Search Results

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

Page 30/530 | < Previous Page | 26 27 28 29 30 31 32 33 34 35 36 37  | Next Page >

  • MySQL Cluster 7.2: Over 8x Higher Performance than Cluster 7.1

    - by Mat Keep
    0 0 1 893 5092 Homework 42 11 5974 14.0 Normal 0 false false false EN-US JA X-NONE /* Style Definitions */ table.MsoNormalTable {mso-style-name:"Table Normal"; mso-tstyle-rowband-size:0; mso-tstyle-colband-size:0; mso-style-noshow:yes; mso-style-priority:99; mso-style-parent:""; mso-padding-alt:0cm 5.4pt 0cm 5.4pt; mso-para-margin:0cm; mso-para-margin-bottom:.0001pt; mso-pagination:widow-orphan; font-size:12.0pt; font-family:Cambria; mso-ascii-font-family:Cambria; mso-ascii-theme-font:minor-latin; mso-hansi-font-family:Cambria; mso-hansi-theme-font:minor-latin; mso-ansi-language:EN-US;} Summary The scalability enhancements delivered by extensions to multi-threaded data nodes enables MySQL Cluster 7.2 to deliver over 8x higher performance than the previous MySQL Cluster 7.1 release on a recent benchmark What’s New in MySQL Cluster 7.2 MySQL Cluster 7.2 was released as GA (Generally Available) in February 2012, delivering many enhancements to performance on complex queries, new NoSQL Key / Value API, cross-data center replication and ease-of-use. These enhancements are summarized in the Figure below, and detailed in the MySQL Cluster New Features whitepaper Figure 1: Next Generation Web Services, Cross Data Center Replication and Ease-of-Use Once of the key enhancements delivered in MySQL Cluster 7.2 is extensions made to the multi-threading processes of the data nodes. Multi-Threaded Data Node Extensions The MySQL Cluster 7.2 data node is now functionally divided into seven thread types: 1) Local Data Manager threads (ldm). Note – these are sometimes also called LQH threads. 2) Transaction Coordinator threads (tc) 3) Asynchronous Replication threads (rep) 4) Schema Management threads (main) 5) Network receiver threads (recv) 6) Network send threads (send) 7) IO threads Each of these thread types are discussed in more detail below. MySQL Cluster 7.2 increases the maximum number of LDM threads from 4 to 16. The LDM contains the actual data, which means that when using 16 threads the data is more heavily partitioned (this is automatic in MySQL Cluster). Each LDM thread maintains its own set of data partitions, index partitions and REDO log. The number of LDM partitions per data node is not dynamically configurable, but it is possible, however, to map more than one partition onto each LDM thread, providing flexibility in modifying the number of LDM threads. The TC domain stores the state of in-flight transactions. This means that every new transaction can easily be assigned to a new TC thread. Testing has shown that in most cases 1 TC thread per 2 LDM threads is sufficient, and in many cases even 1 TC thread per 4 LDM threads is also acceptable. Testing also demonstrated that in some instances where the workload needed to sustain very high update loads it is necessary to configure 3 to 4 TC threads per 4 LDM threads. In the previous MySQL Cluster 7.1 release, only one TC thread was available. This limit has been increased to 16 TC threads in MySQL Cluster 7.2. The TC domain also manages the Adaptive Query Localization functionality introduced in MySQL Cluster 7.2 that significantly enhanced complex query performance by pushing JOIN operations down to the data nodes. Asynchronous Replication was separated into its own thread with the release of MySQL Cluster 7.1, and has not been modified in the latest 7.2 release. To scale the number of TC threads, it was necessary to separate the Schema Management domain from the TC domain. The schema management thread has little load, so is implemented with a single thread. The Network receiver domain was bound to 1 thread in MySQL Cluster 7.1. With the increase of threads in MySQL Cluster 7.2 it is also necessary to increase the number of recv threads to 8. This enables each receive thread to service one or more sockets used to communicate with other nodes the Cluster. The Network send thread is a new thread type introduced in MySQL Cluster 7.2. Previously other threads handled the sending operations themselves, which can provide for lower latency. To achieve highest throughput however, it has been necessary to create dedicated send threads, of which 8 can be configured. It is still possible to configure MySQL Cluster 7.2 to a legacy mode that does not use any of the send threads – useful for those workloads that are most sensitive to latency. The IO Thread is the final thread type and there have been no changes to this domain in MySQL Cluster 7.2. Multiple IO threads were already available, which could be configured to either one thread per open file, or to a fixed number of IO threads that handle the IO traffic. Except when using compression on disk, the IO threads typically have a very light load. Benchmarking the Scalability Enhancements The scalability enhancements discussed above have made it possible to scale CPU usage of each data node to more than 5x of that possible in MySQL Cluster 7.1. In addition, a number of bottlenecks have been removed, making it possible to scale data node performance by even more than 5x. Figure 2: MySQL Cluster 7.2 Delivers 8.4x Higher Performance than 7.1 The flexAsynch benchmark was used to compare MySQL Cluster 7.2 performance to 7.1 across an 8-node Intel Xeon x5670-based cluster of dual socket commodity servers (6 cores each). As the results demonstrate, MySQL Cluster 7.2 delivers over 8x higher performance per data nodes than MySQL Cluster 7.1. More details of this and other benchmarks will be published in a new whitepaper – coming soon, so stay tuned! In a following blog post, I’ll provide recommendations on optimum thread configurations for different types of server processor. You can also learn more from the Best Practices Guide to Optimizing Performance of MySQL Cluster Conclusion MySQL Cluster has achieved a range of impressive benchmark results, and set in context with the previous 7.1 release, is able to deliver over 8x higher performance per node. As a result, the multi-threaded data node extensions not only serve to increase performance of MySQL Cluster, they also enable users to achieve significantly improved levels of utilization from current and future generations of massively multi-core, multi-thread processor designs.

    Read the article

  • How to improve WinForms MSChart performance?

    - by Marcel
    Hi all, I have created some simple charts (of type FastLine) with MSChart and update them with live data, like below: . To do so, I bind an observable collection of a custom type to the chart like so: // set chart data source this._Chart.DataSource = value; //is of type ObservableCollection<SpectrumLevels> //define x and y value members for each series this._Chart.Series[0].XValueMember = "Index"; this._Chart.Series[1].XValueMember = "Index"; this._Chart.Series[0].YValueMembers = "Channel0Level"; this._Chart.Series[1].YValueMembers = "Channel1Level"; // bind data to chart this._Chart.DataBind(); //lasts 1.5 seconds for 8000 points per series At each refresh, the dataset completely changes, it is not a scrolling update! With a profiler I have found that the DataBind() call takes about 1.5 seconds. The other calls are negligible. How can I make this faster? Should I use another type than ObservableCollection? An array probably? Should I use another form of data binding? Is there some tweak for the MSChart that I may have missed? Should I use a sparsed set of date, having one value per pixel only? Have I simply reached the performance limit of MSCharts? From the type of the application to keep it "fluent", we should have multiple refreshes per second. Thanks for any hints!

    Read the article

  • Books and resources for Java Performance tuning - when working with databases, huge lists

    - by Arvind
    Hi All, I am relatively new to working on huge applications in Java. I am working on a Java web service which is pretty heavily used by various clients. The service basically queries the database (hibernate) and then works with a lot of Lists (there are adapters to convert list returned from DB to the interface which the service publishes) and I am seeing lot of issues with the service like high CPU usage or high heap space. While I can troubleshoot the performance issues using a profiler, I want to actually learn about what all I need to take care when I actually write code. Like what kind of List to use or things like using StringBuilder instead of String, etc... Is there any book or blogs which I can refer which will help me while I write new services? Also my application is multithreaded - each service call from a client is a new thread, and I want to know some best practices around that area as well. I did search the web but I found many tips which are not relevant in the latest Java 6 releases, so wanted to know what kind of resources would help a developer starting out now on Java for heavily used applications. Arvind

    Read the article

  • Silverlight combobox performance issue

    - by Vinzz
    Hi, I'm facing a performance issue with a crowded combobox (5000 items). Rendering of the drop down list is really slow (as if it was computing all items before showing any). Do you have any trick to make this dropdown display lazy? Xaml code: <Grid x:Name="LayoutRoot"> <StackPanel Orientation="Horizontal" Width="200" Height="20"> <TextBlock>Test Combo </TextBlock> <ComboBox x:Name="fooCombo" Margin="5,0,0,0"></ComboBox> </StackPanel> </Grid> code behind: public MainPage() { InitializeComponent(); List<string> li = new List<string>(); int Max = 5000; for (int i = 0; i < Max; ++i) li.Add("Item - " + i); fooCombo.ItemsSource = li; }

    Read the article

  • Increase performance on iphone at pdf rendering

    - by burki
    Hi! I have a UITableView, and in every cell there's displayed a UIImage created from a pdf. But now the performance is very bad. Here's my code I use to generate the UIImage from the PDF. Creating CGPDFDocumentRef and UIImageView (in cellForRowAtIndexPath method): ... CFURLRef pdfURL = CFBundleCopyResourceURL(CFBundleGetMainBundle(), (CFStringRef)formula.icon, NULL, NULL); CGPDFDocumentRef documentRef = CGPDFDocumentCreateWithURL((CFURLRef)pdfURL); CFRelease(pdfURL); UIImageView *imageView = [[UIImageView alloc] initWithImage:[self imageFromPDFWithDocumentRef:documentRef]]; ... Generate UIImage: - (UIImage *)imageFromPDFWithDocumentRef:(CGPDFDocumentRef)documentRef { CGPDFPageRef pageRef = CGPDFDocumentGetPage(documentRef, 1); CGRect pageRect = CGPDFPageGetBoxRect(pageRef, kCGPDFCropBox); UIGraphicsBeginImageContext(pageRect.size); CGContextRef context = UIGraphicsGetCurrentContext(); CGContextTranslateCTM(context, CGRectGetMinX(pageRect),CGRectGetMaxY(pageRect)); CGContextScaleCTM(context, 1, -1); CGContextTranslateCTM(context, -(pageRect.origin.x), -(pageRect.origin.y)); CGContextDrawPDFPage(context, pageRef); UIImage *finalImage = UIGraphicsGetImageFromCurrentImageContext(); UIGraphicsEndImageContext(); return finalImage; } What can I do to increas the speed and keep the memory low?

    Read the article

  • Iterator performance contract (and use on non-collections)

    - by polygenelubricants
    If all that you're doing is a simple one-pass iteration (i.e. only hasNext() and next(), no remove()), are you guaranteed linear time performance and/or amortized constant cost per operation? Is this specified in the Iterator contract anywhere? Are there data structures/Java Collection which cannot be iterated in linear time? java.util.Scanner implements Iterator<String>. A Scanner is hardly a data structure (e.g. remove() makes absolutely no sense). Is this considered a design blunder? Is something like PrimeGenerator implements Iterator<Integer> considered bad design, or is this exactly what Iterator is for? (hasNext() always returns true, next() computes the next number on demand, remove() makes no sense). Similarly, would it have made sense for java.util.Random implements Iterator<Double>? Should a type really implement Iterator if it's effectively only using one-third of its API? (i.e. no remove(), always hasNext())

    Read the article

  • MySQL MyISAM table performance... painfully, painfully slow

    - by Salman A
    I've got a table structure that can be summarized as follows: pagegroup * pagegroupid * name has 3600 rows page * pageid * pagegroupid * data references pagegroup; has 10000 rows; can have anything between 1-700 rows per pagegroup; the data column is of type mediumtext and the column contains 100k - 200kbytes data per row userdata * userdataid * pageid * column1 * column2 * column9 references page; has about 300,000 rows; can have about 1-50 rows per page The above structure is pretty straight forwad, the problem is that that a join from userdata to page group is terribly, terribly slow even though I have indexed all columns that should be indexed. The time needed to run a query for such a join (userdata inner_join page inner_join pagegroup) exceeds 3 minutes. This is terribly slow considering the fact that I am not selecting the data column at all. Example of the query that takes too long: SELECT userdata.column1, pagegroup.name FROM userdata INNER JOIN page USING( pageid ) INNER JOIN pagegroup USING( pagegroupid ) Please help by explaining why does it take so long and what can i do to make it faster. Edit #1 Explain returns following gibberish: id select_type table type possible_keys key key_len ref rows Extra 1 SIMPLE userdata ALL pageid 372420 1 SIMPLE page eq_ref PRIMARY,pagegroupid PRIMARY 4 topsecret.userdata.pageid 1 1 SIMPLE pagegroup eq_ref PRIMARY PRIMARY 4 topsecret.page.pagegroupid 1 Edit #2 SELECT u.field2, p.pageid FROM userdata u INNER JOIN page p ON u.pageid = p.pageid; /* 0.07 sec execution, 6.05 sec fecth */ id select_type table type possible_keys key key_len ref rows Extra 1 SIMPLE u ALL pageid 372420 1 SIMPLE p eq_ref PRIMARY PRIMARY 4 topsecret.u.pageid 1 Using index SELECT p.pageid, g.pagegroupid FROM page p INNER JOIN pagegroup g ON p.pagegroupid = g.pagegroupid; /* 9.37 sec execution, 60.0 sec fetch */ id select_type table type possible_keys key key_len ref rows Extra 1 SIMPLE g index PRIMARY PRIMARY 4 3646 Using index 1 SIMPLE p ref pagegroupid pagegroupid 5 topsecret.g.pagegroupid 3 Using where Moral of the story Keep medium/long text columns in a separate table if you run into performance problems such as this one.

    Read the article

  • Varying performance of MSVC release exe

    - by Andrew
    Hello everyone, I am curious what could be the reason for highly varying performance of the same executable. Sometimes, I run it and it takes 20 seconds and sometimes it is 110. Source is compiled with MSVC in Release mode with standard options. The code is here: vector<double> Un; vector<double> Ucur; double *pUn, *pUcur; ... // time marching for (old_time=time-logfreq, time+=dt; time <= end_time; time+=dt) { for (i=1, j=Un.size()-1, pUn=&Un[1], pUcur=&Ucur[1]; i < j; ++i, ++pUn, ++pUcur) { *pUcur = (*pUn)*(1.0-0.5*alpha*( *(pUn+1) - *(pUn-1) )); } Ucur[0] = (Un[0])*(1.0-0.5*alpha*( Un[1] - Un[j] )); Ucur[j] = (Un[j])*(1.0-0.5*alpha*( Un[0] - Un[j-1] )); Un = Ucur; }

    Read the article

  • Python performance improvement request for winkler

    - by Martlark
    I'm a python n00b and I'd like some suggestions on how to improve the algorithm to improve the performance of this method to compute the Jaro-Winkler distance of two names. def winklerCompareP(str1, str2): """Return approximate string comparator measure (between 0.0 and 1.0) USAGE: score = winkler(str1, str2) ARGUMENTS: str1 The first string str2 The second string DESCRIPTION: As described in 'An Application of the Fellegi-Sunter Model of Record Linkage to the 1990 U.S. Decennial Census' by William E. Winkler and Yves Thibaudeau. Based on the 'jaro' string comparator, but modifies it according to whether the first few characters are the same or not. """ # Quick check if the strings are the same - - - - - - - - - - - - - - - - - - # jaro_winkler_marker_char = chr(1) if (str1 == str2): return 1.0 len1 = len(str1) len2 = len(str2) halflen = max(len1,len2) / 2 - 1 ass1 = '' # Characters assigned in str1 ass2 = '' # Characters assigned in str2 #ass1 = '' #ass2 = '' workstr1 = str1 workstr2 = str2 common1 = 0 # Number of common characters common2 = 0 #print "'len1', str1[i], start, end, index, ass1, workstr2, common1" # Analyse the first string - - - - - - - - - - - - - - - - - - - - - - - - - # for i in range(len1): start = max(0,i-halflen) end = min(i+halflen+1,len2) index = workstr2.find(str1[i],start,end) #print 'len1', str1[i], start, end, index, ass1, workstr2, common1 if (index > -1): # Found common character common1 += 1 #ass1 += str1[i] ass1 = ass1 + str1[i] workstr2 = workstr2[:index]+jaro_winkler_marker_char+workstr2[index+1:] #print "str1 analyse result", ass1, common1 #print "str1 analyse result", ass1, common1 # Analyse the second string - - - - - - - - - - - - - - - - - - - - - - - - - # for i in range(len2): start = max(0,i-halflen) end = min(i+halflen+1,len1) index = workstr1.find(str2[i],start,end) #print 'len2', str2[i], start, end, index, ass1, workstr1, common2 if (index > -1): # Found common character common2 += 1 #ass2 += str2[i] ass2 = ass2 + str2[i] workstr1 = workstr1[:index]+jaro_winkler_marker_char+workstr1[index+1:] if (common1 != common2): print('Winkler: Wrong common values for strings "%s" and "%s"' % \ (str1, str2) + ', common1: %i, common2: %i' % (common1, common2) + \ ', common should be the same.') common1 = float(common1+common2) / 2.0 ##### This is just a fix ##### if (common1 == 0): return 0.0 # Compute number of transpositions - - - - - - - - - - - - - - - - - - - - - # transposition = 0 for i in range(len(ass1)): if (ass1[i] != ass2[i]): transposition += 1 transposition = transposition / 2.0 # Now compute how many characters are common at beginning - - - - - - - - - - # minlen = min(len1,len2) for same in range(minlen+1): if (str1[:same] != str2[:same]): break same -= 1 if (same > 4): same = 4 common1 = float(common1) w = 1./3.*(common1 / float(len1) + common1 / float(len2) + (common1-transposition) / common1) wn = w + same*0.1 * (1.0 - w) return wn

    Read the article

  • Haskell math performance

    - by Travis Brown
    I'm in the middle of porting David Blei's original C implementation of Latent Dirichlet Allocation to Haskell, and I'm trying to decide whether to leave some of the low-level stuff in C. The following function is one example—it's an approximation of the second derivative of lgamma: double trigamma(double x) { double p; int i; x=x+6; p=1/(x*x); p=(((((0.075757575757576*p-0.033333333333333)*p+0.0238095238095238) *p-0.033333333333333)*p+0.166666666666667)*p+1)/x+0.5*p; for (i=0; i<6 ;i++) { x=x-1; p=1/(x*x)+p; } return(p); } I've translated this into more or less idiomatic Haskell as follows: trigamma :: Double -> Double trigamma x = snd $ last $ take 7 $ iterate next (x' - 1, p') where x' = x + 6 p = 1 / x' ^ 2 p' = p / 2 + c / x' c = foldr1 (\a b -> (a + b * p)) [1, 1/6, -1/30, 1/42, -1/30, 5/66] next (x, p) = (x - 1, 1 / x ^ 2 + p) The problem is that when I run both through Criterion, my Haskell version is six or seven times slower (I'm compiling with -O2 on GHC 6.12.1). Some similar functions are even worse. I know practically nothing about Haskell performance, and I'm not terribly interested in digging through Core or anything like that, since I can always just call the handful of math-intensive C functions through FFI. But I'm curious about whether there's low-hanging fruit that I'm missing—some kind of extension or library or annotation that I could use to speed up this numeric stuff without making it too ugly.

    Read the article

  • Performance of Java matrix math libraries?

    - by dfrankow
    We are computing something whose runtime is bound by matrix operations. (Some details below if interested.) This experience prompted the following question: Do folk have experience with the performance of Java libraries for matrix math (e.g., multiply, inverse, etc.)? For example: JAMA: http://math.nist.gov/javanumerics/jama/ COLT: http://acs.lbl.gov/~hoschek/colt/ Apache commons math: http://commons.apache.org/math/ I searched and found nothing. Details of our speed comparison: We are using Intel FORTRAN (ifort (IFORT) 10.1 20070913). We have reimplemented it in Java (1.6) using Apache commons math 1.2 matrix ops, and it agrees to all of its digits of accuracy. (We have reasons for wanting it in Java.) (Java doubles, Fortran real*8). Fortran: 6 minutes, Java 33 minutes, same machine. jvisualm profiling shows much time spent in RealMatrixImpl.{getEntry,isValidCoordinate} (which appear to be gone in unreleased Apache commons math 2.0, but 2.0 is no faster). Fortran is using Atlas BLAS routines (dpotrf, etc.). Obviously this could depend on our code in each language, but we believe most of the time is in equivalent matrix operations. In several other computations that do not involve libraries, Java has not been much slower, and sometimes much faster.

    Read the article

  • Performance of map overlay in conjunction with ItemizedOverlay is very poor

    - by oviroa
    I am trying to display one png (drawable) on a map in about 300 points. I am retrieving the coordinates from a Sqlite table, dumping them in a cursor. When I try to display them by parsing through the cursor, it takes for ever for the images to be drawn, about .5 second per image. I find that to be suspiciously slow, so some insight on how I can increase performance would help. Here is the snippet of my code that does the rendering: while (!mFlavorsCursor.isAfterLast()) { Log.d("cursor",""+(i++)); point = new GeoPoint( (int)(mFlavorsCursor.getFloat(mFlavorsCursor.getColumnIndex(DataBaseHelper.KEY_LATITUDE))*1000000), (int)(mFlavorsCursor.getFloat(mFlavorsCursor.getColumnIndex(DataBaseHelper.KEY_LONGITUDE))*1000000)); overlayitem = new OverlayItem(point, "", ""); itemizedoverlay.addOverlay(overlayitem); itemizedoverlay.doPopulate(); mFlavorsCursor.moveToNext(); } mapOverlays.add(itemizedoverlay); I tried to isolate all the steps and it looks like the slow one is this: itemizedoverlay.doPopulate(); This is a public method in my class that extends ItemizedOverlay that runs the private populate() method.

    Read the article

  • Mysql InnoDB performance optimization and indexing

    - by Davide C
    Hello everybody, I have 2 databases and I need to link information between two big tables (more than 3M entries each, continuously growing). The 1st database has a table 'pages' that stores various information about web pages, and includes the URL of each one. The column 'URL' is a varchar(512) and has no index. The 2nd database has a table 'urlHops' defined as: CREATE TABLE urlHops ( dest varchar(512) NOT NULL, src varchar(512) DEFAULT NULL, timestamp timestamp NOT NULL DEFAULT CURRENT_TIMESTAMP, KEY dest_key (dest), KEY src_key (src) ) ENGINE=InnoDB DEFAULT CHARSET=latin1 Now, I need basically to issue (efficiently) queries like this: select p.id,p.URL from db1.pages p, db2.urlHops u where u.src=p.URL and u.dest=? At first, I thought to add an index on pages(URL). But it's a very long column, and I already issue a lot of INSERTs and UPDATEs on the same table (way more than the number of SELECTs I would do using this index). Other possible solutions I thought are: -adding a column to pages, storing the md5 hash of the URL and indexing it; this way I could do queries using the md5 of the URL, with the advantage of an index on a smaller column. -adding another table that contains only page id and page URL, indexing both columns. But this is maybe a waste of space, having only the advantage of not slowing down the inserts and updates I execute on 'pages'. I don't want to slow down the inserts and updates, but at the same time I would be able to do the queries on the URL efficiently. Any advice? My primary concern is performance; if needed, wasting some disk space is not a problem. Thank you, regards Davide

    Read the article

  • C# vs C - Big performance difference

    - by John
    I'm finding massive performance differences between similar code in C anc C#. The C code is: #include <stdio.h> #include <time.h> #include <math.h> main() { int i; double root; clock_t start = clock(); for (i = 0 ; i <= 100000000; i++){ root = sqrt(i); } printf("Time elapsed: %f\n", ((double)clock() - start) / CLOCKS_PER_SEC); } And the C# (console app) is: using System; using System.Collections.Generic; using System.Text; namespace ConsoleApplication2 { class Program { static void Main(string[] args) { DateTime startTime = DateTime.Now; double root; for (int i = 0; i <= 100000000; i++) { root = Math.Sqrt(i); } TimeSpan runTime = DateTime.Now - startTime; Console.WriteLine("Time elapsed: " + Convert.ToString(runTime.TotalMilliseconds/1000)); } } } With the above code, the C# completes in 0.328125 seconds (release version) and the C takes 11.14 seconds to run. The c is being compiled to a windows executable using mingw. I've always been under the assumption that C/C++ were faster or at least comparable to C#.net. What exactly is causing the C to run over 30 times slower? EDIT: It does appear that the C# optimizer was removing the root as it wasn't being used. I changed the root assignment to root += and printed out the total at the end. I've also compiled the C using cl.exe with the /O2 flag set for max speed. The results are now: 3.75 seconds for the C 2.61 seconds for the C# The C is still taking longer, but this is acceptable

    Read the article

  • Does SetFileBandwidthReservation affect memory-mapped file performance?

    - by Ghostrider
    Does this function affect Memory-mapped file performance? Here's the problem I need to solve: I have two applications competing for disk access: "reader" and "updater". Whole system runs on Windows Server 2008 R2 x64 "Updater" constantly accesses disk in a linear manner, updating data. They system is set up in such a way that updater always has infinite data to update. Consider that it is constantly approximating a solution of a huge set of equations that takes up entire 2TB disk drive. Updater uses ReadFile and WriteFile to process data in a linear fashion. "Reader" is occasionally invoked by user to get some pieces of data. Usually user would read several 4kb blocks from the drive and stop. Occasionally user needs to read up to 100mb sequentially. In exceptional cases up to several gigabytes. Reader maps files to memory to get data it needs. What I would like to achieve is for "reader" to have absolute priority so that "updater" would completely stop if needed so that "reader" could get the data user needs ASAP. Is this problem solvable by using SetPriorityClass and SetFileBandwidthReservation calls? I would really hate to put synchronization login in "reader" and "updater" and rather have the OS take care of priorities.

    Read the article

  • sql-server performance optimization by removing print statements

    - by AG
    We're going through a round of sql-server stored procedure optimizations. The one recommendation we've found that clearly applies for us is 'SET NOCOUNT ON' at the top of each procedure. (Yes, I've seen the posts that point out issues with this depending on what client objects you run the stored procedures from but these are not issues for us.) So now I'm just trying to add in a bit of common sense. If the benefit of SET NOCOUNT ON is simply to reduce network traffic by some small amount every time, wouldn't it also make sense to turn off all the PRINT statements we have in the stored procedures that we only use for debugging? I can't see how it can hurt performance. OTOH, it's a bit of a hassle to implement due to the fact that some of the print statements are the only thing within else clauses, so you can't just always comment out the one line and be done. The change carries some amount of risk so I don't want to do it if it isn't going to actually help. But I don't see eliminating print statements mentioned anywhere in articles on optimization. Is that because it is so obvious no one bothers to mention it?

    Read the article

  • ColdFusion 8 Slow Performance

    - by JoeBob
    We have started a new CF8 app and it is running dog slow. A test where we go around ColdFusion (queries within a database utility) show normal speed (80ms). CF8 returns the same query in something like 60 to 80 seconds! I have been looking online and seeing lots of posts about CF8 and performance problems, but don't get any overall sense of a solution; just lots of people trying things and saying that they didn't have the problem with CF7. We are also seeing instability on the server, and some errors relating to garbage collection and the memory heap. We have a number of other applications running on CF8 and they perform adequately...our programmer is not an expert or a guru, he just plugs away. We have isolated this down to a single query that takes forever to return, so it is not a complicated test. Are there any known CF8 problems or obvious tweaks that we should consider trying? If we have to start over and learn a new environment, I will never make deadline. JoeBob

    Read the article

  • WPF drawing performance with large numbers of geometries

    - by MyFaJoArCo
    Hello, I have problems with WPF drawing performance. There are a lot of small EllipseGeometry objects (1024 ellipses, for example), which are added to three separate GeometryGroups with different foreground brushes. After, I render it all on simple Image control. Code: DrawingGroup tmpDrawing = new DrawingGroup(); GeometryGroup onGroup = new GeometryGroup(); GeometryGroup offGroup = new GeometryGroup(); GeometryGroup disabledGroup = new GeometryGroup(); for (int x = 0; x < DisplayWidth; ++x) { for (int y = 0; y < DisplayHeight; ++y) { if (States[x, y] == true) onGroup.Children.Add(new EllipseGeometry(new Rect((double)x * EDGE, (double)y * EDGE, EDGE, EDGE))); else if (States[x, y] == false) offGroup.Children.Add(new EllipseGeometry(new Rect((double)x * EDGE, (double)y * EDGE, EDGE, EDGE))); else disabledGroup.Children.Add(new EllipseGeometry(new Rect((double)x * EDGE, (double)y * EDGE, EDGE, EDGE))); } } tmpDrawing.Children.Add(new GeometryDrawing(OnBrush, null, onGroup)); tmpDrawing.Children.Add(new GeometryDrawing(OffBrush, null, offGroup)); tmpDrawing.Children.Add(new GeometryDrawing(DisabledBrush, null, disabledGroup)); DisplayImage.Source = new DrawingImage(tmpDrawing); It works fine, but takes too much time - 0.5s on Core 2 Quad, 2s on Pentium 4. I need <0.1s everywhere. All Ellipses, how you can see, are equal. Background of control, where is my DisplayImage, is solid (black, for example), so we can use this fact. I tried to use 1024 Ellipse elements instead of Image with EllipseGeometries, and it was working much faster (~0.5s), but not enough. How to speed up it? Regards, Oleg Eremeev P.S. Sorry for my English.

    Read the article

  • visualising piano performance evaluation

    - by Dolphin
    I need to develop a performance evaluator for piano playing. Based on a midi generated from sheet music, I need to evaluate the midi of the actual playing (midi keyboard). I'm planning to evaluate the playing based on note pitch, duration and loudness. The evaluation is I suppose a comparison of the notes of the sheet music and playing in midi. But I have no idea how I can visualise (i.e. show where the person have gone wrong) this evaluation process. i.e. maybe show both the notation and highlight which note has gone wrong. But how can I show any of this in some graphical form? Or more precisely on a stave (a music score) itself. I have note details (pitch, duration) and score details (key and time signature) stored in a table, and I'm using Java. But I have no clue as in how I can put all this into graphical form. Any insight is most gratefully appreciated. Advance thanks

    Read the article

  • MySQL Normalization stored procedure performance

    - by srkiNZ84
    Hi, I've written a stored procedure in MySQL to take values currently in a table and to "Normalize" them. This means that for each value passed to the stored procedure, it checks whether the value is already in the table. If it is, then it stores the id of that row in a variable. If the value is not in the table, it stores the newly inserted value's id. The stored procedure then takes the id's and inserts them into a table which is equivalent to the original de-normailized table, but this table is fully normalized and consists of mainly foreign keys. My problem with this design is that the stored procedure takes approximately 10ms or so to return, which is too long when you're trying to work through some 10million records. My suspicion is that the performance is to do with the way in which I'm doing the inserts. i.e. INSERT INTO TableA (first_value) VALUES (argument_from_sp) ON DUPLICATE KEY UPDATE id=LAST_INSERT_ID(id); SET @TableAId = LAST_INSERT_ID(); The "ON DUPLICATE KEY UPDATE" is a bit of a hack, due to the fact that on a duplicate key I don't want to update anything but rather just return the id value of the row. If you miss this step though, the LAST_INSERT_ID() function returns the wrong value when you're trying to run the "SET ..." statement. Does anyone know of a better way to do this in MySQL? Thank you

    Read the article

  • Performance of VIEW vs. SQL statement

    - by Matt W.
    I have a query that goes something like the following: select <field list> from <table list> where <join conditions> and <condition list> and PrimaryKey in (select PrimaryKey from <table list> where <join list> and <condition list>) and PrimaryKey not in (select PrimaryKey from <table list> where <join list> and <condition list>) The sub-select queries both have multiple sub-select queries of their own that I'm not showing so as not to clutter the statement. One of the developers on my team thinks a view would be better. I disagree in that the SQL statement uses variables passed in by the program (based on the user's login Id). Are there any hard and fast rules on when a view should be used vs. using a SQL statement? What kind of performance gain issues are there in running SQL statements on their own against regular tables vs. against views. (Note that all the joins / where conditions are against indexed columns, so that shouldn't be an issue.) EDIT for clarification... Here's the query I'm working with: select obj_id from object where obj_id in( (select distinct(sec_id) from security where sec_type_id = 494 and ( (sec_usergroup_id = 3278 and sec_usergroup_type_id = 230) or (sec_usergroup_id in (select ug_gi_id from user_group where ug_ui_id = 3278) and sec_usergroup_type_id = 231) ) and sec_obj_id in ( select obj_id from object where obj_ot_id in (select of_ot_id from obj_form left outer join obj_type on ot_id = of_ot_id where ot_app_id = 87 and of_id in (select sec_obj_id from security where sec_type_id = 493 and ( (sec_usergroup_id = 3278 and sec_usergroup_type_id = 230) or (sec_usergroup_id in (select ug_gi_id from user_group where ug_ui_id = 3278) and sec_usergroup_type_id = 231) ) ) and of_usage_type_id = 131 ) ) ) ) or (obj_ot_id in (select of_ot_id from obj_form left outer join obj_type on ot_id = of_ot_id where ot_app_id = 87 and of_id in (select sec_obj_id from security where sec_type_id = 493 and ( (sec_usergroup_id = 3278 and sec_usergroup_type_id = 230) or (sec_usergroup_id in (select ug_gi_id from user_group where ug_ui_id = 3278) and sec_usergroup_type_id = 231) ) ) and of_usage_type_id = 131 ) and obj_id not in (select sec_obj_id from security where sec_type_id = 494) )

    Read the article

  • Possibility of a custom Contacts field with a set list of values and Contacts lookup performance

    - by areyling
    I'm pretty sure it's not viable to do what I'd like to based on some initial research, but I figured it couldn't hurt to ask the community of experts here in case someone knows a way. I'd like to create a custom field for contacts that the user is able to edit from the main Contacts app; however, the user should only be allowed to select from a list of four specific values. A short list of string values would be ideal, but an int with a min/max range would suffice. I'm interested in knowing if it's possible either way, but also wondering if it make sense to go this route performance wise. More specifically, would it be better to look up a contact (based on a phone number) each time a call or SMS message is received or better to store my own set of data (consisting of name, numbers, and the custom field) and just syncing contact info in a thread every so often? Or syncing contacts the first time the app is run and then registering for changes using ContentObserver? Here is a similar question with an answer that explains how to add a custom field to a contact. Thanks in advance.

    Read the article

  • Android map performance with > 800 overlays of KML data

    - by span
    I have some a shape file which I have converted to a KML file that I wish to read coordinates from and then draw paths between the coordinates on a MapView. With the help of this great post: How to draw a path on a map using kml file? I have been able to read the the KML into an ArrayList of "Placemarks". This great blog post then showed how to take a list of GeoPoints and draw a path: http://djsolid.net/blog/android---draw-a-path-array-of-points-in-mapview The example in the above post only draws one path between some points however and since I have many more paths than that I am running into some performance problems. I'm currently adding a new RouteOverlay for each of the separate paths. This results in me having over 800 overlays when they have all been added. This has a performance hit and I would love some input on what I can do to improve it. Here are some options I have considered: Try to add all the points to a List which then can be passed into a class that will extend Overlay. In that new class perhaps it would be possible to add and draw the paths in a single Overlay layer? I'm not sure on how to implement this though since the paths are not always intersecting and they have different start and end points. At the moment I'm adding each path which has several points to it's own list and then I add that to an Overlay. That results in over 700 overlays... Simplify the KML or SHP. Instead of having over 700 different paths, perhaps there is someway to merge them into perhaps 100 paths or less? Since alot of paths are intersected at some point it should be possible to modify the original SHP file so that it merges all intersections. Since I have never worked with these kinds of files before I have not been able to find a way to do this in GQIS. If someone knows how to do this I would love for some input on that. Here is a link to the group of shape files if you are interested: http://danielkvist.net/cprg_bef_cbana_polyline.shp http://danielkvist.net/cprg_bef_cbana_polyline.shx http://danielkvist.net/cprg_bef_cbana_polyline.dbf http://danielkvist.net/cprg_bef_cbana_polyline.prj Anyway, here is the code I'm using to add the Overlays. Many thanks in advance. RoutePathOverlay.java package net.danielkvist; import java.util.List; import android.graphics.Canvas; import android.graphics.Color; import android.graphics.Paint; import android.graphics.Path; import android.graphics.Point; import android.graphics.RectF; import com.google.android.maps.GeoPoint; import com.google.android.maps.MapView; import com.google.android.maps.Overlay; import com.google.android.maps.Projection; public class RoutePathOverlay extends Overlay { private int _pathColor; private final List<GeoPoint> _points; private boolean _drawStartEnd; public RoutePathOverlay(List<GeoPoint> points) { this(points, Color.RED, false); } public RoutePathOverlay(List<GeoPoint> points, int pathColor, boolean drawStartEnd) { _points = points; _pathColor = pathColor; _drawStartEnd = drawStartEnd; } private void drawOval(Canvas canvas, Paint paint, Point point) { Paint ovalPaint = new Paint(paint); ovalPaint.setStyle(Paint.Style.FILL_AND_STROKE); ovalPaint.setStrokeWidth(2); int _radius = 6; RectF oval = new RectF(point.x - _radius, point.y - _radius, point.x + _radius, point.y + _radius); canvas.drawOval(oval, ovalPaint); } public boolean draw(Canvas canvas, MapView mapView, boolean shadow, long when) { Projection projection = mapView.getProjection(); if (shadow == false && _points != null) { Point startPoint = null, endPoint = null; Path path = new Path(); // We are creating the path for (int i = 0; i < _points.size(); i++) { GeoPoint gPointA = _points.get(i); Point pointA = new Point(); projection.toPixels(gPointA, pointA); if (i == 0) { // This is the start point startPoint = pointA; path.moveTo(pointA.x, pointA.y); } else { if (i == _points.size() - 1)// This is the end point endPoint = pointA; path.lineTo(pointA.x, pointA.y); } } Paint paint = new Paint(); paint.setAntiAlias(true); paint.setColor(_pathColor); paint.setStyle(Paint.Style.STROKE); paint.setStrokeWidth(3); paint.setAlpha(90); if (getDrawStartEnd()) { if (startPoint != null) { drawOval(canvas, paint, startPoint); } if (endPoint != null) { drawOval(canvas, paint, endPoint); } } if (!path.isEmpty()) canvas.drawPath(path, paint); } return super.draw(canvas, mapView, shadow, when); } public boolean getDrawStartEnd() { return _drawStartEnd; } public void setDrawStartEnd(boolean markStartEnd) { _drawStartEnd = markStartEnd; } } MyMapActivity package net.danielkvist; import java.util.ArrayList; import java.util.Collection; import java.util.Iterator; import android.graphics.Color; import android.os.Bundle; import android.util.Log; import com.google.android.maps.GeoPoint; import com.google.android.maps.MapActivity; import com.google.android.maps.MapView; public class MyMapActivity extends MapActivity { /** Called when the activity is first created. */ @Override public void onCreate(Bundle savedInstanceState) { super.onCreate(savedInstanceState); setContentView(R.layout.main); MapView mapView = (MapView) findViewById(R.id.mapview); mapView.setBuiltInZoomControls(true); String url = "http://danielkvist.net/cprg_bef_cbana_polyline_simp1600.kml"; NavigationDataSet set = MapService.getNavigationDataSet(url); drawPath(set, Color.parseColor("#6C8715"), mapView); } /** * Does the actual drawing of the route, based on the geo points provided in * the nav set * * @param navSet * Navigation set bean that holds the route information, incl. * geo pos * @param color * Color in which to draw the lines * @param mMapView01 * Map view to draw onto */ public void drawPath(NavigationDataSet navSet, int color, MapView mMapView01) { ArrayList<GeoPoint> geoPoints = new ArrayList<GeoPoint>(); Collection overlaysToAddAgain = new ArrayList(); for (Iterator iter = mMapView01.getOverlays().iterator(); iter.hasNext();) { Object o = iter.next(); Log.d(BikeApp.APP, "overlay type: " + o.getClass().getName()); if (!RouteOverlay.class.getName().equals(o.getClass().getName())) { overlaysToAddAgain.add(o); } } mMapView01.getOverlays().clear(); mMapView01.getOverlays().addAll(overlaysToAddAgain); int totalNumberOfOverlaysAdded = 0; for(Placemark placemark : navSet.getPlacemarks()) { String path = placemark.getCoordinates(); if (path != null && path.trim().length() > 0) { String[] pairs = path.trim().split(" "); String[] lngLat = pairs[0].split(","); // lngLat[0]=longitude // lngLat[1]=latitude // lngLat[2]=height try { if(lngLat.length > 1 && !lngLat[0].equals("") && !lngLat[1].equals("")) { GeoPoint startGP = new GeoPoint( (int) (Double.parseDouble(lngLat[1]) * 1E6), (int) (Double.parseDouble(lngLat[0]) * 1E6)); GeoPoint gp1; GeoPoint gp2 = startGP; geoPoints = new ArrayList<GeoPoint>(); geoPoints.add(startGP); for (int i = 1; i < pairs.length; i++) { lngLat = pairs[i].split(","); gp1 = gp2; if (lngLat.length >= 2 && gp1.getLatitudeE6() > 0 && gp1.getLongitudeE6() > 0 && gp2.getLatitudeE6() > 0 && gp2.getLongitudeE6() > 0) { // for GeoPoint, first:latitude, second:longitude gp2 = new GeoPoint( (int) (Double.parseDouble(lngLat[1]) * 1E6), (int) (Double.parseDouble(lngLat[0]) * 1E6)); if (gp2.getLatitudeE6() != 22200000) { geoPoints.add(gp2); } } } totalNumberOfOverlaysAdded++; mMapView01.getOverlays().add(new RoutePathOverlay(geoPoints)); } } catch (NumberFormatException e) { Log.e(BikeApp.APP, "Cannot draw route.", e); } } } Log.d(BikeApp.APP, "Total overlays: " + totalNumberOfOverlaysAdded); mMapView01.setEnabled(true); } @Override protected boolean isRouteDisplayed() { // TODO Auto-generated method stub return false; } } Edit: There are of course some more files I'm using but that I have not posted. You can download the complete Eclipse project here: http://danielkvist.net/se.zip

    Read the article

  • How to display many SVGs in Java with high performance

    - by Oak
    What I want My goal is to be able to display a large number of SVG images on a single drawing area in Java, each with its own translation/rotation/scale values. I'm looking for the simplest solution allowing this, optionally even using OpenGL to speed things up. What I've Tried My initial naive approach was to use SVGSalamander to draw directly on a JPanel, but the performance was pathetic. I poked around around and learned that I should do something like manually convert each SVG into a BufferedImage created with createCompatibleImage, then do the transformations I want, then draw it using double buffering. I ran into some troubles here, and before I continued I tried looking for frameworks to simplify things. What I've Looked At I've been a bit overwhelmed by the available options, which is why I'm turning to SO for help. I've looked at: Cairo (with Glitz maybe?) Libart - not sure if this actually supports SVGs FengGUI Slick - looks promising but a bit of an overkill But couldn't decide what is best for me to start working with, and I hope someone here as experience with any of these doing similar things.

    Read the article

  • Performance of Multiple Joins

    - by geeko
    Greetings Overflowers, I need to query against objects with many/complex spacial conditions. In relational databases that is translated to many joins (possibly 10+). I'm new to this business and wondering whether to go with MS SQL Server 2008 R2 or Oracle 11g or document-based solutions such as RavenDB or simply go with some spacial database (GIS)... Any thoughts ? Regards UPDATE: Thank you all for your answers. Would anybody opt for document/spatial databases ? My database would consist of tens of millions to few billion records. Mostly read-only. Almost no updates unless in case of mistakes in input. Overnight inserts and not that frequent. The join tables are predicted beforehand but the number of self joins (tables joining themselves multiple times) is not. Small pages of results from such queries are going to be viewed on an highly interactive website so response time is critical. Any predictions on how this can perform on MS SQL Server 2008 R2 or Oracle 11g ? I'm also concerned about boosting performance by adding more servers, which one scales better ? How about PostgresQL ?

    Read the article

< Previous Page | 26 27 28 29 30 31 32 33 34 35 36 37  | Next Page >