm going to start converting files to a easier to use format (probably JSON). Thing is, I already have the code to parse the file in Javascript, so I may as well use it as my exporter too! The problem is saving. Now, I know that I can parse the file, send the result to the server, and have the browser request the file back from the server as a download. But in reality the server has nothing to do with this particular process, so why get it involved? I already have the contents of the desired file in memory. Is there any way that I could present the user with a download using pure javascript? (I doubt it, but might as well ask...) And to be clear: I am not trying to access the filesystem without the users knowledge! The user will provide a file (probably via drag and drop), the script will transform the file in memory, and the user will be prompted to download the result. All of which should be "safe" activities as far as the browser is concerned.

  Read the article

• What is a NULL value

  - by Adi

  I am wondering , what exactly is stored in the memory when we say a particular variable pointer to be NULL. suppose I have a structure, say typdef struct MEM_LIST MEM_INSTANCE; struct MEM_LIST { char *start_addr; int size; MEM_INSTANCE *next; }; MEM_INSTANCE *front; front = (MEM_INSTANCE*)malloc(sizeof(MEM_INSTANCE*)); -1) If I make front=NULL. What will be the value which actually gets stored in the different fields of the front, say front-size ,front-start_addr. Is it 0 or something else. I have limited knowledge in this NULL thing. -2) If I do a free(front); It frees the memory which is pointed out by front. So what exactly free means here, does it make it NULL or make it all 0. -3) What can be a good strategy to deal with initialization of pointers and freeing them . Thanks in advance

  Read the article

• Problem with writing mobilesubstrate plugins for iOS

  - by overboming

  I am trying to hooking message sending for iOS 3.2, I implement my own hook on a working ExampleHook program I find on the web. But my hook apparently caused segmentation fault everytime it hooks and I don't know why. I want to hook to [NSURL initWithString:(NSString *)URLString relativeToURL:(NSURL *)baseURL; and here is my related implementation static id __$GFWInterceptor_NSURL_initWithString2(NSURL<GFWInterceptor> *_NSURL, NSString *URLString, NSURL* baseURL){ NSLog(@"We have intercepted this url: %@",URLString); [_NSURL __HelloNSURL_initWithString:URLString relativeToURL:baseURL]; establish hook Class _$NSURL = objc_getClass("NSURL"); MSHookMessage(_$NSURL, @selector(initWithString:relativeToURL:), (IMP) &__$GFWInterceptor_NSURL_initWithString2, "__HelloNSURL_"); original method declaration - (void)__HelloNSURL_initWithString:(NSString *)URLString relativeToURL:(NSURL *)baseURL; and here is my gdb backtrace Program received signal EXC_BAD_ACCESS, Could not access memory. Reason: KERN_INVALID_ADDRESS at address: 0x74696e71 0x335625f8 in objc_msgSend () (gdb) bt 0 0x335625f8 in objc_msgSend () 1 0x32c05b1a in CFStringGetLength () 2 0x32c108a8 in _CFStringIsLegalURLString () 3 0x32b1c32a in -[NSURL initWithString:relativeToURL:] () 4 0x000877c0 in __$GFWInterceptor_NSURL_initWithString2 () 5 0x32b1c220 in +[NSURL URLWithString:relativeToURL:] () 6 0x32b1c1f4 in +[NSURL URLWithString:] () 7 0x3061c614 in createUniqueWebDataURL () 8 0x3061c212 in +[WebFrame(WebInternal) _createMainFrameWithSimpleHTMLDocumentWithPage:frameView:withStyle:editable:] () and apparently it hooks, but there is some memory issue there and I can't find anything to blame now

  Read the article

• Lookup table size reduction

  - by Ryan

  Hello: I have an application in which I have to store a couple of millions of integers, I have to store them in a Look up table, obviously I cannot store such amount of data in memory and in my requirements I am very limited I have to store the data in an embebedded system so I am very limited in the space, so I would like to ask you about recommended methods that I can use for the reduction of the look up table. I cannot use function approximation such as neural networks, the values needs to be in a table. The range of the integers is not known at the moment. When I say integers I mean a 32 bit value. Basically the idea is use some copmpression method to reduce the amount of memory but without losing many precision. This thing needs to run in hardware so the computation overhead cannot be very high. In my algorithm I have to access to one value of the table do some operations with it and after update the value. In the end what I should have is a function which I pass an index to it and then I get a value, and after I have to use another function to write a value in the table. I found one called tile coding http://www.cs.ualberta.ca/~sutton/book/8/node6.html, this one is based on several look up tables, does anyone know any other method?. Thanks.

  Read the article

• Handling user interface in a multi-threaded application (or being forced to have a UI-only main thre

  - by Patrick

  In my application, I have a 'logging' window, which shows all the logging, warnings, errors of the application. Last year my application was still single-threaded so this worked [quite] good. Now I am introducing multithreading. I quickly noticed that it's not a good idea to update the logging window from different threads. Reading some articles on keeping the UI in the main thread, I made a communication buffer, in which the other threads are adding their logging messages, and from which the main thread takes the messages and shows them in the logging window (this is done in the message loop). Now, in a part of my application, the memory usage increases dramatically, because the separate threads are generating lots of logging messages, and the main thread cannot empty the communication buffer quickly enough. After the while the memory decreases again (if the other threads have finished their work and the main thread gradually empties the communication buffer). I solved this problem by having a maximum size on the communication buffer, but then I run into a problem in the following situation: the main thread has to perform a complex action the main thread takes some parts of the action and let's separate threads execute this while the seperate threads are executing their logic, the main thread processes the results from the other threads and continues with its work if the other threads are finished Problem is that in this situation, if the other threads perform logging, there is no UI-message loop, and so the communication buffer is filled, but not emptied. I see two solutions in solving this problem: require the main thread to do regular polling of the communication buffer only performing user interface logic in the main thread (no other logic) I think the second solution seems the best, but this may not that easy to introduce in a big application (in my case it performs mathematical simulations). Are there any other solutions or tips? Or is one of the two proposed the best, easiest, most-pragmatic solution? Thanks, Patrick

  Read the article

• pure/const functions in C++

  - by Albert

  Hi, I'm thinking of using pure/const functions more heavily in my C++ code. (pure/const attribute in GCC) However, I am curious how strict I should be about it and what could possibly break. The most obvious case are debug outputs (in whatever form, could be on cout, in some file or in some custom debug class). I probably will have a lot of functions, which don't have any side effects despite this sort of debug output. No matter if the debug output is made or not, this will absolutely have no effect on the rest of my application. Or another case I'm thinking of is the use of my own SmartPointer class. In debug mode, my SmartPointer class has some global register where it does some extra checks. If I use such an object in a pure/const function, it does have some slight side effects (in the sense that some memory probably will be different) which should not have any real side effects though (in the sense that the behaviour is in any way different). Similar also for mutexes and other stuff. I can think of many complex cases where it has some side effects (in the sense of that some memory will be different, maybe even some threads are created, some filesystem manipulation is made, etc) but has no computational difference (all those side effects could very well be left out and I would even prefer that). How does it work out in practice? If I mark such functions as pure/const, could it break anything (considering that the code is all correct)?

  Read the article

• Using memcache together with conventional cache

  - by Industrial

  Hi! Here's the deal. We would have taken the complete static html road to solve performance issues, but since the site will be partially dynamic, this won't work out for us. What we have thought of instead is using memcache + eAccelerator to speed up PHP and take care of caching for the most used data. Here's our two approaches that we have thought of right now: Using memcache on all<< major queries and leaving it alone to do what it does best. Usinc memcache for most commonly retrieved data, and combining with a standard harddrive-stored cache for further usage. The major advantage of only using memcache is of course the performance, but as users increases, the memory usage gets heavy. Combining the two sounds like a more natural approach to us, even though the theoretical compromize in performance. Memcached appears to have some replication features available as well, which may come handy when it's time to increase the nodes. What approach should we use? - Is it stupid to compromize and combine the two methods? Should we insted be focusing on utilizing memcache and instead focusing on upgrading the memory as the load increases with the number of users? Thanks a lot!

  Read the article

• How to know why an animation stutters?

  - by Patrick Klug

  I have a few fairly simple animations (moving text around, moving ellipses etc.) and running in full screen (1920x1080 minus the task bar) the WPF Performance Suite reports a good framerate around 50 FPS throughout the animation. Dirty Rect Addition is somewhere around 300 rect/s, the SW frames are between 0 and 4 and the HW frames are between 3 and 5. Video memory usage is around 80 MB. Problem is that the animations stutters every other half second. My machine is a new Dell laptop XPS 15 with the GeForce GT 435 with 2GB memory. - The drivers are up to date. (The same behavior occurs on my netbook (in full screen) as well so I don't think it is hardware related.) If I make the window smaller the stutter goes away. The stutter occurs with the simplest of animations - even with just a couple of elements but adding more elements certainly makes it more noticeable. How can I find out what causes this stutter? When I think of it, I have not actually seen any WPF animations which run smoothly in full screen. Is this even possible?

  Read the article

• Spring Integration 1.0 RC2: Streaming file content?

  - by gdm

  I've been trying to find information on this, but due to the immaturity of the Spring Integration framework I haven't had much luck. Here is my desired work flow: New files are placed in an 'Incoming' directory Files are picked up using a file:inbound-channel-adapter The file content is streamed, N lines at a time, to a 'Stage 1' channel, which parses the line into an intermediary (shared) representation. This parsed line is routed to multiple 'Stage 2' channels. Each 'Stage 2' channel does its own processing on the N available lines to convert them to a final representation. This channel must have a queue which ensures no Stage 2 channel is overwhelmed in the event that one channel processes significantly slower than the others. The final representation of the N lines is written to a file. There will be as many output files as there were routing destinations in step 4. *'N' above stands for any reasonable number of lines to read at a time, from [1, whatever I can fit into memory reasonably], but is guaranteed to always be less than the number of lines in the full file. How can I accomplish streaming (steps 3, 4, 5) in Spring Integration? It's fairly easy to do without streaming the files, but my files are large enough that I cannot read the entire file into memory. As a side note, I have a working implementation of this work flow without Spring Integration, but since we're using Spring Integration in other places in our project, I'd like to try it here to see how it performs and how the resulting code compares for length and clarity.

  Read the article

• Using shared_ptr to implement RCU (read-copy-update)?

  - by yongsun

  I'm very interested in the user-space RCU (read-copy-update), and trying to simulate one via tr1::shared_ptr, here is the code, while I'm really a newbie in concurrent programming, would some experts help me to review? The basic idea is, reader calls get_reading_copy() to gain the pointer of current protected data (let's say it's generation one, or G1). writer calls get_updating_copy() to gain a copy of the G1 (let's say it's G2), and only one writer is allowed to enter the critical section. After the updating is done, writer calls update() to do a swap, and make the m_data_ptr pointing to data G2. The ongoing readers and the writer now hold the shared_ptr of G1, and either a reader or a writer will eventually deallocate the G1 data. Any new readers would get the pointer to G2, and a new writer would get the copy of G2 (let's say G3). It's possible the G1 is not released yet, so multiple generations of data my co-exists. template <typename T> class rcu_protected { public: typedef T type; typedef std::tr1::shared_ptr<type> rcu_pointer; rcu_protected() : m_data_ptr (new type()) {} rcu_pointer get_reading_copy () { spin_until_eq (m_is_swapping, 0); return m_data_ptr; } rcu_pointer get_updating_copy () { spin_until_eq (m_is_swapping, 0); while (!CAS (m_is_writing, 0, 1)) {/* do sleep for back-off when exceeding maximum retry times */} rcu_pointer new_data_ptr(new type(*m_data_ptr)); // as spin_until_eq does not have memory barrier protection, // we need to place a read barrier to protect the loading of // new_data_ptr not to be re-ordered before its construction _ReadBarrier(); return new_data_ptr; } void update (rcu_pointer new_data_ptr) { while (!CAS (m_is_swapping, 0, 1)) {} m_data_ptr.swap (new_data_ptr); // as spin_until_eq does not have memory barrier protection, // we need to place a write barrier to protect the assignments of // m_is_writing/m_is_swapping be re-ordered bofore the swapping _WriteBarrier(); m_is_writing = 0; m_is_swapping = 0; } private: volatile long m_is_writing; volatile long m_is_swapping; rcu_pointer m_data_ptr; };

  Read the article

• Efficient algorithm for Next button on a MySQL result set

  - by David Grayson

  I have a website that lets people view rows in a table (each row is a picture). There are more than 100,000 rows. You can view different subsets of the rows, and you can view them with different sort orders. While you are viewing one of the rows, you can click the "Next" or "Previous" buttons to go the next/previous row in the list. How would you implement the "Next" and "Previous" features of the website? More specifically, if you have an arbitrary query that returns a list of up to 100,000+ rows, and you know some information about the current row someone is viewing, how do you determine the NEXT row efficiently? Here is the pseudo-code of the solution I came up with when the website was young, and it worked well when there were only 1000 rows, but now that there are 100,000 rows I think it is eating up too much memory. int nextRowId(string query, int currentRowId) { array allRowIds = mysql_query(query); // Takes up a lot of memory! int currentIndex = (index of currentRowId in allRowIds); // Takes time! return allRowIds[currentIndex+1]; } While you are thinking about this problem, remember that the website can store more information about the current row than just its ID (for example, the position of the current row in the result set), and this information can be used as a hint to help determine the ID of the next row. Edit: Sorry for not mentioning this earlier, but this isn't just a static website: rows can often be added to the list, and rows can be re-ordered in the list. (Much rarer, rows can be removed from the list.) I think that I should worry about that kind of thing, but maybe you can convince me otherwise.

  Read the article

• Auto increment with a Unit Of Work

  - by Derick

  Context I'm building a persistence layer to abstract different types of databases that I'll be needing. On the relational part I have mySQL, Oracle and PostgreSQL. Let's take the following simplified MySQL tables: CREATE TABLE Contact ( ID varchar(15), NAME varchar(30) ); CREATE TABLE Address ( ID varchar(15), CONTACT_ID varchar(15), NAME varchar(50) ); I use code to generate system specific alpha numeric unique ID's fitting 15 chars in this case. Thus, if I insert a Contact record with it's Addresses I have my generated Contact.ID and Address.CONTACT_IDs before committing. I've created a Unit of Work (amongst others) as per Martin Fowler's patterns to add transaction support. I'm using a key based Identity Map in the UoW to track the changed records in memory. It works like a charm for the scenario above, all pretty standard stuff so far. The question scenario comes in when I have a database that is not under my control and the ID fields are auto-increment (or in Oracle sequences). In this case I do not have the db generated Contact.ID beforehand, so when I create my Address I do not have a value for Address.CONTACT_ID. The transaction has not been started on the DB session since all is kept in the Identity Map in memory. Question: What is a good approach to address this? (Avoiding unnecessary db round trips) Some ideas: Retrieve the last ID: I can do a call to the database to retrieve the last Id like: SELECT Auto_increment FROM information_schema.tables WHERE table_name='Contact'; But this is MySQL specific and probably something similar can be done for the other databases. If do this then would need to do the 1st insert, get the ID and then update the children (Address.CONTACT_IDs) – all in the current transaction context.

  Read the article

• How is a referencing environment generally implemented for closures?

  - by Alexandr Kurilin

  Let's say I have a statically/lexically scoped language with deep binding and I create a closure. The closure will consist of the statements I want executed plus the so called referencing environment, or, to quote this post, the collection of variables which can be used. What does this referencing environment actually look like implementation-wise? I was recently reading about ObjectiveC's implementation of blocks, and the author suggests that behind the scenes you get a copy of all of the variables on the stack and also of all the references to heap objects. The explanation claims that you get a "snapshot" of the referencing environment at the point in time of the closure's creation. Is that more or less what happens, or did I misread that? Is anything done to "freeze" a separate copy of the heap objects, or is it safe to assume that if they get modified between closure creation and the closure executing, the closure will no longer be operating on the original version of the object? If indeed there's copying being made, are there memory usage considerations in situations where one might want to create plenty of closures and store them somewhere? I think that misunderstanding of some of these concepts might lead to tricky issues like the ones Eric Lippert mentions in this blog post. It's interesting because you'd think that it wouldn't make sense to keep a reference to a value type that might be gone by the time the closure is called, but I'm guessing that in C# the compiler will figure out that the variable is needed later and put it into the heap instead. It seems that in most memory-managed languages everything is a reference and thus ObjectiveC is a somewhat unique situation with having to deal with copying what's on the stack.

  Read the article

• What about race condition in multithreaded reading?

  - by themoob

  Hi, According to an article on IBM.com, "a race condition is a situation in which two or more threads or processes are reading or writing some shared data, and the final result depends on the timing of how the threads are scheduled. Race conditions can lead to unpredictable results and subtle program bugs." . Although the article concerns Java, I have in general been taught the same definition. As far as I know, simple operation of reading from RAM is composed of setting the states of specific input lines (address, read etc.) and reading the states of output lines. This is an operation that obviously cannot be executed simultaneously by two devices and has to be serialized. Now let's suppose we have a situation when a couple of threads access an object in memory. In theory, this access should be serialized in order to prevent race conditions. But e.g. the readers/writers algorithm assumes that an arbitrary number of readers can use the shared memory at the same time. So, the question is: does one have to implement an exclusive lock for read when using multithreading (in WinAPI e.g.)? If not, why? Where is this control implemented - OS, hardware? Best regards, Kuba

  Read the article

< Previous Page | 246 247 248 249 250 251 252 253 254 255 256 257  | Next Page >