Search Results

Search found 13249 results on 530 pages for 'performance tuning'.

Page 468/530 | < Previous Page | 464 465 466 467 468 469 470 471 472 473 474 475  | Next Page >

• is mysql index useful on column 'state' when only doing bit-operations on the column?

  - by Geert-Jan

  I have a lot of domain entities (stored in mysql) which undergo lots of different operations. Each operation is executed from a different program. I need to keep (flow)-state for these entities which I implemented in as a long field 'flowstate' used as a bitset. to query mysql for entities which have undergone a certain operation I do something like: select * from entities where state >> 7 & 1 = 1 Indicating bit 7 (cooresponding to operation 7) has run. (<-- simplified) Anyway, I really didn't pay attention to the performance implications of this setup in the beginning, and I think I'm in a bit of trouble since queries as the above run pretty slow. What I'd like to know: Does an mysql index on 'flowstate' help at all? After all it's not a single value Mysql can quickly find using a binary sort or whatever. If it doesn't, are there any other things I could do to speed things up? . Are there special 'mask-indices' for fields with use-cases as the above? TIA, Geert-jan

  Read the article

• PHP autoloader: ignoring non-existing include

  - by Bart van Heukelom

  I have a problem with my autoloader: public function loadClass($className) { $file = str_replace(array('_', '\\'), '/', $className) . '.php'; include_once $file; } As you can see, it's quite simple. I just deduce the filename of the class and try to include it. I have a problem though; I get an exception when trying to load a non-existing class (because I have an error handler which throws exceptions). This is inconvenient, because it's also fired when you use class_exists() on a non-existing class. You don't want an exception there, just a "false" returned. I fixed this earlier by putting an @ before the include (supressing all errors). The big drawback with this, though, is that any parser/compiler errors (that are fatal) in this include won't show up (not even in the logs), resulting in a hard to find bug. What would be the best way to solve both problems at once? The easiest way would be to include something like this in the autoloader (pseudocode): foreach (path in the include_path) { if (is_readable(the path + the class name)) readable = true; } if (!readable) return; But I worry about the performance there. Would it hurt a lot? (Solved) Made it like this: public function loadClass($className) { $file = str_replace(array('_', '\\'), '/', $className) . '.php'; $paths = explode(PATH_SEPARATOR, get_include_path()); foreach ($paths as $path) { if (is_readable($path . '/' . $file)) { include_once $file; return; } } }

  Read the article

• Caching page by parts; how to pass variables calculated in cached parts into never-cached parts?

  - by Kirzilla

  Hello, Let's imagine that I have a code like this... if (!$data = $cache->load("part1_cache_id")) { $item_id = $model->getItemId(); ob_start(); echo 'Here is the cached item id: '.$item_id; $data = ob_get_contents(); ob_end_clean(); $cache->save($data, "part1_cache_id"); } echo $data; echo never_cache_function($item_id); if (!$data_2 = $cache->load("part2_cache_id")) { ob_start(); echo 'Here is the another cached part of the page...'; $data_2 = ob_get_contents(); ob_end_clean(); $cache->save("part2_cache_id"); } echo $data_2; As far as you can see I need to pass $item_id variable into never_cache_function, but if fist part is cached (part1_cache_id) then I have no way to get $item_id value. I see the only solution - serialize all data from fist part (including $item_id value); then cache serialized string and unserialize it everytime when script is executed... Something like this... if (!$data = $cache->load("part1_cache_id")) { $item_id = $model->getItemId(); $data['item_id'] = $item_id; ob_start(); echo 'Here is the cached item id: '.$item_id; $data['html'] = ob_get_contents(); ob_end_clean(); $cache->save( serialize($data), "part1_cache_id" ); } $data = unserialize($data); echo $data['html'] echo never_cache_function($data['item_id']); Is there any other ways for doing such trick? I'm looking for the most high performance solution. Thank you UPDATED And another question is - how to implement such caching into controller without separating page into two templates? Is it possible? PS: Please, do not suggest Smarty, I'm really interested in implementing custom caching.

  Read the article

• how do i work bool!

  - by user350217

  HI! im a beginner and im really getting frustrated with this whole concept. i dont understand how to exactly use the bool function. i know what it is but i cantfigure out how to make it work in my program. Im trying to say that if my variable "selection" is any letter beween 'A' and 'I' then it is valid and can continue on to the next function which is called calcExchangeAmt(amtExchanged, selection). if it is false i want it to ask the user if they want to repeat the program and if they agree to repeat i want it to clear the screen and restart to the main function. please help edit my work! this is my bool function: bool isSelectionValid(char selection, char yesNo, double amtExchanged) { bool validData; validData = true; if((selection >= 'a' && selection <= 'i') || (selection >= 'A' && selection <= 'I')) { validData = calcExchangeAmt (amtExchanged, selection); } else(validData == false); { cout<<"Do you wish to continue? (Y for Yes / N for No)"; cin>>yesNo; } do { main(); } while ((yesNo =='y')||(yesNo == 'Y')); { system("cls"); } return 0; } this is the warning that im gettng: warning C4800: 'double' : forcing value to bool 'true' or 'false' (performance warning)

  Read the article

• DataReader-DataSet Hybrid solution

  - by G33kKahuna

  My solution architects and I have exhausted both pure Dataset and Datareader solutions. Basically we have a Microsoft.NET 2.0 windows service application that pulls data based on a query and processes additional tasks per record; almost a poor mans workflow system. The recordsets are broader (in terms of the columns) and deeper (in terms of number of records). We observed that DataSet performs much better in terms of performance but runs into contraints as # of records increase say 100K+ we start seeing System.OutOfMemoryException on a 4G machine with processModel configured to run at memoryLimit set to 85. Since this is a multi-threaded app, there could be multiple threads processing different queries and building different DataSets, so we run into the exception sooner in that case DataReader on the other hand works but is a lot slower and hits other contraints; if there is some sort of disconnect it has to start over again or leaves open connections on the DB side and worst case takes down the service completely etc. So, we decided the best option would be some sort of hybrid solution. I'm open to guidance and suggestions. Are there any hybrid solutions available? Any other suggestions

  Read the article

• How to react when asked a question you already know during an interview

  - by DevNull

  The short story:- If you are asked a tough algorithmic/puzzle question during an interview, whose solution is already known to you, do you:- Honestly tell the interviewer that you know this question already? -- this could result in bursting the interviewer's ego and him increasing the complexity level of the subsequent questions. Do an Oscar deserving performance and act as if you are thinking and trying hard and slowly getting to the solution? -- depending on your acting skills, could majorly impress the interviewer making the rest of the interview easier. Long story:- OK, this question comes as a result of what happened to me in a recent telephonic interview that I gave - the interview was supposed to be all algorithmic. The interviewer started with an algorithmic question which I had luckily already seen here on Stackoverflow. The best solution to that problem is not very intuitive and is more of a you-get-it-if-you-know-it kind. Now, just to not disappoint the interviewer too much, I took a few seconds as if I was pondering on the problem and then blurted out the answer which I knew too well having read and admired it on SO already. But I guess that gave it away to the interviewer that I already knew this question and since then, he started asking me for more efficient solutions and I kept coming up with approaches (even if not correct or more efficient, but I did touch a lot of different data structures and algos) and he kept asking for more efficient solutions and generally seemed put off by my initial salvo which was unexpected. What should I have done? Cheers!

  Read the article

• [Java] Nested methods vs "piped" methods, which is better?

  - by Michael Mao

  Hi: Since uni, I've programming in Java for 3 years, although I am not fully dedicated to this language, I have spent quite some time in it, nevertheless. I understand both ways, just curious which style do you prefer. public class Test{ public static void main(String[] args) { System.out.println(getAgent().getAgentName()); } private static Agent getAgent() { return new Agent(); }} class Agent{ private String getAgentName() { return "John Smith"; }} I am pretty happy with nested method calls such like the following public class Test{ public static void main(String[] args) { getAgentName(getAgent()); } private static void getAgentName(Agent agent) { System.out.println(agent.getName()); } private static Agent getAgent() { return new Agent(); }} class Agent { public String getName(){ return "John Smith"; }} They have identical output I saw "John Smith" twice. I wonder, if one way of doing this has better performance or other advantages over the other. Personally I prefer the latter, since for nested methods I can certainly tell which starts first, and which is after. The above code is but a sample, The code that I am working with now is much more complicated, a bit like a maze... So switching between the two styles often blows my head in no time.

  Read the article

• Does this mimic perfectly a function template specialization?

  - by zeroes00

  Since the function template in the following code is a member of a class template, it can't be specialized without specializing the enclosing class. But if the compiler's full optimizations are on (assume Visual Studio 2010), will the if-else-statement in the following code get optimized out? And if it does, wouldn't it mean that for all practical purposes this IS a function template specialization without any performance cost? template<typename T> struct Holder { T data; template<int Number> void saveReciprocalOf(); }; template<typename T> template<int Number> void Holder<T>::saveReciprocalOf() { //Will this if-else-statement get completely optimized out if(Number == 0) data = (T)0; else data = (T)1 / Number; } //----------------------------------- void main() { Holder<float> holder; holder.saveReciprocalOf<2>(); cout << holder.data << endl; }

  Read the article

• Question about creating device-compatible bitmaps in C#

  - by MusiGenesis

  I am storing bitmap-like data in a two-dimensional int array. To convert this array into a GDI-compatible bitmap (for use with BitBlt), I am using this function: public IntPtr GetGDIBitmap(int[,] data) { int w = data.GetLength(0); int h = data.GetLength(1); IntPtr ret = IntPtr.Zero; using (Bitmap bmp = new Bitmap(w, h)) { for (int x = 0; x < w; x++) { for (int y = 0; y < h; y++) { Color color = Color.FromArgb(data[x, y]); bmp.SetPixel(x, y, color); } } ret = bmp.GetHbitmap(); } return ret; } This works as expected, but the call to bmp.GetHbitmap() has to allocate memory for the returned bitmap. I'd like to modify this method in two (probably related) ways: I'd like to remove the intermediate Bitmap from the above code entirely, and go directly from my int[,] array to the device-compatible bitmap (i.e. the IntPtr). I presume this would involve calling CreateCompatibleBitmap, but I don't know how to go from that call to actually manipulating the pixel values. This should logically follow from the answer to the first, but I'd also like my method to re-use existing GDI bitmap handles (instead of creating a new bitmap each time). How can I do this? NOTE: I don't really use Bitmap.SetPixel(), as its performance could best be described as "glacial". The code is just for illustration.

  Read the article

• How to lazy load scripts in YUI that accompany ajax html fragments

  - by Chris Beck

  I have a web app with Tabs for Messages and Contacts (think gmail). Each Tab has a Y.on('click') event listener that retrieves an HTML fragment via Y.io() and renders the fragment via node.setContent(). However, the Contact Tab also requires contact.js to enable an Edit button in the fragment. How do I defer the cost of loading contact.js until the user clicks on the Contacts tab? How should contact.js add it's listener to the Edit button? The Complete function of my Tab's on('click') event could serialize Get.script('contact.js') after Y.io('fragment'). However, for better performance, I would prefer to download the script in parallel to downloading the HTML fragment. In that case, I would need to defer adding an event listener to the Edit button until the Edit button is available. This seems like a common RIA design pattern. What is the best way to implement this with YUI? How should I get the script? How should I defer sections of the script until elements in the fragment are available in the DOM?

  Read the article

• Custom UIView using CALayers disappears after 180º rotation or navigation controller pop

  - by Steve Madsen

  I have a created a custom UIView subclass that is exhibiting some strange behavior. It is a spinning wheel selector, and for performance reasons it is drawn entirely into two CALayer instances. The bottom layer is the wheel itself, which is rotated using setAffineTransform: according to touches. The top layer is eye candy. drawRect: is fairly simple. If the control hasn't been drawn yet (or it's been invalidated), it calls a method that creates the images and assigns them to the layer contents property. - (void) drawRect:(CGRect)rect { if (imageLayer == nil) { [self drawIntoImageLayer]; } [self updateWheelRotation]; } When the view controller using this view first appears, everything is fine. There are two instances where the view completely disappears, however: If the device is rotated a full 180°. After a view controller is popped off the navigation stack and the view becomes visible again. drawRect: is not called either time. Interestingly enough, it IS called after a 90° orientation change, and that causes the view to re-appear. How can I ensure that a custom view using CALayers is redrawn properly in these situations?

  Read the article

• Mysql Server Optimization

  - by Ish Kumar

  Hi Geeks, We are having serious MySQL(InnoDB) performance issues at a moment when we do: (10-20) insertions on TABLE1 (10-20) updates on TABLE2 Note: Both above operations happens within fraction of a second. And this occurs every few (10-15) minutes. And all online users (approx 400-600) doing read operation on join of TABLE1 & TABLE2 every 1 second. Here is our mysql configuration info: http://docs.google.com/View?id=dfrswh7c_117fmgcmb44 Issues: Lot queries wait and expire later (saw it from phpmyadmin / processes). My poor MySQL server crashes sometimes Questions Q1: Any suggestions to optimize at MySQL level? Q2: I thinking to use persistent connections at application level, is it right? Info Added Later: Database Engine: InnoDB TABLE1 : 400,000 rows (inserting 8,000 daily) & TABLE2: 8,000 rows 1 second query: SELECT b.id, b.user_id, b.description, b.debit, b.created, b.price, u.username, u.email, u.mobile FROM TABLE1 b, TABLE2 u WHERE b.credit = 0 AND b.user_id = u.id AND b.auction_id = "12345" ORDER BY b.id DESC LIMIT 10; // there are few more but they are not so critical. Indexing is good, we are using them wisely. In above query all id's are indexed And TABLE1 has frequent insertions and TABLE2 has frequent updates.

  Read the article

• Is it safe to reuse javax.xml.ws.Service objects

  - by Noel Ang

  I have JAX-WS style web service client that was auto-generated with the NetBeans IDE. The generated proxy factory (extends javax.xml.ws.Service) delegates proxy creation to the various Service.getPort methods. The application that I am maintaining instantiates the factory and obtains a proxy each time it calls the targetted service. Creating the new proxy factory instances repeatedly has been shown to be expensive, given that the WSDL documentation supplied to the factory constructor, an HTTP URI, is re-retrieved for each instantiation. We had success in improving the performance by caching the WSDL. But this has ugly maintenance and packaging implications for us. I would like to explore the suitability of caching the proxy factory itself. Is it safe, e.g., can two different client classes, executing on the same JVM and targetting the same web service, safely use the same factory to obtain distinct proxy objects (or a shared, reentrant one)? I've been unable to find guidance from either the JAX-WS specification nor the javax.xml.ws API documentation. The factory-proxy multiplicity is unclear to me. Having Service.getPort rather than Service.createPort does not inspire confidence.

  Read the article

• How/when to hire new programmers, and how to integrate them?

  - by Shaul

  Hiring new programmers, especially in a small company, can often present a Catch-22 situation. We have too much work to do, so we need to hire new programmers. But we can't hire new programmers now, because they will need mentoring and several months of learning curve in your industry/product/environment before they're useful, and none of the programmers has time to be a mentor to a new programmer, because they're all completely swamped with the current work load. That may be a slightly frivolous way of describing the situation, but nevertheless, it's difficult for a small company on a tight budget to justify hiring someone who is not only going to be unproductive for a long time, but will also take away from the performance of the current programmers. How have you dealt with this kind of situation? When is the best time to hire someone? What are the best tasks to assign to a new team member so that they can learn their way around your code base and start getting their hands dirty as quickly as possible? How do you get the new guy useful without bogging your existing programmers down in too much mentoring? Any comments & suggestions you have are much appreciated!

  Read the article

• sem-dynamic cdn

  - by dwi kristianto

  i'm developing couple of websites using php (directory script, etc.) and wordpress as cms. i need to improve its performance, by using cdn for static files (css, js, images). the problem is, css and javascript files are generated on the fly. i did that due to yahoo and some expert advice to combine the files into one file. also changing basic color of css files. for the time being, i use couple of small vps but still its not fast enough. i already contact maxcdn and the support guy said that they dont have such kind of services. what i need is: a cdn that will serve the request from user/visitor and there's no file in local disk, the cdn will redirect/fetch it from another domain/server. in vps, it could be done easily using combination of .htaccess and php, but NOT in the cdn. most of cdn only support purely static files. is there any such cdn that will server semi-dynamic files?

  Read the article

• Commit is VERY slow in my NHibernate / SQLite project

  - by Tom Bushell

  I've just started doing some real-world performance testing on my Fluent NHibernate / SQLite project, and am experiencing some serious delays when when I Commit to the database. By serious, I mean taking 20 - 30 seconds to Commit 30 K of data! This delay seems to get worse as the database grows. When the SQLite DB file is empty, commits happen almost instantly, but when it grows to 10 Meg, I see these huge delays. The database has 16 tables, averaging 10 columns each. One possible problem is that I'm storing a dozen or so IList members, but they are typically only 200 elements long. But this is a recent addition to Fluent NHibernate automapping, which stores each float in a single table row, so maybe that's a potential problem. Any suggestions on how to track this down? I suspect SQLite is the culprit, but maybe it's NHibernate? I don't have any experience with profilers, but am thinking of getting one. I'm aware of NHibernate Profiler - any recommendations for profilers that work well with SQLite? Here's the method that saves the data - it's just a SaveOrUpdate call and a Commit, if you ignore all the error handling and debug logging. public static void SaveMeasurement(object measurement) { Debug.WriteLine("\r\n---SaveMeasurement---"); // Get the application's database session var session = GetSession(); using (var transaction = session.BeginTransaction()) { try { session.SaveOrUpdate(measurement); } catch (Exception e) { throw new ApplicationException( "\r\n SaveMeasurement->SaveOrUpdate failed\r\n\r\n", e); } try { Debug.WriteLine("\r\n---Commit---"); transaction.Commit(); Debug.WriteLine("\r\n---Commit Complete---"); } catch (Exception e) { throw new ApplicationException( "\r\n SaveMeasurement->Commit failed\r\n\r\n", e); } } }

  Read the article

• javascript pausing consistently. How do I find what is causing it to pause?

  - by pedalpete

  I've got a fairly ajax heavy site and I'm trying to tune the performance. I have a function that runs between 20 & 200 times, depending on the user. I'm outputting the time the function takes to execute via console.time in firefox. The function takes about 4-6ms to complete. The strange thing is that on my larger test with 200 or runs through that function, it runs through the first 31, then seems to pause for almost a second before completing the last 170 or so. However, that 'pause' doesn't show up in the console.time logs, and I'm not running any other functions, and the object that gets passed to the function looks the same as all other objects that get passed in. The function is called like this for (var s in thisGroup.events){ showEvent(thisGroup.events[s]) } so, I don't see how or why it would suddenly pause near the beginning. but only pause once and then continue through. The pause ALWAYS happens on the 31st time through the function. I've taken a close look at the 'thisGroup.events[s]' that it is being run through, and it looks like this for #31 "eventId":"5106", "sid":"68", "gid":"29", "uid":"70","type":"event", "startDate":"2010-03-22","startTime":"6:00 PM","endDate":"2010-03-22","endTime":"11:00 PM","durationLength":"5", "durationTime":"5:00", "note":"", "desc":"event" The event immediately after the pause, #32 looks like this "eventId":"5111", "sid":"68", "gid":"29", "uid":"71","type":"event", "startDate":"2010-03-22","startTime":"6:00 PM","endDate":"2010-03-22","endTime":"11:00 PM","durationLength":"5", "durationTime":"5:00", "note":"", "desc":"event" another event that runs through no problem looks like this "eventId":"5113", "sid":"68", "gid":"29", "uid":"72","type":"event", "startDate":"2010-03-22","startTime":"4:30 PM","endDate":"2010-03-22","endTime":"11:00 PM","durationLength":"6.5", "durationTime":"6:30", "note":"", "desc":"event" From the console outputs, it doesn't appear as there is anything hanging or taking up time in the function itself, as the console.time for each event including #31,32 is 4ms. Another strange thing here is that the total time running the for loop across the entire object is coming out as 1014ms which is right for 200 events at 4-6ms each. Any suggestions on how to find this 'pause'? I find it very interesting that it is consistently happening between #31 & #32 only!

  Read the article

• Help on understanding multiple columns on an index?

  - by Xaisoft

  Assume I have a table called "table" and I have 3 columns, a, b, and c. What does it mean to have a non-clustered index on columns a,b? Is a nonclustered index on columns a,b the same as a nonclustered index on columns b,a? (Note the order). Also, Is a nonclustered index on column a the same as a nonclustered index on a,c? I was looking at the website sqlserver performance and they had these dmv scripts where it would tell you if you had overlapping indexes and I believe it was saying that having an index on a is the same as a,b, so it is redundant. Is this true about indexes? One last question is why is the clustered index put on the primary key. Most of the time the primary key is not queried against, so shouldn't the clustered index be on the most queried column. I am probably missing something here like having it on the primary key speeds up joins? Great explanations. Should I turn this into a wiki and change the title index explanation?

  Read the article

• How to choose light version of databse system

  - by adopilot

  I am starting one POS (Point of sale) project. Targeting system is going to be written in C# .NET 2 WinForms and as main database server We are going to use MS-SQL Server. As we have a lot of POS devices in chain for one store I will love to have backend local data base system on each POS device. Scenario are following: When main server goes down!! POS application should continue working "off-line" with local database, until connection to main server come up again. Now I am in dilemma which local database is going to be most adoptable for me. Here is some notes for helping me point me in right direction: To be Light "My POS devices art usually old and suffering with performances" To be Free "I have a lot of devices and I do not wont additional cost beside main SQL serer" One day Ill love to try all that port on Mono and Linux OS. Here is what I've researched so far: Simple XML "Light but I am afraid of performance, My main table of items is average of 10K records" SQL-Expres "I am afraid that my POS devices is poor with hardware for SQLExpres, and also hard to install on each device and configure" Less known Advantage Database Server have free distribution of offline ADT system. DBF with extended Library,"Respect for good old DBFs but that era is behind Me with clipper and DBFs" MS Access Sqlite "Mostly like for now, but I am afraid how it is going to pair with MS SQL do they have same Data taypes". I know that in this SO is a lot of subjective data, but at least can someone recommended some others lite database system, or things that I shod most take attention before I choice database.

  Read the article

• Algorithm to produce Cartesian product of arrays in depth-first order

  - by Yuri Gadow

  I'm looking for an example of how, in Ruby, a C like language, or pseudo code, to create the Cartesian product of a variable number of arrays of integers, each of differing length, and step through the results in a particular order: So given, [1,2,3],[1,2,3],[1,2,3]: [1, 1, 1] [2, 1, 1] [1, 2, 1] [1, 1, 2] [2, 2, 1] [1, 2, 2] [2, 1, 2] [2, 2, 2] [3, 1, 1] [1, 3, 1] etc. Instead of the typical result I've seen (including the example I give below): [1, 1, 1] [2, 1, 1] [3, 1, 1] [1, 2, 1] [2, 2, 1] [3, 2, 1] [1, 3, 1] [2, 3, 1] etc. The problem with this example is that the third position isn't explored at all until all combinations of of the first two are tried. In the code that uses this, that means even though the right answer is generally (the much larger equivalent of) 1,1,2 it will examine a few million possibilities instead of just a few thousand before finding it. I'm dealing with result sets of one million to hundreds of millions, so generating them and then sorting isn't doable here and would defeat the reason for ordering them in the first example, which is to find the correct answer sooner and so break out of the cartesian product generation earlier. Just in case it helps clarify any of the above, here's how I do this now (this has correct results and right performance, but not the order I want, i.e., it creates results as in the second listing above): def cartesian(a_of_a) a_of_a_len = a_of_a.size result = Array.new(a_of_a_len) j, k, a2, a2_len = nil, nil, nil, nil i = 0 while 1 do j, k = i, 0 while k < a_of_a_len a2 = a_of_a[k] a2_len = a2.size result[k] = a2[j % a2_len] j /= a2_len k += 1 end return if j > 0 yield result i += 1 end end UPDATE: I didn't make it very clear that I'm after a solution where all the combinations of 1,2 are examined before 3 is added in, then all 3 and 1, then all 3, 2 and 1, then all 3,2. In other words, explore all earlier combinations "horizontally" before "vertically." The precise order in which those possibilities are explored, i.e., 1,1,2 or 2,1,1, doesn't matter, just that all 2 and 1 are explored before mixing in 3 and so on.

  Read the article

• How should I avoid memoization causing bugs in Ruby?

  - by Andrew Grimm

  Is there a consensus on how to avoid memoization causing bugs due to mutable state? In this example, a cached result had its state mutated, and therefore gave the wrong result the second time it was called. class Greeter def initialize @greeting_cache = {} end def expensive_greeting_calculation(formality) case formality when :casual then "Hi" when :formal then "Hello" end end def greeting(formality) unless @greeting_cache.has_key?(formality) @greeting_cache[formality] = expensive_greeting_calculation(formality) end @greeting_cache[formality] end end def memoization_mutator greeter = Greeter.new first_person = "Bob" # Mildly contrived in this case, # but you could encounter this in more complex scenarios puts(greeter.greeting(:casual) << " " << first_person) # => Hi Bob second_person = "Sue" puts(greeter.greeting(:casual) << " " << second_person) # => Hi Bob Sue end memoization_mutator Approaches I can see to avoid this are: greeting could return a dup or clone of @greeting_cache[formality] greeting could freeze the result of @greeting_cache[formality]. That'd cause an exception to be raised when memoization_mutator appends strings to it. Check all code that uses the result of greeting to ensure none of it does any mutating of the string. Is there a consensus on the best approach? Is the only disadvantage of doing (1) or (2) decreased performance? (I also suspect freezing an object may not work fully if it has references to other objects) Side note: this problem doesn't affect the main application of memoization: as Fixnums are immutable, calculating Fibonacci sequences doesn't have problems with mutable state. :)

  Read the article

• Maintaining a pool of DAO Class instances vs doing new operator

  - by Fazal

  we have been trying to benchmark our application performance in multiple way for sometime now. I always believed that object creation in java using Class.newInstance() was not slow (at least after 1.4 version of java). But we anyways did a test to use newInstance method vs mainitain an object pool of 1000 objects. We did about 200K iterations of loading data from DB using JDBC and populating these objects. I was amazed (even shocked) to see that newInstance code compared to object pool code was almost 10 times slower. These objects represent tables with about 50 fields and all string type. Can someone share there thoughts on this issue as now I am more confused if object pooling of atleast some DAO instances is a better option. The pool size as I see right now should be large enough to meet size of average requests. There is a flip side as my memory footprint will go up but I am beginning to wonder if this kind of idea makes sense atleast for some of the DAO entities representing tables of about 50 or more columns Please share your ideas and let me know if this has been tried by someone or am I missing some point here

  Read the article

• Count rows against to SQL server (2005) table?

  - by David.Chu.ca

  I have a simple question with two options to get count of rows in a SQL server (2005). I am using VS 2005. There are two options to get the count: SELECT id FROM Table1 WHERE dt >= startDt AND dt < endDt;; I get a list of ids from above call in cache and then I get count by List.Count. Another option is SELECT COUNT(*) FROM Table1 WHERE dt >= startDt AND dt < endDt; The above call will get the count directly. The issue is that I had several cases of exceptions with the second method: timeout. What I found is that the table1 is too big with millions of data. When I used the first option, it seems OK. I am confused by the fact that Count() takes more time than getting all the rows(is that true?). Not sure if the aggregation call with Count() would cause SQL server to create temporary table or cache on server side and it would result in slow performance when table is too big? I am not sure what is the best way to get the count?

  Read the article

• MKL Accelerated Math Libraries for Java...

  - by Kaopua

  I've looked at the related threads on StackOverflow and Googled with not much luck. I'm also very new to Java (I'm coming from a C# and .NET background) so please bear with me. There is so much available in the Java world it's pretty overwhelming. I'm starting on a new Java-on-Linux project that requires some heavy and highly repetitious numerical calculations (i.e. statistics, FFT, Linear Algebra, Matrices, etc.). So maximizing the performance of the mathematical operations is a requirement, as is ensuring the math is correct. So hence I have an interest in finding a Java library that perhaps leverages native acceleration such as MKL, and is proven (so commercial options are definitely a possibility here). In the .NET space there are highly optimized and MKL accelerated commercial Mathematical libraries such as Centerspace NMath and Extreme Optimization. Is there anything comparable in Java? Most of the math libraries I have found for Java either do not seem to be actively maintained (such as Colt) or do not appear to leverage MKL or other native acceleration (such as Apache Commons Math). I have considered trying to leverage MKL directly from Java myself (e.g. JNI), but me being new to Java (let alone interoperating between Java and native libraries) it seemed smarter finding a Java library that has already done this correctly, efficiently, and is proven. Again I apologize if I am mistaken or misguided (even in regarding any libraries I've mentioned) and my ignorance of the Java offerings. It's a whole new world for me coming from the heavily commercialized Microsoft stock so I could easily be mistaken on where to look and regarding the Java libraries I've mentioned. I would greatly appreciate any help or advice.

  Read the article

• Stack usage with MMX intrinsics and Microsoft C++

  - by arik-funke

  I have an inline assembler loop that cumulatively adds elements from an int32 data array with MMX instructions. In particular, it uses the fact that the MMX registers can accommodate 16 int32s to calculate 16 different cumulative sums in parallel. I would now like to convert this piece of code to MMX intrinsics but I am afraid that I will suffer a performance penalty because one cannot explicitly intruct the compiler to use the 8 MMX registers to accomulate 16 independent sums. Can anybody comment on this and maybe propose a solution on how to convert the piece of code below to use intrinsics? == inline assembler (only part within the loop) == paddd mm0, [esi+edx+8*0] ; add first & second pair of int32 elements paddd mm1, [esi+edx+8*1] ; add third & fourth pair of int32 elements ... paddd mm2, [esi+edx+8*2] paddd mm3, [esi+edx+8*3] paddd mm4, [esi+edx+8*4] paddd mm5, [esi+edx+8*5] paddd mm6, [esi+edx+8*6] paddd mm7, [esi+edx+8*7] ; add 15th & 16th pair of int32 elements esi points to the beginning of the data array edx provides the offset in the data array for the current loop iteration the data array is arranged such that the elements for the 16 independent sums are interleaved.

  Read the article

< Previous Page | 464 465 466 467 468 469 470 471 472 473 474 475  | Next Page >