Search Results

Search found 21107 results on 845 pages for 'size optimization'.

Page 49/845 | < Previous Page | 45 46 47 48 49 50 51 52 53 54 55 56  | Next Page >

  • mySQL Optimization Suggestions

    - by Brian Schroeter
    I'm trying to optimize our mySQL configuration for our large Magento website. The reason I believe that mySQL needs to be configured further is because New Relic has shown that our SELECT queries are taking a long time (20,000+ ms) in some categories. I ran MySQLTuner 1.3.0 and got the following results... (Disclaimer: I restarted mySQL earlier after tweaking some settings, and so the results here may not be 100% accurate): >> MySQLTuner 1.3.0 - Major Hayden <[email protected]> >> Bug reports, feature requests, and downloads at http://mysqltuner.com/ >> Run with '--help' for additional options and output filtering [OK] Currently running supported MySQL version 5.5.37-35.0 [OK] Operating on 64-bit architecture -------- Storage Engine Statistics ------------------------------------------- [--] Status: +ARCHIVE +BLACKHOLE +CSV -FEDERATED +InnoDB +MRG_MYISAM [--] Data in MyISAM tables: 7G (Tables: 332) [--] Data in InnoDB tables: 213G (Tables: 8714) [--] Data in PERFORMANCE_SCHEMA tables: 0B (Tables: 17) [--] Data in MEMORY tables: 0B (Tables: 353) [!!] Total fragmented tables: 5492 -------- Security Recommendations ------------------------------------------- [!!] User '@host5.server1.autopartsnetwork.com' has no password set. [!!] User '@localhost' has no password set. [!!] User 'root@%' has no password set. -------- Performance Metrics ------------------------------------------------- [--] Up for: 5h 3m 4s (5M q [317.443 qps], 42K conn, TX: 18B, RX: 2B) [--] Reads / Writes: 95% / 5% [--] Total buffers: 35.5G global + 184.5M per thread (1024 max threads) [!!] Maximum possible memory usage: 220.0G (174% of installed RAM) [OK] Slow queries: 0% (6K/5M) [OK] Highest usage of available connections: 5% (61/1024) [OK] Key buffer size / total MyISAM indexes: 512.0M/3.1G [OK] Key buffer hit rate: 100.0% (102M cached / 45K reads) [OK] Query cache efficiency: 66.9% (3M cached / 5M selects) [!!] Query cache prunes per day: 3486361 [OK] Sorts requiring temporary tables: 0% (0 temp sorts / 812K sorts) [!!] Joins performed without indexes: 1328 [OK] Temporary tables created on disk: 11% (126K on disk / 1M total) [OK] Thread cache hit rate: 99% (61 created / 42K connections) [!!] Table cache hit rate: 19% (9K open / 49K opened) [OK] Open file limit used: 2% (712/25K) [OK] Table locks acquired immediately: 100% (5M immediate / 5M locks) [!!] InnoDB buffer pool / data size: 32.0G/213.4G [OK] InnoDB log waits: 0 -------- Recommendations ----------------------------------------------------- General recommendations: Run OPTIMIZE TABLE to defragment tables for better performance MySQL started within last 24 hours - recommendations may be inaccurate Reduce your overall MySQL memory footprint for system stability Enable the slow query log to troubleshoot bad queries Increasing the query_cache size over 128M may reduce performance Adjust your join queries to always utilize indexes Increase table_cache gradually to avoid file descriptor limits Read this before increasing table_cache over 64: http://bit.ly/1mi7c4C Variables to adjust: *** MySQL's maximum memory usage is dangerously high *** *** Add RAM before increasing MySQL buffer variables *** query_cache_size (> 512M) [see warning above] join_buffer_size (> 128.0M, or always use indexes with joins) table_cache (> 12288) innodb_buffer_pool_size (>= 213G) My my.cnf configuration is as follows... [client] port = 3306 [mysqld_safe] nice = 0 [mysqld] tmpdir = /var/lib/mysql/tmp user = mysql port = 3306 skip-external-locking character-set-server = utf8 collation-server = utf8_general_ci event_scheduler = 0 key_buffer = 512M max_allowed_packet = 64M thread_stack = 512K thread_cache_size = 512 sort_buffer_size = 24M read_buffer_size = 8M read_rnd_buffer_size = 24M join_buffer_size = 128M # for some nightly processes client sessions set the join buffer to 8 GB auto-increment-increment = 1 auto-increment-offset = 1 myisam-recover = BACKUP max_connections = 1024 # max connect errors artificially high to support behaviors of NetScaler monitors max_connect_errors = 999999 concurrent_insert = 2 connect_timeout = 5 wait_timeout = 180 net_read_timeout = 120 net_write_timeout = 120 back_log = 128 # this table_open_cache might be too low because of MySQL bugs #16244691 and #65384) table_open_cache = 12288 tmp_table_size = 512M max_heap_table_size = 512M bulk_insert_buffer_size = 512M open-files-limit = 8192 open-files = 1024 query_cache_type = 1 # large query limit supports SOAP and REST API integrations query_cache_limit = 4M # larger than 512 MB query cache size is problematic; this is typically ~60% full query_cache_size = 512M # set to true on read slaves read_only = false slow_query_log_file = /var/log/mysql/slow.log slow_query_log = 0 long_query_time = 0.2 expire_logs_days = 10 max_binlog_size = 1024M binlog_cache_size = 32K sync_binlog = 0 # SSD RAID10 technically has a write capacity of 10000 IOPS innodb_io_capacity = 400 innodb_file_per_table innodb_table_locks = true innodb_lock_wait_timeout = 30 # These servers have 80 CPU threads; match 1:1 innodb_thread_concurrency = 48 innodb_commit_concurrency = 2 innodb_support_xa = true innodb_buffer_pool_size = 32G innodb_file_per_table innodb_flush_log_at_trx_commit = 1 innodb_log_buffer_size = 2G skip-federated [mysqldump] quick quote-names single-transaction max_allowed_packet = 64M I have a monster of a server here to power our site because our catalog is very large (300,000 simple SKUs), and I'm just wondering if I'm missing anything that I can configure further. :-) Thanks!

    Read the article

  • Design advice for avoiding change in several classes

    - by Anders Svensson
    Hi, I'm trying to figure out how to design a small application more elegantly, and make it more resistant to change. Basically it is a sort of project price calculator, and the problem is that there are many parameters that can affect the pricing. I'm trying to avoid cluttering the code with a lot of if-clauses for each parameter, but still I have e.g. if-clauses in two places checking for the value of the size parameter. I have the Head First Design Patterns book, and have tried to find ideas there, but the closest I got was the decorator pattern, which has an example where starbuzz coffee sets prices depending first on condiments added, and then later in an exercise by adding a size parameter (Tall, Grande, Venti). But that didn't seem to help, because adding that parameter still seemed to add if-clause complexity in a lot of places (and this being an exercise they didn't explain that further). What I am trying to avoid is having to change several classes if a parameter were to change or a new parameter added, or at least change in as few places as possible (there's some fancy design principle word for this that I don't rememeber :-)). Here below is the code. Basically it calculates the price for a project that has the tasks "Writing" and "Analysis" with a size parameter and different pricing models. There will be other parameters coming in later too, like "How new is the product?" (New, 1-5 years old, 6-10 years old), etc. Any advice on the best design would be greatly appreciated, whether a "design pattern" or just good object oriented principles that would make it resistant to change (e.g. adding another size, or changing one of the size values, and only have to change in one place rather than in several if-clauses): public class Project { private readonly int _numberOfProducts; protected Size _size; public Task Analysis { get; set; } public Task Writing { get; set; } public Project(int numberOfProducts) { _numberOfProducts = numberOfProducts; _size = GetSize(); Analysis = new AnalysisTask(numberOfProducts, _size); Writing = new WritingTask(numberOfProducts, _size); } private Size GetSize() { if (_numberOfProducts <= 2) return Size.small; if (_numberOfProducts <= 8) return Size.medium; return Size.large; } public double GetPrice() { return Analysis.GetPrice() + Writing.GetPrice(); } } public abstract class Task { protected readonly int _numberOfProducts; protected Size _size; protected double _pricePerHour; protected Dictionary<Size, int> _hours; public abstract int TotalHours { get; } public double Price { get; set; } protected Task(int numberOfProducts, Size size) { _numberOfProducts = numberOfProducts; _size = size; } public double GetPrice() { return _pricePerHour * TotalHours; } } public class AnalysisTask : Task { public AnalysisTask(int numberOfProducts, Size size) : base(numberOfProducts, size) { _pricePerHour = 850; _hours = new Dictionary<Size, int>() { { Size.small, 56 }, { Size.medium, 104 }, { Size.large, 200 } }; } public override int TotalHours { get { return _hours[_size]; } } } public class WritingTask : Task { public WritingTask(int numberOfProducts, Size size) : base(numberOfProducts, size) { _pricePerHour = 650; _hours = new Dictionary<Size, int>() { { Size.small, 125 }, { Size.medium, 100 }, { Size.large, 60 } }; } public override int TotalHours { get { if (_size == Size.small) return _hours[_size] * _numberOfProducts; if (_size == Size.medium) return (_hours[Size.small] * 2) + (_hours[Size.medium] * (_numberOfProducts - 2)); return (_hours[Size.small] * 2) + (_hours[Size.medium] * (8 - 2)) + (_hours[Size.large] * (_numberOfProducts - 8)); } } } public enum Size { small, medium, large } public partial class Form1 : Form { public Form1() { InitializeComponent(); List<int> quantities = new List<int>(); for (int i = 0; i < 100; i++) { quantities.Add(i); } comboBoxNumberOfProducts.DataSource = quantities; } private void comboBoxNumberOfProducts_SelectedIndexChanged(object sender, EventArgs e) { Project project = new Project((int)comboBoxNumberOfProducts.SelectedItem); labelPrice.Text = project.GetPrice().ToString(); labelWriterHours.Text = project.Writing.TotalHours.ToString(); labelAnalysisHours.Text = project.Analysis.TotalHours.ToString(); } } At the end is a simple current calling code in the change event for a combobox that set size... (BTW, I don't like the fact that I have to use several dots to get to the TotalHours at the end here either, as far as I can recall, that violates the "principle of least knowledge" or "the law of demeter", so input on that would be appreciated too, but it's not the main point of the question) Regards, Anders

    Read the article

  • Social Engagement: One Size Doesn't Fit Anyone

    - by Mike Stiles
    The key to achieving meaningful social engagement is to know who you’re talking to, know what they like, and consistently deliver that kind of material to them. Every magazine for women knows this. When you read the article titles promoted on their covers, there’s no mistaking for whom that magazine is intended. And yet, confusion still reigns at many brands as to exactly whom they want to talk to, what those people want to hear, and what kind of content they should be creating for them. In most instances, the root problem is brands want to be all things to all people. Their target audience…the world! Good luck with that. It’s 2012, the age of aggregation and custom content delivery. To cope with the modern day barrage of information, people have constructed technological filters so that content they regard as being “for them” is mostly what gets through. Even if your brand is for men and women, young and old, you may want to consider social properties that divide men from women, and young from old. Yes, a man might find something in a women’s magazine that interests him. But that doesn’t mean he’s going to subscribe to it, or buy even one issue. In fact he’ll probably never see the article he’d otherwise be interested in, because in his mind, “This isn’t for me.” It wasn’t packaged for him. News Flash: men and women are different. So it’s a tall order to craft your Facebook Page or Twitter handle to simultaneously exude the motivators for both. The Harris Interactive study “2012 Connecting and Communicating Online: State of Social Media” sheds light on the differing social behaviors and drivers. -65% of women (vs. 59% of men) stay glued to social because they don’t want to miss anything. -25% of women check social when they wake up, before they check email. Only 18% of men check social before e-mail. -95% of women surveyed belong to Facebook vs. 86% of men. -67% of women log in to Facebook once a day or more vs. 54% of men. -Conventional wisdom is Pinterest is mostly a woman-thing, right? That may be true for viewing, but not true for sharing. Men are actually more likely to share on Pinterest than women, 23% to 10%. -The sharing divide extends to YouTube. 68% of women use it mainly for consumption, as opposed to 52% of men. -Women are as likely to have a Twitter account as men, but they’re much less likely to check it often. 54% of women check it once a week compared to 2/3 of men. Obviously, there are some takeaways from this depending on your target. Women don’t want to miss out on anything, so serialized content might be a good idea, right? Promotional posts that lead to a big payoff could keep them hooked. Posts for women might be better served first thing in the morning. If sharing is your goal, maybe male-targeted content is more likely to get those desired shares. And maybe Twitter is a better place to aim your male-targeted content than Facebook. Some grocery stores started experimenting with male-only aisles. The results have been impressive. Why? Because while it’s true men were finding those same items in the store just fine before, now something has been created just for them. They have a place in the store where they belong. Each brand’s strategy and targets are going to differ. The point is…know who you’re talking to, know how they behave, know what they like, and deliver content using any number of social relationship management targeting tools that meets their expectations. If, however, you’re committed to a one-size-fits-all, “our content is for everybody” strategy (or even worse, a “this is what we want to put out and we expect everybody to love it” strategy), your content will miss the mark for more often than it hits. @mikestilesPhoto via stock.schng

    Read the article

  • Optimizing sorting container of objects with heap-allocated buffers - how to avoid hard-copying buff

    - by Kache4
    I was making sure I knew how to do the op= and copy constructor correctly in order to sort() properly, so I wrote up a test case. After getting it to work, I realized that the op= was hard-copying all the data_. I figure if I wanted to sort a container with this structure (its elements have heap allocated char buffer arrays), it'd be faster to just swap the pointers around. Is there a way to do that? Would I have to write my own sort/swap function? #include <deque> //#include <string> //#include <utility> //#include <cstdlib> #include <cstring> #include <iostream> //#include <algorithm> // I use sort(), so why does this still compile when commented out? #include <boost/filesystem.hpp> #include <boost/foreach.hpp> using namespace std; namespace fs = boost::filesystem; class Page { public: // constructor Page(const char* path, const char* data, int size) : path_(fs::path(path)), size_(size), data_(new char[size]) { // cout << "Creating Page..." << endl; strncpy(data_, data, size); // cout << "done creating Page..." << endl; } // copy constructor Page(const Page& other) : path_(fs::path(other.path())), size_(other.size()), data_(new char[other.size()]) { // cout << "Copying Page..." << endl; strncpy(data_, other.data(), size_); // cout << "done copying Page..." << endl; } // destructor ~Page() { delete[] data_; } // accessors const fs::path& path() const { return path_; } const char* data() const { return data_; } int size() const { return size_; } // operators Page& operator = (const Page& other) { if (this == &other) return *this; char* newImage = new char[other.size()]; strncpy(newImage, other.data(), other.size()); delete[] data_; data_ = newImage; path_ = fs::path(other.path()); size_ = other.size(); return *this; } bool operator < (const Page& other) const { return path_ < other.path(); } private: fs::path path_; int size_; char* data_; }; class Book { public: Book(const char* path) : path_(fs::path(path)) { cout << "Creating Book..." << endl; cout << "pushing back #1" << endl; pages_.push_back(Page("image1.jpg", "firstImageData", 14)); cout << "pushing back #3" << endl; pages_.push_back(Page("image3.jpg", "thirdImageData", 14)); cout << "pushing back #2" << endl; pages_.push_back(Page("image2.jpg", "secondImageData", 15)); cout << "testing operator <" << endl; cout << pages_[0].path().string() << (pages_[0] < pages_[1]? " < " : " > ") << pages_[1].path().string() << endl; cout << pages_[1].path().string() << (pages_[1] < pages_[2]? " < " : " > ") << pages_[2].path().string() << endl; cout << pages_[0].path().string() << (pages_[0] < pages_[2]? " < " : " > ") << pages_[2].path().string() << endl; cout << "sorting" << endl; BOOST_FOREACH (Page p, pages_) cout << p.path().string() << endl; sort(pages_.begin(), pages_.end()); cout << "done sorting\n"; BOOST_FOREACH (Page p, pages_) cout << p.path().string() << endl; cout << "checking datas" << endl; BOOST_FOREACH (Page p, pages_) { char data[p.size() + 1]; strncpy((char*)&data, p.data(), p.size()); data[p.size()] = '\0'; cout << p.path().string() << " " << data << endl; } cout << "done Creating Book" << endl; } private: deque<Page> pages_; fs::path path_; }; int main() { Book* book = new Book("/some/path/"); }

    Read the article

  • How to cache dynamic javascript/jquery/ajax/json content with Akamai

    - by Starfs
    Trying to wrap my head around how things are cached on a CDN and it is new territory for me. In the document we received about sending in environment requests, it says "Dynamically-generated content will not benefit much from EdgeSuite". I feel like this is a simplified statement and there has to be a way to make it so you cache dynamically generated content if the tools are configured correctly. The site we are working with runs off a wordpress database, and uses javascript and ajax to build the pages, based on the json objects that php scripts have generated. The process - user's browser this URL, browser talks to edgesuite tools which will have cached certain pre-defined elements, and then requests from the host web server anything that is not cached, once edgesuite has compiled a combination of the two, it sends that information back to the browser. Can we not simply cache all json objects (and of course images, js, css) and therefore the web browser never has to hit the host server's database, at which point in essence, we have cached our dynamic content? Does anyone have any pointers on the most efficient configuration for this type of system -- Akamai/CDN -- to served javascript/ajax/json generated pages that ideally already hit pre-cached json data? Any and all feedback is welcome!

    Read the article

  • How to Reduce the Size of Your WinSXS Folder on Windows 7 or 8

    - by Chris Hoffman
    The WinSXS folder at C:\Windows\WinSXS is massive and continues to grow the longer you have Windows installed. This folder builds up unnecessary files over time, such as old versions of system components. This folder also contains files for uninstalled, disabled Windows components. Even if you don’t have a Windows component installed, it will be present in your WinSXS folder, taking up space. Why the WinSXS Folder Gets to Big The WinSXS folder contains all Windows system components. In fact, component files elsewhere in Windows are just links to files contained in the WinSXS folder. The WinSXS folder contains every operating system file. When Windows installs updates, it drops the new Windows component in the WinSXS folder and keeps the old component in the WinSXS folder. This means that every Windows Update you install increases the size of your WinSXS folder. This allows you to uninstall operating system updates from the Control Panel, which can be useful in the case of a buggy update — but it’s a feature that’s rarely used. Windows 7 dealt with this by including a feature that allows Windows to clean up old Windows update files after you install a new Windows service pack. The idea was that the system could be cleaned up regularly along with service packs. However, Windows 7 only saw one service pack — Service Pack 1 — released in 2010. Microsoft has no intention of launching another. This means that, for more than three years, Windows update uninstallation files have been building up on Windows 7 systems and couldn’t be easily removed. Clean Up Update Files To fix this problem, Microsoft recently backported a feature from Windows 8 to Windows 7. They did this without much fanfare — it was rolled out in a typical minor operating system update, the kind that don’t generally add new features. To clean up such update files, open the Disk Cleanup wizard (tap the Windows key, type “disk cleanup” into the Start menu, and press Enter). Click the Clean up System Files button, enable the Windows Update Cleanup option and click OK. If you’ve been using your Windows 7 system for a few years, you’ll likely be able to free several gigabytes of space. The next time you reboot after doing this, Windows will take a few minutes to clean up system files before you can log in and use your desktop. If you don’t see this feature in the Disk Cleanup window, you’re likely behind on your updates — install the latest updates from Windows Update. Windows 8 and 8.1 include built-in features that do this automatically. In fact, there’s a StartComponentCleanup scheduled task included with Windows that will automatically run in the background, cleaning up components 30 days after you’ve installed them. This 30-day period gives you time to uninstall an update if it causes problems. If you’d like to manually clean up updates, you can also use the Windows Update Cleanup option in the Disk Usage window, just as you can on Windows 7. (To open it, tap the Windows key, type “disk cleanup” to perform a search, and click the “Free up disk space by removing unnecessary files” shortcut that appears.) Windows 8.1 gives you more options, allowing you to forcibly remove all previous versions of uninstalled components, even ones that haven’t been around for more than 30 days. These commands must be run in an elevated Command Prompt — in other words, start the Command Prompt window as Administrator. For example, the following command will uninstall all previous versions of components without the scheduled task’s 30-day grace period: DISM.exe /online /Cleanup-Image /StartComponentCleanup The following command will remove files needed for uninstallation of service packs. You won’t be able to uninstall any currently installed service packs after running this command: DISM.exe /online /Cleanup-Image /SPSuperseded The following command will remove all old versions of every component. You won’t be able to uninstall any currently installed service packs or updates after this completes: DISM.exe /online /Cleanup-Image /StartComponentCleanup /ResetBase Remove Features on Demand Modern versions of Windows allow you to enable or disable Windows features on demand. You’ll find a list of these features in the Windows Features window you can access from the Control Panel. Even features you don’t have installed — that is, the features you see unchecked in this window — are stored on your hard drive in your WinSXS folder. If you choose to install them, they’ll be made available from your WinSXS folder. This means you won’t have to download anything or provide Windows installation media to install these features. However, these features take up space. While this shouldn’t matter on typical computers, users with extremely low amounts of storage or Windows server administrators who want to slim their Windows installs down to the smallest possible set of system files may want to get these files off their hard drives. For this reason, Windows 8 added a new option that allows you to remove these uninstalled components from the WinSXS folder entirely, freeing up space. If you choose to install the removed components later, Windows will prompt you to download the component files from Microsoft. To do this, open a Command Prompt window as Administrator. Use the following command to see the features available to you: DISM.exe /Online /English /Get-Features /Format:Table You’ll see a table of feature names and their states. To remove a feature from your system, you’d use the following command, replacing NAME with the name of the feature you want to remove. You can get the feature name you need from the table above. DISM.exe /Online /Disable-Feature /featurename:NAME /Remove If you run the /GetFeatures command again, you’ll now see that the feature has a status of “Disabled with Payload Removed” instead of just “Disabled.” That’s how you know it’s not taking up space on your computer’s hard drive. If you’re trying to slim down a Windows system as much as possible, be sure to check out our lists of ways to free up disk space on Windows and reduce the space used by system files.     

    Read the article

  • SQL SERVER – Introduction to Rollup Clause

    - by pinaldave
    In this article we will go over basic understanding of Rollup clause in SQL Server. ROLLUP clause is used to do aggregate operation on multiple levels in hierarchy. Let us understand how it works by using an example. Consider a table with the following structure and data: CREATE TABLE tblPopulation ( Country VARCHAR(100), [State] VARCHAR(100), City VARCHAR(100), [Population (in Millions)] INT ) GO INSERT INTO tblPopulation VALUES('India', 'Delhi','East Delhi',9 [...]

    Read the article

  • Python — Time complexity of built-in functions versus manually-built functions in finite fields

    - by stackuser
    Generally, I'm wondering about the advantages versus disadvantages of using the built-in arithmetic functions versus rolling your own in Python. Specifically, I'm taking in GF(2) finite field polynomials in string format, converting to base 2 values, performing arithmetic, then output back into polynomials as string format. So a small example of this is in multiplication: Rolling my own: def multiply(a,b): bitsa = reversed("{0:b}".format(a)) g = [(b<<i)*int(bit) for i,bit in enumerate(bitsa)] return reduce(lambda x,y: x+y,g) Versus the built-in: def multiply(a,b): # a,b are GF(2) polynomials in binary form .... return a*b #returns product of 2 polynomials in gf2 Currently, operations like multiplicative inverse (with for example 20 bit exponents) take a long time to run in my program as it's using all of Python's built-in mathematical operations like // floor division and % modulus, etc. as opposed to making my own division, remainder, etc. I'm wondering how much of a gain in efficiency and performance I can get by building these manually (as shown above). I realize the gains are dependent on how well the manual versions are built, that's not the question. I'd like to find out 'basically' how much advantage there is over the built-in's. So for instance, if multiplication (as in the example above) is well-suited for base 10 (decimal) arithmetic but has to jump through more hoops to change bases to binary and then even more hoops in operating (so it's lower efficiency), that's what I'm wondering. Like, I'm wondering if it's possible to bring the time down significantly by building them myself in ways that maybe some professionals here have already come across.

    Read the article

  • Page Speed and it&rsquo;s affect on your business

    - by ihaynes
    Originally posted on: http://geekswithblogs.net/ihaynes/archive/2014/05/29/page-speed-and-itrsquos-affect-on-your-business.aspxPage speed was an important issue 10 years ago, when we all had slow modems, but became less so with the advent of fast broadband connections that even seemed to make Ajax unnecessary.  Then along came the mobile internet and we’re back to a world where page speed and asset optimisation are critical again. If you doubt this an article on SitePoint discussing ’Page Speed and Business Metrics’ may change your mind. http://www.sitepoint.com/page-speed-business-metrics/ Here are some of the figures it quotes: Walmart – saw a 2% increase in conversions for every second of improvement in page load time. Put another way, accumulated growth of revenues went up 1% for every 100 milliseconds of load time improvement. Yahoo – for every 400 milliseconds of improvement, the site traffic increased by 9%. The bottom line is that if people have to wait more than a few seconds for a page to fully render, particularly on a mobile device, they’ll probably go elsewhere. Ignore this at your peril.   For two previous posts on the subject see: Page Weight: 10 easy fixes Xat.com Image Optimiser – Useful for RWD/Mobile

    Read the article

  • OpenGL's matrix stack vs Hand multiplying

    - by deft_code
    Which is more efficient using OpenGL's transformation stack or applying the transformations by hand. I've often heard that you should minimize the number of state transitions in your graphics pipeline. Pushing and popping translation matrices seem like a big change. However, I wonder if the graphics card might be able to more than make up for pipeline hiccup by using its parallel execution hardware to bulk multiply the vertices. My specific case. I have font rendered to a sprite sheet. The coordinates of each character or a string are calculated and added to a vertex buffer. Now I need to move that string. Would it be better to iterate through the vertex buffer and adjust each of the vertices by hand or temporarily push a new translation matrix?

    Read the article

  • Minecraft Style Chunk building problem

    - by David Torrey
    I'm having some problems with speed in my chunk engine. I timed it out, and in its current state it takes a total ~5 seconds per chunk to fill each face's list. I have a check to see if each face of a block is visible and if it is not visible, it skips it and moves on. I'm using a dictionary (unordered map) because it makes sense memorywise to just not have an entry if there is no block. I've tracked my problem down to testing if there is an entry, and accessing an entry if it does exist. If I remove the tests to see if there is an entry in the dictionary for an adjacent block, or if the block type itself is seethrough, it runs within about 2-4 milliseconds. so here's my question: Is there a faster way to check for an entry in a dictionary than .ContainsKey()? As an aside, I tried TryGetValue() and it doesn't really help with the speed that much. If I remove the ContainsKey() and keep the test where it does the IsSeeThrough for each block, it halves the time, but it's still about 2-3 seconds. It only drops to 2-4ms if I remove BOTH checks. Here is my code: using System; using System.Collections.Generic; using System.Linq; using System.Text; using System.Threading.Tasks; using System.Runtime.InteropServices; using OpenTK; using OpenTK.Graphics.OpenGL; using System.Drawing; namespace Anabelle_Lee { public enum BlockEnum { air = 0, dirt = 1, } [StructLayout(LayoutKind.Sequential,Pack=1)] public struct Coordinates<T1> { public T1 x; public T1 y; public T1 z; public override string ToString() { return "(" + x + "," + y + "," + z + ") : " + typeof(T1); } } public struct Sides<T1> { public T1 left; public T1 right; public T1 top; public T1 bottom; public T1 front; public T1 back; } public class Block { public int blockType; public bool SeeThrough() { switch (blockType) { case 0: return true; } return false ; } public override string ToString() { return ((BlockEnum)(blockType)).ToString(); } } class Chunk { private Dictionary<Coordinates<byte>, Block> mChunkData; //stores the block data private Sides<List<Coordinates<byte>>> mVBOVertexBuffer; private Sides<int> mVBOHandle; //private bool mIsChanged; private const byte mCHUNKSIZE = 16; public Chunk() { } public void InitializeChunk() { //create VBO references #if DEBUG Console.WriteLine ("Initializing Chunk"); #endif mChunkData = new Dictionary<Coordinates<byte> , Block>(); //mIsChanged = true; GL.GenBuffers(1, out mVBOHandle.left); GL.GenBuffers(1, out mVBOHandle.right); GL.GenBuffers(1, out mVBOHandle.top); GL.GenBuffers(1, out mVBOHandle.bottom); GL.GenBuffers(1, out mVBOHandle.front); GL.GenBuffers(1, out mVBOHandle.back); //make new list of vertexes for each face mVBOVertexBuffer.top = new List<Coordinates<byte>>(); mVBOVertexBuffer.bottom = new List<Coordinates<byte>>(); mVBOVertexBuffer.left = new List<Coordinates<byte>>(); mVBOVertexBuffer.right = new List<Coordinates<byte>>(); mVBOVertexBuffer.front = new List<Coordinates<byte>>(); mVBOVertexBuffer.back = new List<Coordinates<byte>>(); #if DEBUG Console.WriteLine("Chunk Initialized"); #endif } public void GenerateChunk() { #if DEBUG Console.WriteLine("Generating Chunk"); #endif for (byte i = 0; i < mCHUNKSIZE; i++) { for (byte j = 0; j < mCHUNKSIZE; j++) { for (byte k = 0; k < mCHUNKSIZE; k++) { Random blockLoc = new Random(); Coordinates<byte> randChunk = new Coordinates<byte> { x = i, y = j, z = k }; mChunkData.Add(randChunk, new Block()); mChunkData[randChunk].blockType = blockLoc.Next(0, 1); } } } #if DEBUG Console.WriteLine("Chunk Generated"); #endif } public void DeleteChunk() { //delete VBO references #if DEBUG Console.WriteLine("Deleting Chunk"); #endif GL.DeleteBuffers(1, ref mVBOHandle.left); GL.DeleteBuffers(1, ref mVBOHandle.right); GL.DeleteBuffers(1, ref mVBOHandle.top); GL.DeleteBuffers(1, ref mVBOHandle.bottom); GL.DeleteBuffers(1, ref mVBOHandle.front); GL.DeleteBuffers(1, ref mVBOHandle.back); //clear all vertex buffers ClearPolyLists(); #if DEBUG Console.WriteLine("Chunk Deleted"); #endif } public void UpdateChunk() { #if DEBUG Console.WriteLine("Updating Chunk"); #endif ClearPolyLists(); //prepare buffers //for every entry in mChunkData map foreach(KeyValuePair<Coordinates<byte>,Block> feBlockData in mChunkData) { Coordinates<byte> checkBlock = new Coordinates<byte> { x = feBlockData.Key.x, y = feBlockData.Key.y, z = feBlockData.Key.z }; //check for polygonson the left side of the cube if (checkBlock.x > 0) { //check to see if there is a key for current x - 1. if not, add the vector if (!IsVisible(checkBlock.x - 1, checkBlock.y, checkBlock.z)) { //add polygon AddPoly(checkBlock.x, checkBlock.y, checkBlock.z, mVBOHandle.left); } } else { //polygon is far left and should be added AddPoly(checkBlock.x, checkBlock.y, checkBlock.z, mVBOHandle.left); } //check for polygons on the right side of the cube if (checkBlock.x < mCHUNKSIZE - 1) { if (!IsVisible(checkBlock.x + 1, checkBlock.y, checkBlock.z)) { //add poly AddPoly(checkBlock.x, checkBlock.y, checkBlock.z, mVBOHandle.right); } } else { //poly for right add AddPoly(checkBlock.x, checkBlock.y, checkBlock.z, mVBOHandle.right); } if (checkBlock.y > 0) { //check to see if there is a key for current x - 1. if not, add the vector if (!IsVisible(checkBlock.x, checkBlock.y - 1, checkBlock.z)) { //add polygon AddPoly(checkBlock.x, checkBlock.y, checkBlock.z, mVBOHandle.bottom); } } else { //polygon is far left and should be added AddPoly(checkBlock.x, checkBlock.y, checkBlock.z, mVBOHandle.bottom); } //check for polygons on the right side of the cube if (checkBlock.y < mCHUNKSIZE - 1) { if (!IsVisible(checkBlock.x, checkBlock.y + 1, checkBlock.z)) { //add poly AddPoly(checkBlock.x, checkBlock.y, checkBlock.z, mVBOHandle.top); } } else { //poly for right add AddPoly(checkBlock.x, checkBlock.y, checkBlock.z, mVBOHandle.top); } if (checkBlock.z > 0) { //check to see if there is a key for current x - 1. if not, add the vector if (!IsVisible(checkBlock.x, checkBlock.y, checkBlock.z - 1)) { //add polygon AddPoly(checkBlock.x, checkBlock.y, checkBlock.z, mVBOHandle.back); } } else { //polygon is far left and should be added AddPoly(checkBlock.x, checkBlock.y, checkBlock.z, mVBOHandle.back); } //check for polygons on the right side of the cube if (checkBlock.z < mCHUNKSIZE - 1) { if (!IsVisible(checkBlock.x, checkBlock.y, checkBlock.z + 1)) { //add poly AddPoly(checkBlock.x, checkBlock.y, checkBlock.z, mVBOHandle.front); } } else { //poly for right add AddPoly(checkBlock.x, checkBlock.y, checkBlock.z, mVBOHandle.front); } } BuildBuffers(); #if DEBUG Console.WriteLine("Chunk Updated"); #endif } public void RenderChunk() { } public void LoadChunk() { #if DEBUG Console.WriteLine("Loading Chunk"); #endif #if DEBUG Console.WriteLine("Chunk Deleted"); #endif } public void SaveChunk() { #if DEBUG Console.WriteLine("Saving Chunk"); #endif #if DEBUG Console.WriteLine("Chunk Saved"); #endif } private bool IsVisible(int pX,int pY,int pZ) { Block testBlock; Coordinates<byte> checkBlock = new Coordinates<byte> { x = Convert.ToByte(pX), y = Convert.ToByte(pY), z = Convert.ToByte(pZ) }; if (mChunkData.TryGetValue(checkBlock,out testBlock )) //if data exists { if (testBlock.SeeThrough() == true) //if existing data is not seethrough { return true; } } return true; } private void AddPoly(byte pX, byte pY, byte pZ, int BufferSide) { //create temp array GL.BindBuffer(BufferTarget.ArrayBuffer, BufferSide); if (BufferSide == mVBOHandle.front) { //front face mVBOVertexBuffer.front.Add(new Coordinates<byte> { x = (byte)(pX) , y = (byte)(pY + 1), z = (byte)(pZ + 1) }); mVBOVertexBuffer.front.Add(new Coordinates<byte> { x = (byte)(pX) , y = (byte)(pY) , z = (byte)(pZ + 1) }); mVBOVertexBuffer.front.Add(new Coordinates<byte> { x = (byte)(pX + 1), y = (byte)(pY) , z = (byte)(pZ + 1) }); mVBOVertexBuffer.front.Add(new Coordinates<byte> { x = (byte)(pX + 1), y = (byte)(pY) , z = (byte)(pZ + 1) }); mVBOVertexBuffer.front.Add(new Coordinates<byte> { x = (byte)(pX + 1), y = (byte)(pY + 1), z = (byte)(pZ + 1) }); mVBOVertexBuffer.front.Add(new Coordinates<byte> { x = (byte)(pX) , y = (byte)(pY + 1), z = (byte)(pZ + 1) }); } else if (BufferSide == mVBOHandle.right) { //back face mVBOVertexBuffer.back.Add(new Coordinates<byte> { x = (byte)(pX + 1), y = (byte)(pY + 1), z = (byte)(pZ) }); mVBOVertexBuffer.back.Add(new Coordinates<byte> { x = (byte)(pX + 1), y = (byte)(pY) , z = (byte)(pZ) }); mVBOVertexBuffer.back.Add(new Coordinates<byte> { x = (byte)(pX) , y = (byte)(pY) , z = (byte)(pZ) }); mVBOVertexBuffer.back.Add(new Coordinates<byte> { x = (byte)(pX) , y = (byte)(pY) , z = (byte)(pZ) }); mVBOVertexBuffer.back.Add(new Coordinates<byte> { x = (byte)(pX) , y = (byte)(pY + 1), z = (byte)(pZ) }); mVBOVertexBuffer.back.Add(new Coordinates<byte> { x = (byte)(pX + 1), y = (byte)(pY + 1), z = (byte)(pZ) }); } else if (BufferSide == mVBOHandle.top) { //left face mVBOVertexBuffer.left.Add(new Coordinates<byte> { x = (byte)(pX), y = (byte)(pY + 1), z = (byte)(pZ) }); mVBOVertexBuffer.left.Add(new Coordinates<byte> { x = (byte)(pX), y = (byte)(pY) , z = (byte)(pZ) }); mVBOVertexBuffer.left.Add(new Coordinates<byte> { x = (byte)(pX), y = (byte)(pY) , z = (byte)(pZ + 1) }); mVBOVertexBuffer.left.Add(new Coordinates<byte> { x = (byte)(pX), y = (byte)(pY) , z = (byte)(pZ + 1) }); mVBOVertexBuffer.left.Add(new Coordinates<byte> { x = (byte)(pX), y = (byte)(pY + 1), z = (byte)(pZ + 1) }); mVBOVertexBuffer.left.Add(new Coordinates<byte> { x = (byte)(pX), y = (byte)(pY + 1), z = (byte)(pZ) }); } else if (BufferSide == mVBOHandle.bottom) { //right face mVBOVertexBuffer.right.Add(new Coordinates<byte> { x = (byte)(pX + 1), y = (byte)(pY + 1), z = (byte)(pZ + 1) }); mVBOVertexBuffer.right.Add(new Coordinates<byte> { x = (byte)(pX + 1), y = (byte)(pY) , z = (byte)(pZ + 1) }); mVBOVertexBuffer.right.Add(new Coordinates<byte> { x = (byte)(pX + 1), y = (byte)(pY) , z = (byte)(pZ) }); mVBOVertexBuffer.right.Add(new Coordinates<byte> { x = (byte)(pX + 1), y = (byte)(pY) , z = (byte)(pZ) }); mVBOVertexBuffer.right.Add(new Coordinates<byte> { x = (byte)(pX + 1), y = (byte)(pY + 1), z = (byte)(pZ) }); mVBOVertexBuffer.right.Add(new Coordinates<byte> { x = (byte)(pX + 1), y = (byte)(pY + 1), z = (byte)(pZ + 1) }); } else if (BufferSide == mVBOHandle.front) { //top face mVBOVertexBuffer.top.Add(new Coordinates<byte> { x = (byte)(pX) , y = (byte)(pY + 1), z = (byte)(pZ) }); mVBOVertexBuffer.top.Add(new Coordinates<byte> { x = (byte)(pX) , y = (byte)(pY + 1), z = (byte)(pZ + 1) }); mVBOVertexBuffer.top.Add(new Coordinates<byte> { x = (byte)(pX + 1), y = (byte)(pY + 1), z = (byte)(pZ + 1) }); mVBOVertexBuffer.top.Add(new Coordinates<byte> { x = (byte)(pX + 1), y = (byte)(pY + 1), z = (byte)(pZ + 1) }); mVBOVertexBuffer.top.Add(new Coordinates<byte> { x = (byte)(pX + 1), y = (byte)(pY + 1), z = (byte)(pZ) }); mVBOVertexBuffer.top.Add(new Coordinates<byte> { x = (byte)(pX) , y = (byte)(pY + 1), z = (byte)(pZ) }); } else if (BufferSide == mVBOHandle.back) { //bottom face mVBOVertexBuffer.bottom.Add(new Coordinates<byte> { x = (byte)(pX) , y = (byte)(pY), z = (byte)(pZ + 1) }); mVBOVertexBuffer.bottom.Add(new Coordinates<byte> { x = (byte)(pX) , y = (byte)(pY), z = (byte)(pZ) }); mVBOVertexBuffer.bottom.Add(new Coordinates<byte> { x = (byte)(pX + 1), y = (byte)(pY), z = (byte)(pZ) }); mVBOVertexBuffer.bottom.Add(new Coordinates<byte> { x = (byte)(pX + 1), y = (byte)(pY), z = (byte)(pZ) }); mVBOVertexBuffer.bottom.Add(new Coordinates<byte> { x = (byte)(pX + 1), y = (byte)(pY), z = (byte)(pZ + 1) }); mVBOVertexBuffer.bottom.Add(new Coordinates<byte> { x = (byte)(pX) , y = (byte)(pY), z = (byte)(pZ + 1) }); } } private void BuildBuffers() { #if DEBUG Console.WriteLine("Building Chunk Buffers"); #endif GL.BindBuffer(BufferTarget.ArrayBuffer, mVBOHandle.front); GL.BufferData(BufferTarget.ArrayBuffer, (IntPtr)(Marshal.SizeOf(new Coordinates<byte>()) * mVBOVertexBuffer.front.Count), mVBOVertexBuffer.front.ToArray(), BufferUsageHint.StaticDraw); GL.BindBuffer(BufferTarget.ArrayBuffer, mVBOHandle.back); GL.BufferData(BufferTarget.ArrayBuffer, (IntPtr)(Marshal.SizeOf(new Coordinates<byte>()) * mVBOVertexBuffer.back.Count), mVBOVertexBuffer.back.ToArray(), BufferUsageHint.StaticDraw); GL.BindBuffer(BufferTarget.ArrayBuffer, mVBOHandle.left); GL.BufferData(BufferTarget.ArrayBuffer, (IntPtr)(Marshal.SizeOf(new Coordinates<byte>()) * mVBOVertexBuffer.left.Count), mVBOVertexBuffer.left.ToArray(), BufferUsageHint.StaticDraw); GL.BindBuffer(BufferTarget.ArrayBuffer, mVBOHandle.right); GL.BufferData(BufferTarget.ArrayBuffer, (IntPtr)(Marshal.SizeOf(new Coordinates<byte>()) * mVBOVertexBuffer.right.Count), mVBOVertexBuffer.right.ToArray(), BufferUsageHint.StaticDraw); GL.BindBuffer(BufferTarget.ArrayBuffer, mVBOHandle.top); GL.BufferData(BufferTarget.ArrayBuffer, (IntPtr)(Marshal.SizeOf(new Coordinates<byte>()) * mVBOVertexBuffer.top.Count), mVBOVertexBuffer.top.ToArray(), BufferUsageHint.StaticDraw); GL.BindBuffer(BufferTarget.ArrayBuffer, mVBOHandle.bottom); GL.BufferData(BufferTarget.ArrayBuffer, (IntPtr)(Marshal.SizeOf(new Coordinates<byte>()) * mVBOVertexBuffer.bottom.Count), mVBOVertexBuffer.bottom.ToArray(), BufferUsageHint.StaticDraw); GL.BindBuffer(BufferTarget.ArrayBuffer,0); #if DEBUG Console.WriteLine("Chunk Buffers Built"); #endif } private void ClearPolyLists() { #if DEBUG Console.WriteLine("Clearing Polygon Lists"); #endif mVBOVertexBuffer.top.Clear(); mVBOVertexBuffer.bottom.Clear(); mVBOVertexBuffer.left.Clear(); mVBOVertexBuffer.right.Clear(); mVBOVertexBuffer.front.Clear(); mVBOVertexBuffer.back.Clear(); #if DEBUG Console.WriteLine("Polygon Lists Cleared"); #endif } }//END CLASS }//END NAMESPACE

    Read the article

  • Performance impact: What is the optimal payload for SqlBulkCopy.WriteToServer()?

    - by Linchi Shea
    For many years, I have been using a C# program to generate the TPC-C compliant data for testing. The program relies on the SqlBulkCopy class to load the data generated by the program into the SQL Server tables. In general, the performance of this C# data loader is satisfactory. Lately however, I found myself in a situation where I needed to generate a much larger amount of data than I typically do and the data needed to be loaded within a confined time frame. So I was driven to look into the code...(read more)

    Read the article

  • How often is software speed evident in the eyes of customers?

    - by rwong
    In theory, customers should be able to feel the software performance improvements from first-hand experience. In practice, sometimes the improvements are not noticible enough, such that in order to monetize from the improvements, it is necessary to use quotable performance figures in marketing in order to attract customers. We already know the difference between perceived performance (GUI latency, etc) and server-side performance (machines, networks, infrastructure, etc). How often is it that programmers need to go the extra length to "write up" performance analyses for which the audience is not fellow programmers, but managers and customers?

    Read the article

  • CPU Architecture and floating-point math

    - by Jo-Herman Haugholt
    I'm trying to wrap my head around some details about how floating point math is performed on the CPU, trying to better understand what data types to use etc. I think I have a fairly good understanding of how integer math is performed. If I've understood correctly, and disregarding SIMD, a 32-bit CPU will generally perform integer math at at least 32-bit precision etc. Is it correct that floating-point math is dependent on the presence of a FPU? And that the FPU on the x86 is 80-bit, so floating point math is performed at this precision unless using SIMD? What about ARM?

    Read the article

  • "Optimal" game loop for 2D side-scroller

    - by MrDatabase
    Is it possible to describe an "optimal" (in terms of performance) layout for a 2D side-scroller's game loop? In this context the "game loop" takes user input, updates the states of game objects and draws the game objects. For example having a GameObject base class with a deep inheritance hierarchy could be good for maintenance... you can do something like the following: foreach(GameObject g in gameObjects) g.update(); However I think this approach can create performance issues. On the other hand all game objects' data and functions could be global. Which would be a maintenance headache but might be closer to an optimally performing game loop. Any thoughts? I'm interested in practical applications of near optimal game loop structure... even if I get a maintenance headache in exchange for great performance.

    Read the article

  • OpenGL Performance Questions

    - by Daniel
    This subject, as with any optimisation problem, gets hit on a lot, but I just couldn't find what I (think) I want. A lot of tutorials, and even SO questions have similar tips; generally covering: Use GL face culling (the OpenGL function, not the scene logic) Only send 1 matrix to the GPU (projectionModelView combination), therefore decreasing the MVP calculations from per vertex to once per model (as it should be). Use interleaved Vertices Minimize as many GL calls as possible, batch where appropriate And possibly a few/many others. I am (for curiosity reasons) rendering 28 million triangles in my application using several vertex buffers. I have tried all the above techniques (to the best of my knowledge), and received almost no performance change. Whilst I am receiving around 40FPS in my implementation, which is by no means problematic, I am still curious as to where these optimisation 'tips' actually come into use? My CPU is idling around 20-50% during rendering, therefore I assume I am GPU bound for increasing performance. Note: I am looking into gDEBugger at the moment Cross posted at StackOverflow

    Read the article

  • How to go about unused CSS issues

    - by Saif Bechan
    I am running some speedtests on a blog, and I always get complaints about unused CSS. But this is not CSS that I never use, it is just not used on that particular page. Now I work in a structured way, but there still has to be some CSS in the file that will not be used, because you need it on another page. I do not think that using different CSS files on different pages is the way to go, I think you are much better off just creating one big file that can be cached. Now is there an elegant way of dealing with this, or do you just stick with it.

    Read the article

  • Open source vs commercial game engines

    - by Vanangamudi
    How commercial game accomplsih stunnning graphics with smooth game play? I am a huge die hard fan and follower of GNU Stallman and his philosophies and other Libre people Cmon how wud I miss Linus. but I got to admit commercial games does excellent jobs. One such good example is Assasin's Creed from Ubisoft. It has good quality graphcis and plays smoothly in my Dual core CPU with Nvidia Geforce 8400ES. Rockstar GTA4 has awesome graphcis but it's slower than AC considering the graphics quality tradeoff. Age of Empires from Ensemble studios, does include Massive crowd AI simulation, yet it plays so smoothly with eyecandy graphics and very large weapon sets and different techtree elements on the other hand. Open source games like Glest, 0A.D(still in alpha :) are not so smooth even though they have very restricted abilities? Coming to question: how do game companies achieve such optmizations, or the open source community is not doing optimizations, or there are any propriarity technological elements that benefits only the companies exists huh?? e.g the OpenSubDiv from Pixar just released open to community?? something like that. and why it is hard to implement optimizations? are there any legal restrictions???

    Read the article

< Previous Page | 45 46 47 48 49 50 51 52 53 54 55 56  | Next Page >