Search Results

Search found 7490 results on 300 pages for 'algorithm analysis'.

Page 150/300 | < Previous Page | 146 147 148 149 150 151 152 153 154 155 156 157  | Next Page >

• multiple move operations and data processes in work thread

  - by younevertell

  main thread-- start workthread--StartStage(get list of positions for data process) -- move to one position -- data sampling*strong text*-- data collection--data analysis------data sampling*strong text* basically, work thread does the data sampling*strong text*-- data collection--data analysis------data sampling*strong text* loop for one positioin until press stop or target is obtained. my questions: After work thread finishs the loop for one positioin, it would end itself. now how to make the work thread moves to the next position to do the data process loop after work thread finish one position work, would not end itself until data process for all the positions are done? Thanks in advance!

  Read the article

• Can I execute a "variable statements" within a function and without defines.

  - by René Nyffenegger

  I am facing a problem that I cannot see how it is solvable without #defines or incuring a performance impact although I am sure that someone can point me to a solution. I have an algorithm that sort of produces a (large) series of values. For simplicity's sake, in the following I pretend it's a for loop in a for loop, although in my code it's more complex than that. In the core of the loop I need to do calculations with the values being produced. Although the algorithm for the values stays the same, the calculations vary. So basically, what I have is: void normal() { // "Algorithm" producing numbers (x and y): for (int x=0 ; x<1000 ; x++) { for (int y=0 ; y<1000 ; y++) { // Calculation with numbers being produced: if ( x+y == 800 && y > 790) { std::cout << x << ", " << y << std::endl; } // end of calculation }} } So, the only part I need to change is if ( x+y == 800 && y > 790) { std::cout << x << ", " << y << std::endl; } So, in order to solve that, I could construct an abstract base class: class inner_0 { public: virtual void call(int x, int y) = 0; }; and derive a "callable" class from it: class inner : public inner_0 { public: virtual void call(int x, int y) { if ( x+y == 800 && y > 790) { std::cout << x << ", " << y << std::endl; } } }; I can then pass an instance of the class to the "algorithm" like so: void O(inner i) { for (int x=0 ; x<1000 ; x++) { for (int y=0 ; y<1000 ; y++) { i.call(x,y); }} } // somewhere else.... inner I; O(I); In my case, I incur a performance hit because there is an indirect call via virtual function table. So I was thinking about a way around it. It's possible with two #defines: #define OUTER \ for (int x=0 ; x<1000 ; x++) { \ for (int y=0 ; y<1000 ; y++) { \ INNER \ }} // later... #define INNER \ if (x + y == 800 && y > 790) \ std::cout << x << ", " << y << std::endl; OUTER While this certainly works, I am not 100% happy with it because I don't necessarly like #defines. So, my question: is there a better way for what I want to achieve?

  Read the article

• Running [R] on a Netbook

  - by Thomas

  I am interested in purchasing a netbook to do field research in another country. My hardware specifications for the nebtook are fairly basic: Be rugged enough to survive a bit of wear and tear Fairly fast processing (the ability to upgrade from 1GB of RAM to 2GB) A battery life of longer than 6 hours At least a 10 inch screen A decent camera for Skyping However, I am mainly concerned about being able to do basic statistical analysis in conjunction with R Be able run a Spreadsheet program to do basic data input (like Excel or Open Office) Use R to do basic data analysis (Regression, some simulation (nothing crazy), data cleaning, and some of the functionality) Word Processing (Word or Open Office) Do you have any suggestions on which models or brands my fit my needs? Some of the models I am considering: Samsung NB-30 Toshiba NB 305 Asus Eee PC 1005HA Lenovo S10-2 Does anyone use R on a netbook, and if so do you have any recommendations on how best to optimize it? This article from Lifehacker mentions some OS. Anybody use these in conjunction with R? Any help would be much appreciated.

  Read the article

• Determining whether a class implements a generic list in a T4 template

  - by James Hollingworth

  I'm writing a T4 template which loads some classes from an assembly, does some analysis of the classes and then generates some code. One particular bit of analysis I need to do is to determine whether the class implements a generic list. I can do this pretty simply in C#, e.g. public class Foo : List<string> { } var t = typeof(Foo); if (t.BaseType != null && t.BaseType.IsGenericType && t.BaseType.GetGenericTypeDefinition() == typeof(List<>))) Console.WriteLine("Win"); However T4 templates use the FXCop introspection engine and so you do not have access to the .net reflection API. I've spent the past couple of hours in Reflector but still can't figure it out. Does anyone have any clues about how to do this?

  Read the article

• displaying python's autodoc to the user (python 3.3)

  - by Plotinus

  I'm writing a simple command line math game, and I'm using python's autodoc for my math algorithms to help me remember, for example, what a proth number is while i'm writing the algorithm, but later on I'll want to tell that information to the user as well, so they'll know what the answer was. So, for example I have: def is_proth(): """Proth numbers and numbers that fit the formula k×2^n + 1, where k are odd positive integers, and 2^n > k.""" [snip] return proths and then I tried to make a dictionary, like so: definitions = {"proths" : help(is_proth)} But it doesn't work. It prints this when I start the program, one for each item in the dictionary, and then it errors out on one of them that returns a set. And anyway, I don't want it displayed to the user until after they've played the game. Help on function is_proth in module __main__: is_proth() Proth numbers and numbers that fit the formula k×2^n + 1, where k are odd positive integers, and 2^n > k. (END) I understand the purpose of autodoc is more for helping programmers who are calling a function than for generating userdoc, but it seems inefficient to have to type out the definition of what a proth number is twice, once in a comment to help me remember what an algorithm does and then once to tell the user the answer to the game they were playing after they've won or lost.

  Read the article

• Parallelism in .NET – Part 4, Imperative Data Parallelism: Aggregation

  - by Reed

  In the article on simple data parallelism, I described how to perform an operation on an entire collection of elements in parallel.  Often, this is not adequate, as the parallel operation is going to be performing some form of aggregation. Simple examples of this might include taking the sum of the results of processing a function on each element in the collection, or finding the minimum of the collection given some criteria.  This can be done using the techniques described in simple data parallelism, however, special care needs to be taken into account to synchronize the shared data appropriately.  The Task Parallel Library has tools to assist in this synchronization. The main issue with aggregation when parallelizing a routine is that you need to handle synchronization of data.  Since multiple threads will need to write to a shared portion of data.  Suppose, for example, that we wanted to parallelize a simple loop that looked for the minimum value within a dataset: double min = double.MaxValue; foreach(var item in collection) { double value = item.PerformComputation(); min = System.Math.Min(min, value); } .csharpcode, .csharpcode pre { font-size: small; color: black; font-family: consolas, "Courier New", courier, monospace; background-color: #ffffff; /*white-space: pre;*/ } .csharpcode pre { margin: 0em; } .csharpcode .rem { color: #008000; } .csharpcode .kwrd { color: #0000ff; } .csharpcode .str { color: #006080; } .csharpcode .op { color: #0000c0; } .csharpcode .preproc { color: #cc6633; } .csharpcode .asp { background-color: #ffff00; } .csharpcode .html { color: #800000; } .csharpcode .attr { color: #ff0000; } .csharpcode .alt { background-color: #f4f4f4; width: 100%; margin: 0em; } .csharpcode .lnum { color: #606060; } This seems like a good candidate for parallelization, but there is a problem here.  If we just wrap this into a call to Parallel.ForEach, we’ll introduce a critical race condition, and get the wrong answer.  Let’s look at what happens here: // Buggy code! Do not use! double min = double.MaxValue; Parallel.ForEach(collection, item => { double value = item.PerformComputation(); min = System.Math.Min(min, value); }); This code has a fatal flaw: min will be checked, then set, by multiple threads simultaneously.  Two threads may perform the check at the same time, and set the wrong value for min.  Say we get a value of 1 in thread 1, and a value of 2 in thread 2, and these two elements are the first two to run.  If both hit the min check line at the same time, both will determine that min should change, to 1 and 2 respectively.  If element 1 happens to set the variable first, then element 2 sets the min variable, we’ll detect a min value of 2 instead of 1.  This can lead to wrong answers. Unfortunately, fixing this, with the Parallel.ForEach call we’re using, would require adding locking.  We would need to rewrite this like: // Safe, but slow double min = double.MaxValue; // Make a "lock" object object syncObject = new object(); Parallel.ForEach(collection, item => { double value = item.PerformComputation(); lock(syncObject) min = System.Math.Min(min, value); }); This will potentially add a huge amount of overhead to our calculation.  Since we can potentially block while waiting on the lock for every single iteration, we will most likely slow this down to where it is actually quite a bit slower than our serial implementation.  The problem is the lock statement – any time you use lock(object), you’re almost assuring reduced performance in a parallel situation.  This leads to two observations I’ll make: When parallelizing a routine, try to avoid locks. That being said: Always add any and all required synchronization to avoid race conditions. These two observations tend to be opposing forces – we often need to synchronize our algorithms, but we also want to avoid the synchronization when possible.  Looking at our routine, there is no way to directly avoid this lock, since each element is potentially being run on a separate thread, and this lock is necessary in order for our routine to function correctly every time. However, this isn’t the only way to design this routine to implement this algorithm.  Realize that, although our collection may have thousands or even millions of elements, we have a limited number of Processing Elements (PE).  Processing Element is the standard term for a hardware element which can process and execute instructions.  This typically is a core in your processor, but many modern systems have multiple hardware execution threads per core.  The Task Parallel Library will not execute the work for each item in the collection as a separate work item. Instead, when Parallel.ForEach executes, it will partition the collection into larger “chunks” which get processed on different threads via the ThreadPool.  This helps reduce the threading overhead, and help the overall speed.  In general, the Parallel class will only use one thread per PE in the system. Given the fact that there are typically fewer threads than work items, we can rethink our algorithm design.  We can parallelize our algorithm more effectively by approaching it differently.  Because the basic aggregation we are doing here (Min) is communitive, we do not need to perform this in a given order.  We knew this to be true already – otherwise, we wouldn’t have been able to parallelize this routine in the first place.  With this in mind, we can treat each thread’s work independently, allowing each thread to serially process many elements with no locking, then, after all the threads are complete, “merge” together the results. This can be accomplished via a different set of overloads in the Parallel class: Parallel.ForEach<TSource,TLocal>.  The idea behind these overloads is to allow each thread to begin by initializing some local state (TLocal).  The thread will then process an entire set of items in the source collection, providing that state to the delegate which processes an individual item.  Finally, at the end, a separate delegate is run which allows you to handle merging that local state into your final results. To rewriting our routine using Parallel.ForEach<TSource,TLocal>, we need to provide three delegates instead of one.  The most basic version of this function is declared as: public static ParallelLoopResult ForEach<TSource, TLocal>( IEnumerable<TSource> source, Func<TLocal> localInit, Func<TSource, ParallelLoopState, TLocal, TLocal> body, Action<TLocal> localFinally ) The first delegate (the localInit argument) is defined as Func<TLocal>.  This delegate initializes our local state.  It should return some object we can use to track the results of a single thread’s operations. The second delegate (the body argument) is where our main processing occurs, although now, instead of being an Action<T>, we actually provide a Func<TSource, ParallelLoopState, TLocal, TLocal> delegate.  This delegate will receive three arguments: our original element from the collection (TSource), a ParallelLoopState which we can use for early termination, and the instance of our local state we created (TLocal).  It should do whatever processing you wish to occur per element, then return the value of the local state after processing is completed. The third delegate (the localFinally argument) is defined as Action<TLocal>.  This delegate is passed our local state after it’s been processed by all of the elements this thread will handle.  This is where you can merge your final results together.  This may require synchronization, but now, instead of synchronizing once per element (potentially millions of times), you’ll only have to synchronize once per thread, which is an ideal situation. Now that I’ve explained how this works, lets look at the code: // Safe, and fast! double min = double.MaxValue; // Make a "lock" object object syncObject = new object(); Parallel.ForEach( collection, // First, we provide a local state initialization delegate. () => double.MaxValue, // Next, we supply the body, which takes the original item, loop state, // and local state, and returns a new local state (item, loopState, localState) => { double value = item.PerformComputation(); return System.Math.Min(localState, value); }, // Finally, we provide an Action<TLocal>, to "merge" results together localState => { // This requires locking, but it's only once per used thread lock(syncObj) min = System.Math.Min(min, localState); } ); Although this is a bit more complicated than the previous version, it is now both thread-safe, and has minimal locking.  This same approach can be used by Parallel.For, although now, it’s Parallel.For<TLocal>.  When working with Parallel.For<TLocal>, you use the same triplet of delegates, with the same purpose and results. Also, many times, you can completely avoid locking by using a method of the Interlocked class to perform the final aggregation in an atomic operation.  The MSDN example demonstrating this same technique using Parallel.For uses the Interlocked class instead of a lock, since they are doing a sum operation on a long variable, which is possible via Interlocked.Add. By taking advantage of local state, we can use the Parallel class methods to parallelize algorithms such as aggregation, which, at first, may seem like poor candidates for parallelization.  Doing so requires careful consideration, and often requires a slight redesign of the algorithm, but the performance gains can be significant if handled in a way to avoid excessive synchronization.

  Read the article

• A Nondeterministic Engine written in VB.NET 2010

  - by neil chen

  When I'm reading SICP (Structure and Interpretation of Computer Programs) recently, I'm very interested in the concept of an "Nondeterministic Algorithm". According to wikipedia:  In computer science, a nondeterministic algorithm is an algorithm with one or more choice points where multiple different continuations are possible, without any specification of which one will be taken. For example, here is an puzzle came from the SICP: Baker, Cooper, Fletcher, Miller, and Smith live on different floors of an apartment housethat contains only five floors. Baker does not live on the top floor. Cooper does not live onthe bottom floor. Fletcher does not live on either the top or the bottom floor. Miller lives ona higher floor than does Cooper. Smith does not live on a floor adjacent to Fletcher's.Fletcher does not live on a floor adjacent to Cooper's. Where does everyone live? After reading this I decided to build a simple nondeterministic calculation engine with .NET. The rough idea is that we can use an iterator to track each set of possible values of the parameters, and then we implement some logic inside the engine to automate the statemachine, so that we can try one combination of the values, then test it, and then move to the next. We also used a backtracking algorithm to go back when we are running out of choices at some point. Following is the core code of the engine itself: Code highlighting produced by Actipro CodeHighlighter (freeware)http://www.CodeHighlighter.com/--Public Class NonDeterministicEngine Private _paramDict As New List(Of Tuple(Of String, IEnumerator)) 'Private _predicateDict As New List(Of Tuple(Of Func(Of Object, Boolean), IEnumerable(Of String))) Private _predicateDict As New List(Of Tuple(Of Object, IList(Of String))) Public Sub AddParam(ByVal name As String, ByVal values As IEnumerable) _paramDict.Add(New Tuple(Of String, IEnumerator)(name, values.GetEnumerator())) End Sub Public Sub AddRequire(ByVal predicate As Func(Of Object, Boolean), ByVal paramNames As IList(Of String)) CheckParamCount(1, paramNames) _predicateDict.Add(New Tuple(Of Object, IList(Of String))(predicate, paramNames)) End Sub Public Sub AddRequire(ByVal predicate As Func(Of Object, Object, Boolean), ByVal paramNames As IList(Of String)) CheckParamCount(2, paramNames) _predicateDict.Add(New Tuple(Of Object, IList(Of String))(predicate, paramNames)) End Sub Public Sub AddRequire(ByVal predicate As Func(Of Object, Object, Object, Boolean), ByVal paramNames As IList(Of String)) CheckParamCount(3, paramNames) _predicateDict.Add(New Tuple(Of Object, IList(Of String))(predicate, paramNames)) End Sub Public Sub AddRequire(ByVal predicate As Func(Of Object, Object, Object, Object, Boolean), ByVal paramNames As IList(Of String)) CheckParamCount(4, paramNames) _predicateDict.Add(New Tuple(Of Object, IList(Of String))(predicate, paramNames)) End Sub Public Sub AddRequire(ByVal predicate As Func(Of Object, Object, Object, Object, Object, Boolean), ByVal paramNames As IList(Of String)) CheckParamCount(5, paramNames) _predicateDict.Add(New Tuple(Of Object, IList(Of String))(predicate, paramNames)) End Sub Public Sub AddRequire(ByVal predicate As Func(Of Object, Object, Object, Object, Object, Object, Boolean), ByVal paramNames As IList(Of String)) CheckParamCount(6, paramNames) _predicateDict.Add(New Tuple(Of Object, IList(Of String))(predicate, paramNames)) End Sub Public Sub AddRequire(ByVal predicate As Func(Of Object, Object, Object, Object, Object, Object, Object, Boolean), ByVal paramNames As IList(Of String)) CheckParamCount(7, paramNames) _predicateDict.Add(New Tuple(Of Object, IList(Of String))(predicate, paramNames)) End Sub Public Sub AddRequire(ByVal predicate As Func(Of Object, Object, Object, Object, Object, Object, Object, Object, Boolean), ByVal paramNames As IList(Of String)) CheckParamCount(8, paramNames) _predicateDict.Add(New Tuple(Of Object, IList(Of String))(predicate, paramNames)) End Sub Sub CheckParamCount(ByVal count As Integer, ByVal paramNames As IList(Of String)) If paramNames.Count <> count Then Throw New Exception("Parameter count does not match.") End If End Sub Public Property IterationOver As Boolean Private _firstTime As Boolean = True Public ReadOnly Property Current As Dictionary(Of String, Object) Get If IterationOver Then Return Nothing Else Dim _nextResult = New Dictionary(Of String, Object) For Each item In _paramDict Dim iter = item.Item2 _nextResult.Add(item.Item1, iter.Current) Next Return _nextResult End If End Get End Property Function MoveNext() As Boolean If IterationOver Then Return False End If If _firstTime Then For Each item In _paramDict Dim iter = item.Item2 iter.MoveNext() Next _firstTime = False Return True Else Dim canMoveNext = False Dim iterIndex = _paramDict.Count - 1 canMoveNext = _paramDict(iterIndex).Item2.MoveNext If canMoveNext Then Return True End If Do While Not canMoveNext iterIndex = iterIndex - 1 If iterIndex = -1 Then Return False IterationOver = True End If canMoveNext = _paramDict(iterIndex).Item2.MoveNext If canMoveNext Then For i = iterIndex + 1 To _paramDict.Count - 1 Dim iter = _paramDict(i).Item2 iter.Reset() iter.MoveNext() Next Return True End If Loop End If End Function Function GetNextResult() As Dictionary(Of String, Object) While MoveNext() Dim result = Current If Satisfy(result) Then Return result End If End While Return Nothing End Function Function Satisfy(ByVal result As Dictionary(Of String, Object)) As Boolean For Each item In _predicateDict Dim pred = item.Item1 Select Case item.Item2.Count Case 1 Dim p1 = DirectCast(pred, Func(Of Object, Boolean)) Dim v1 = result(item.Item2(0)) If Not p1(v1) Then Return False End If Case 2 Dim p2 = DirectCast(pred, Func(Of Object, Object, Boolean)) Dim v1 = result(item.Item2(0)) Dim v2 = result(item.Item2(1)) If Not p2(v1, v2) Then Return False End If Case 3 Dim p3 = DirectCast(pred, Func(Of Object, Object, Object, Boolean)) Dim v1 = result(item.Item2(0)) Dim v2 = result(item.Item2(1)) Dim v3 = result(item.Item2(2)) If Not p3(v1, v2, v3) Then Return False End If Case 4 Dim p4 = DirectCast(pred, Func(Of Object, Object, Object, Object, Boolean)) Dim v1 = result(item.Item2(0)) Dim v2 = result(item.Item2(1)) Dim v3 = result(item.Item2(2)) Dim v4 = result(item.Item2(3)) If Not p4(v1, v2, v3, v4) Then Return False End If Case 5 Dim p5 = DirectCast(pred, Func(Of Object, Object, Object, Object, Object, Boolean)) Dim v1 = result(item.Item2(0)) Dim v2 = result(item.Item2(1)) Dim v3 = result(item.Item2(2)) Dim v4 = result(item.Item2(3)) Dim v5 = result(item.Item2(4)) If Not p5(v1, v2, v3, v4, v5) Then Return False End If Case 6 Dim p6 = DirectCast(pred, Func(Of Object, Object, Object, Object, Object, Object, Boolean)) Dim v1 = result(item.Item2(0)) Dim v2 = result(item.Item2(1)) Dim v3 = result(item.Item2(2)) Dim v4 = result(item.Item2(3)) Dim v5 = result(item.Item2(4)) Dim v6 = result(item.Item2(5)) If Not p6(v1, v2, v3, v4, v5, v6) Then Return False End If Case 7 Dim p7 = DirectCast(pred, Func(Of Object, Object, Object, Object, Object, Object, Object, Boolean)) Dim v1 = result(item.Item2(0)) Dim v2 = result(item.Item2(1)) Dim v3 = result(item.Item2(2)) Dim v4 = result(item.Item2(3)) Dim v5 = result(item.Item2(4)) Dim v6 = result(item.Item2(5)) Dim v7 = result(item.Item2(6)) If Not p7(v1, v2, v3, v4, v5, v6, v7) Then Return False End If Case 8 Dim p8 = DirectCast(pred, Func(Of Object, Object, Object, Object, Object, Object, Object, Object, Boolean)) Dim v1 = result(item.Item2(0)) Dim v2 = result(item.Item2(1)) Dim v3 = result(item.Item2(2)) Dim v4 = result(item.Item2(3)) Dim v5 = result(item.Item2(4)) Dim v6 = result(item.Item2(5)) Dim v7 = result(item.Item2(6)) Dim v8 = result(item.Item2(7)) If Not p8(v1, v2, v3, v4, v5, v6, v7, v8) Then Return False End If Case Else Throw New NotSupportedException End Select Next Return True End FunctionEnd Class    And now we can use the engine to solve the problem we mentioned above:   Code highlighting produced by Actipro CodeHighlighter (freeware)http://www.CodeHighlighter.com/--Sub Test2() Dim engine = New NonDeterministicEngine() engine.AddParam("baker", {1, 2, 3, 4, 5}) engine.AddParam("cooper", {1, 2, 3, 4, 5}) engine.AddParam("fletcher", {1, 2, 3, 4, 5}) engine.AddParam("miller", {1, 2, 3, 4, 5}) engine.AddParam("smith", {1, 2, 3, 4, 5}) engine.AddRequire(Function(baker) As Boolean Return baker <> 5 End Function, {"baker"}) engine.AddRequire(Function(cooper) As Boolean Return cooper <> 1 End Function, {"cooper"}) engine.AddRequire(Function(fletcher) As Boolean Return fletcher <> 1 And fletcher <> 5 End Function, {"fletcher"}) engine.AddRequire(Function(miller, cooper) As Boolean 'Return miller = cooper + 1 Return miller > cooper End Function, {"miller", "cooper"}) engine.AddRequire(Function(smith, fletcher) As Boolean Return smith <> fletcher + 1 And smith <> fletcher - 1 End Function, {"smith", "fletcher"}) engine.AddRequire(Function(fletcher, cooper) As Boolean Return fletcher <> cooper + 1 And fletcher <> cooper - 1 End Function, {"fletcher", "cooper"}) engine.AddRequire(Function(a, b, c, d, e) As Boolean Return a <> b And a <> c And a <> d And a <> e And b <> c And b <> d And b <> e And c <> d And c <> e And d <> e End Function, {"baker", "cooper", "fletcher", "miller", "smith"}) Dim result = engine.GetNextResult() While Not result Is Nothing Console.WriteLine(String.Format("baker: {0}, cooper: {1}, fletcher: {2}, miller: {3}, smith: {4}", result("baker"), result("cooper"), result("fletcher"), result("miller"), result("smith"))) result = engine.GetNextResult() End While Console.WriteLine("Calculation ended.")End Sub   Also, this engine can solve the classic 8 queens puzzle and find out all 92 results for me.   Code highlighting produced by Actipro CodeHighlighter (freeware)http://www.CodeHighlighter.com/--Sub Test3() ' The 8-Queens problem. Dim engine = New NonDeterministicEngine() ' Let's assume that a - h represents the queens in row 1 to 8, then we just need to find out the column number for each of them. engine.AddParam("a", {1, 2, 3, 4, 5, 6, 7, 8}) engine.AddParam("b", {1, 2, 3, 4, 5, 6, 7, 8}) engine.AddParam("c", {1, 2, 3, 4, 5, 6, 7, 8}) engine.AddParam("d", {1, 2, 3, 4, 5, 6, 7, 8}) engine.AddParam("e", {1, 2, 3, 4, 5, 6, 7, 8}) engine.AddParam("f", {1, 2, 3, 4, 5, 6, 7, 8}) engine.AddParam("g", {1, 2, 3, 4, 5, 6, 7, 8}) engine.AddParam("h", {1, 2, 3, 4, 5, 6, 7, 8}) Dim NotInTheSameDiagonalLine = Function(cols As IList) As Boolean For i = 0 To cols.Count - 2 For j = i + 1 To cols.Count - 1 If j - i = Math.Abs(cols(j) - cols(i)) Then Return False End If Next Next Return True End Function engine.AddRequire(Function(a, b, c, d, e, f, g, h) As Boolean Return a <> b AndAlso a <> c AndAlso a <> d AndAlso a <> e AndAlso a <> f AndAlso a <> g AndAlso a <> h AndAlso b <> c AndAlso b <> d AndAlso b <> e AndAlso b <> f AndAlso b <> g AndAlso b <> h AndAlso c <> d AndAlso c <> e AndAlso c <> f AndAlso c <> g AndAlso c <> h AndAlso d <> e AndAlso d <> f AndAlso d <> g AndAlso d <> h AndAlso e <> f AndAlso e <> g AndAlso e <> h AndAlso f <> g AndAlso f <> h AndAlso g <> h AndAlso NotInTheSameDiagonalLine({a, b, c, d, e, f, g, h}) End Function, {"a", "b", "c", "d", "e", "f", "g", "h"}) Dim result = engine.GetNextResult() While Not result Is Nothing Console.WriteLine("(1,{0}), (2,{1}), (3,{2}), (4,{3}), (5,{4}), (6,{5}), (7,{6}), (8,{7})", result("a"), result("b"), result("c"), result("d"), result("e"), result("f"), result("g"), result("h")) result = engine.GetNextResult() End While Console.WriteLine("Calculation ended.")End Sub (Chinese version of the post: http://www.cnblogs.com/RChen/archive/2010/05/17/1737587.html) Cheers,  

  Read the article

• Does a site's bounce rate influence Google rankings?

  - by Joel Spolsky

  Does Google consider bounce rate or something similar in ranking sites? Background: here at Stack Exchange we noticed that the latest Google algorithm changes resulted in about a 20% dip in traffic to Server Fault (and a much smaller dip in traffic to Super User). Stack Overflow traffic was not affected. There was an article on WebProNews which hypothesized that bounce rate might be a ranking signal in Google's latest Panda update. According to Google Analytics, these are our bounce rates over the last month: Site Bounce Rate Avg Time on Site ------------- ----------- ---------------- SuperUser 84.67% 01:16 ServerFault 83.76% 00:53 Stack Overflow 63.63% 04:12 Now, technically, Google has no way to know the bounce rate. If you go to Google, search for something, and click on the first result, Google can't tell the difference between: a user who turns off their computer a user who goes to a completely different web site a user who spends hours clicking around on the website they landed on What Google does know is how long it takes the user to come back to Google and do another search. According to the book In The Plex (page 47), Google distinguishes between what they call "short clicks" and "long clicks": A short click is a search where the user quickly comes back to Google and does another search. Google interprets this as a signal that the first search results were unsatisfactory. A long click is a search where the user doesn't search again for a long time. The book says that Google uses this information internally, to judge the quality of their own algorithms. It also said that short click data in which someone retypes a slight variation of the search is used to fuel the "Did you mean...?" spell checking algorithm. So, my hypothesis is that Google has recently decided to use long click rates as a signal of a high quality site. Does anyone have any evidence of this? Have you seen any high-bounce-rate sites which lost traffic (or vice-versa)?

  Read the article

• Need some advice regarding collision detection with the sprite changing its width and height

  - by Frank Scott

  So I'm messing around with collision detection in my tile-based game and everything works fine and dandy using this method. However, now I am trying to implement sprite sheets so my character can have a walking and jumping animation. For one, I'd like to to be able to have each frame of variable size, I think. I want collision detection to be accurate and during a jumping animation the sprite's height will be shorter (because of the calves meeting the hamstrings). Again, this also works fine at the moment. I can get the character to animate properly each frame and cycle through animations. The problems arise when the width and height of the character change. Often times its position will be corrected by the collision detection system and the character will be rubber-banded to random parts of the map or even go outside the map bounds. For some reason with the linked collision detection algorithm, when the width or height of the sprite is changed on the fly, the entire algorithm breaks down. The solution I found so far is to have a single width and height of the sprite that remains constant, and only adjust the source rectangle for drawing. However, I'm not sure exactly what to set as the sprite's constant bounding box because it varies so much with the different animations. So now I'm not sure what to do. I'm toying with the idea of pixel-perfect collision detection but I'm not sure if it would really be worth it. Does anyone know how Braid does their collision detection? My game is also a 2D sidescroller and I was quite impressed with how it was handled in that game. Thanks for reading.

  Read the article

• Optimizing Solaris 11 SHA-1 on Intel Processors

  - by danx

  SHA-1 is a "hash" or "digest" operation that produces a 160 bit (20 byte) checksum value on arbitrary data, such as a file. It is intended to uniquely identify text and to verify it hasn't been modified. Max Locktyukhin and others at Intel have improved the performance of the SHA-1 digest algorithm using multiple techniques. This code has been incorporated into Solaris 11 and is available in the Solaris Crypto Framework via the libmd(3LIB), the industry-standard libpkcs11(3LIB) library, and Solaris kernel module sha1. The optimized code is used automatically on systems with a x86 CPU supporting SSSE3 (Intel Supplemental SSSE3). Intel microprocessor architectures that support SSSE3 include Nehalem, Westmere, Sandy Bridge microprocessor families. Further optimizations are available for microprocessors that support AVX (such as Sandy Bridge). Although SHA-1 is considered obsolete because of weaknesses found in the SHA-1 algorithm—NIST recommends using at least SHA-256, SHA-1 is still widely used and will be with us for awhile more. Collisions (the same SHA-1 result for two different inputs) can be found with moderate effort. SHA-1 is used heavily though in SSL/TLS, for example. And SHA-1 is stronger than the older MD5 digest algorithm, another digest option defined in SSL/TLS. Optimizations Review SHA-1 operates by reading an arbitrary amount of data. The data is read in 512 bit (64 byte) blocks (the last block is padded in a specific way to ensure it's a full 64 bytes). Each 64 byte block has 80 "rounds" of calculations (consisting of a mixture of "ROTATE-LEFT", "AND", and "XOR") applied to the block. Each round produces a 32-bit intermediate result, called W[i]. Here's what each round operates: The first 16 rounds, rounds 0 to 15, read the 512 bit block 32 bits at-a-time. These 32 bits is used as input to the round. The remaining rounds, rounds 16 to 79, use the results from the previous rounds as input. Specifically for round i it XORs the results of rounds i-3, i-8, i-14, and i-16 and rotates the result left 1 bit. The remaining calculations for the round is a series of AND, XOR, and ROTATE-LEFT operators on the 32-bit input and some constants. The 32-bit result is saved as W[i] for round i. The 32-bit result of the final round, W[79], is the SHA-1 checksum. Optimization: Vectorization The first 16 rounds can be vectorized (computed in parallel) because they don't depend on the output of a previous round. As for the remaining rounds, because of step 2 above, computing round i depends on the results of round i-3, W[i-3], one can vectorize 3 rounds at-a-time. Max Locktyukhin found through simple factoring, explained in detail in his article referenced below, that the dependencies of round i on the results of rounds i-3, i-8, i-14, and i-16 can be replaced instead with dependencies on the results of rounds i-6, i-16, i-28, and i-32. That is, instead of initializing intermediate result W[i] with: W[i] = (W[i-3] XOR W[i-8] XOR W[i-14] XOR W[i-16]) ROTATE-LEFT 1 Initialize W[i] as follows: W[i] = (W[i-6] XOR W[i-16] XOR W[i-28] XOR W[i-32]) ROTATE-LEFT 2 That means that 6 rounds could be vectorized at once, with no additional calculations, instead of just 3! This optimization is independent of Intel or any other microprocessor architecture, although the microprocessor has to support vectorization to use it, and exploits one of the weaknesses of SHA-1. Optimization: SSSE3 Intel SSSE3 makes use of 16 %xmm registers, each 128 bits wide. The 4 32-bit inputs to a round, W[i-6], W[i-16], W[i-28], W[i-32], all fit in one %xmm register. The following code snippet, from Max Locktyukhin's article, converted to ATT assembly syntax, computes 4 rounds in parallel with just a dozen or so SSSE3 instructions: movdqa W_minus_04, W_TMP pxor W_minus_28, W // W equals W[i-32:i-29] before XOR // W = W[i-32:i-29] ^ W[i-28:i-25] palignr $8, W_minus_08, W_TMP // W_TMP = W[i-6:i-3], combined from // W[i-4:i-1] and W[i-8:i-5] vectors pxor W_minus_16, W // W = (W[i-32:i-29] ^ W[i-28:i-25]) ^ W[i-16:i-13] pxor W_TMP, W // W = (W[i-32:i-29] ^ W[i-28:i-25] ^ W[i-16:i-13]) ^ W[i-6:i-3]) movdqa W, W_TMP // 4 dwords in W are rotated left by 2 psrld $30, W // rotate left by 2 W = (W >> 30) | (W << 2) pslld $2, W_TMP por W, W_TMP movdqa W_TMP, W // four new W values W[i:i+3] are now calculated paddd (K_XMM), W_TMP // adding 4 current round's values of K movdqa W_TMP, (WK(i)) // storing for downstream GPR instructions to read A window of the 32 previous results, W[i-1] to W[i-32] is saved in memory on the stack. This is best illustrated with a chart. Without vectorization, computing the rounds is like this (each "R" represents 1 round of SHA-1 computation): RRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRR With vectorization, 4 rounds can be computed in parallel: RRRRRRRRRRRRRRRRRRRR RRRRRRRRRRRRRRRRRRRR RRRRRRRRRRRRRRRRRRRR RRRRRRRRRRRRRRRRRRRR Optimization: AVX The new "Sandy Bridge" microprocessor architecture, which supports AVX, allows another interesting optimization. SSSE3 instructions have two operands, a input and an output. AVX allows three operands, two inputs and an output. In many cases two SSSE3 instructions can be combined into one AVX instruction. The difference is best illustrated with an example. Consider these two instructions from the snippet above: pxor W_minus_16, W // W = (W[i-32:i-29] ^ W[i-28:i-25]) ^ W[i-16:i-13] pxor W_TMP, W // W = (W[i-32:i-29] ^ W[i-28:i-25] ^ W[i-16:i-13]) ^ W[i-6:i-3]) With AVX they can be combined in one instruction: vpxor W_minus_16, W, W_TMP // W = (W[i-32:i-29] ^ W[i-28:i-25] ^ W[i-16:i-13]) ^ W[i-6:i-3]) This optimization is also in Solaris, although Sandy Bridge-based systems aren't widely available yet. As an exercise for the reader, AVX also has 256-bit media registers, %ymm0 - %ymm15 (a superset of 128-bit %xmm0 - %xmm15). Can %ymm registers be used to parallelize the code even more? Optimization: Solaris-specific In addition to using the Intel code described above, I performed other minor optimizations to the Solaris SHA-1 code: Increased the digest(1) and mac(1) command's buffer size from 4K to 64K, as previously done for decrypt(1) and encrypt(1). This size is well suited for ZFS file systems, but helps for other file systems as well. Optimized encode functions, which byte swap the input and output data, to copy/byte-swap 4 or 8 bytes at-a-time instead of 1 byte-at-a-time. Enhanced the Solaris mdb(1) and kmdb(1) debuggers to display all 16 %xmm and %ymm registers (mdb "$x" command). Previously they only displayed the first 8 that are available in 32-bit mode. Can't optimize if you can't debug :-). Changed the SHA-1 code to allow processing in "chunks" greater than 2 Gigabytes (64-bits) Performance I measured performance on a Sun Ultra 27 (which has a Nehalem-class Xeon 5500 Intel W3570 microprocessor @3.2GHz). Turbo mode is disabled for consistent performance measurement. Graphs are better than words and numbers, so here they are: The first graph shows the Solaris digest(1) command before and after the optimizations discussed here, contained in libmd(3LIB). I ran the digest command on a half GByte file in swapfs (/tmp) and execution time decreased from 1.35 seconds to 0.98 seconds. The second graph shows the the results of an internal microbenchmark that uses the Solaris libpkcs11(3LIB) library. The operations are on a 128 byte buffer with 10,000 iterations. The results show operations increased from 320,000 to 416,000 operations per second. Finally the third graph shows the results of an internal kernel microbenchmark that uses the Solaris /kernel/crypto/amd64/sha1 module. The operations are on a 64Kbyte buffer with 100 iterations. third graph shows the results of an internal kernel microbenchmark that uses the Solaris /kernel/crypto/amd64/sha1 module. The operations are on a 64Kbyte buffer with 100 iterations. The results show for 1 kernel thread, operations increased from 410 to 600 MBytes/second. For 8 kernel threads, operations increase from 1540 to 1940 MBytes/second. Availability This code is in Solaris 11 FCS. It is available in the 64-bit libmd(3LIB) library for 64-bit programs and is in the Solaris kernel. You must be running hardware that supports Intel's SSSE3 instructions (for example, Intel Nehalem, Westmere, or Sandy Bridge microprocessor architectures). The easiest way to determine if SSSE3 is available is with the isainfo(1) command. For example, nehalem $ isainfo -v $ isainfo -v 64-bit amd64 applications sse4.2 sse4.1 ssse3 popcnt tscp ahf cx16 sse3 sse2 sse fxsr mmx cmov amd_sysc cx8 tsc fpu 32-bit i386 applications sse4.2 sse4.1 ssse3 popcnt tscp ahf cx16 sse3 sse2 sse fxsr mmx cmov sep cx8 tsc fpu If the output also shows "avx", the Solaris executes the even-more optimized 3-operand AVX instructions for SHA-1 mentioned above: sandybridge $ isainfo -v 64-bit amd64 applications avx xsave pclmulqdq aes sse4.2 sse4.1 ssse3 popcnt tscp ahf cx16 sse3 sse2 sse fxsr mmx cmov amd_sysc cx8 tsc fpu 32-bit i386 applications avx xsave pclmulqdq aes sse4.2 sse4.1 ssse3 popcnt tscp ahf cx16 sse3 sse2 sse fxsr mmx cmov sep cx8 tsc fpu No special configuration or setup is needed to take advantage of this code. Solaris libraries and kernel automatically determine if it's running on SSSE3 or AVX-capable machines and execute the correctly-tuned code for that microprocessor. Summary The Solaris 11 Crypto Framework, via the sha1 kernel module and libmd(3LIB) and libpkcs11(3LIB) libraries, incorporated a useful SHA-1 optimization from Intel for SSSE3-capable microprocessors. As with other Solaris optimizations, they come automatically "under the hood" with the current Solaris release. References "Improving the Performance of the Secure Hash Algorithm (SHA-1)" by Max Locktyukhin (Intel, March 2010). The source for these SHA-1 optimizations used in Solaris "SHA-1", Wikipedia Good overview of SHA-1 FIPS 180-1 SHA-1 standard (FIPS, 1995) NIST Comments on Cryptanalytic Attacks on SHA-1 (2005, revised 2006)

  Read the article

• Use depth bias for shadows in deferred shading

  - by cubrman

  We are building a deferred shading engine and we have a problem with shadows. To add shadows we use two maps: the first one stores the depth of the scene captured by the player's camera and the second one stores the depth of the scene captured by the light's camera. We then ran a shader that analyzes the two maps and outputs the third one with the ready shadow areas for the current frame. The problem we face is a classic one: Self-Shadowing: A standard way to solve this is to use the slope-scale depth bias and depth offsets, however as we are doing things in a deferred way we cannot employ this algorithm. Any attempts to set depth bias when capturing light's view depth produced no or unsatisfying results. So here is my question: MSDN article has a convoluted explanation of the slope-scale: bias = (m × SlopeScaleDepthBias) + DepthBias Where m is the maximum depth slope of the triangle being rendered, defined as: m = max( abs(delta z / delta x), abs(delta z / delta y) ) Could you explain how I can implement this algorithm manually in a shader? Maybe there are better ways to fix this problem for deferred shadows?

  Read the article

• Oracle Solaris 11 ZFS Lab for Openworld 2012

  - by user12626122

  Preface This is the content from the Oracle Openworld 2012 ZFS lab. It was well attended - the feedback was that it was a little short - thats probably because in writing it I bacame very time-concious after the ASM/ACFS on Solaris extravaganza I ran last year which was almost too long for mortal man to finish in the 1 hour session. Enjoy. Table of Contents Exercise Z.1: ZFS Pools Exercise Z.2: ZFS File Systems Exercise Z.3: ZFS Compression Exercise Z.4: ZFS Deduplication Exercise Z.5: ZFS Encryption Exercise Z.6: Solaris 11 Shadow Migration Introduction This set of exercises is designed to briefly demonstrate new features in Solaris 11 ZFS file system: Deduplication, Encryption and Shadow Migration. Also included is the creation of zpools and zfs file systems - the basic building blocks of the technology, and also Compression which is the compliment of Deduplication. The exercises are just introductions - you are referred to the ZFS Adminstration Manual for further information. From Solaris 11 onward the online manual pages consist of zpool(1M) and zfs(1M) with further feature-specific information in zfs_allow(1M), zfs_encrypt(1M) and zfs_share(1M). The lab is easily carried out in a VirtualBox running Solaris 11 with 6 virtual 3 Gb disks to play with. Exercise Z.1: ZFS Pools Task: You have several disks to use for your new file system. Create a new zpool and a file system within it. Lab: You will check the status of existing zpools, create your own pool and expand it. Your Solaris 11 installation already has a root ZFS pool. It contains the root file system. Check this: root@solaris:~# zpool list NAME SIZE ALLOC FREE CAP DEDUP HEALTH ALTROOT rpool 15.9G 6.62G 9.25G 41% 1.00x ONLINE - root@solaris:~# zpool status pool: rpool state: ONLINE scan: none requested config: NAME STATE READ WRITE CKSUM rpool ONLINE 0 0 0 c3t0d0s0 ONLINE 0 0 0 errors: No known data errors Note the disk device the root pool is on - c3t0d0s0 Now you will create your own ZFS pool. First you will check what disks are available: root@solaris:~# echo | format Searching for disks...done AVAILABLE DISK SELECTIONS: 0. c3t0d0 <ATA-VBOX HARDDISK-1.0 cyl 2085 alt 2 hd 255 sec 63> /pci@0,0/pci8086,2829@d/disk@0,0 1. c3t2d0 <ATA-VBOX HARDDISK-1.0 cyl 1534 alt 2 hd 128 sec 32> /pci@0,0/pci8086,2829@d/disk@2,0 2. c3t3d0 <ATA-VBOX HARDDISK-1.0 cyl 1534 alt 2 hd 128 sec 32> /pci@0,0/pci8086,2829@d/disk@3,0 3. c3t4d0 <ATA-VBOX HARDDISK-1.0 cyl 1534 alt 2 hd 128 sec 32> /pci@0,0/pci8086,2829@d/disk@4,0 4. c3t5d0 <ATA-VBOX HARDDISK-1.0 cyl 1534 alt 2 hd 128 sec 32> /pci@0,0/pci8086,2829@d/disk@5,0 5. c3t6d0 <ATA-VBOX HARDDISK-1.0 cyl 1534 alt 2 hd 128 sec 32> /pci@0,0/pci8086,2829@d/disk@6,0 6. c3t7d0 <ATA-VBOX HARDDISK-1.0 cyl 1534 alt 2 hd 128 sec 32> /pci@0,0/pci8086,2829@d/disk@7,0 Specify disk (enter its number): Specify disk (enter its number): The root disk is numbered 0. The others are free for use. Try creating a simple pool and observe the error message: root@solaris:~# zpool create mypool c3t2d0 c3t3d0 'mypool' successfully created, but with no redundancy; failure of one device will cause loss of the pool So destroy that pool and create a mirrored pool instead: root@solaris:~# zpool destroy mypool root@solaris:~# zpool create mypool mirror c3t2d0 c3t3d0 root@solaris:~# zpool status mypool pool: mypool state: ONLINE scan: none requested config: NAME STATE READ WRITE CKSUM mypool ONLINE 0 0 0 mirror-0 ONLINE 0 0 0 c3t2d0 ONLINE 0 0 0 c3t3d0 ONLINE 0 0 0 errors: No known data errors Back to topExercise Z.2: ZFS File Systems Task: You have to create file systems for later exercises. You can see that when a pool is created, a file system of the same name is created: root@solaris:~# zfs list NAME USED AVAIL REFER MOUNTPOINT mypool 86.5K 2.94G 31K /mypool Create your filesystems and mountpoints as follows: root@solaris:~# zfs create -o mountpoint=/data1 mypool/mydata1 The -o option sets the mount point and automatically creates the necessary directory. root@solaris:~# zfs list mypool/mydata1 NAME USED AVAIL REFER MOUNTPOINT mypool/mydata1 31K 2.94G 31K /data1 Back to top Exercise Z.3: ZFS Compression Task:Try out different forms of compression available in ZFS Lab:Create 2nd filesystem with compression, fill both file systems with the same data, observe results You can see from the zfs(1) manual page that there are several types of compression available to you, set with the property=value syntax: compression=on | off | lzjb | gzip | gzip-N | zle Controls the compression algorithm used for this dataset. The lzjb compression algorithm is optimized for performance while providing decent data compression. Setting compression to on uses the lzjb compression algorithm. The gzip compression algorithm uses the same compression as the gzip(1) command. You can specify the gzip level by using the value gzip-N where N is an integer from 1 (fastest) to 9 (best compression ratio). Currently, gzip is equivalent to gzip-6 (which is also the default for gzip(1)). Create a second filesystem with compression turned on. Note how you set and get your values separately: root@solaris:~# zfs create -o mountpoint=/data2 mypool/mydata2 root@solaris:~# zfs set compression=gzip-9 mypool/mydata2 root@solaris:~# zfs get compression mypool/mydata1 NAME PROPERTY VALUE SOURCE mypool/mydata1 compression off default root@solaris:~# zfs get compression mypool/mydata2 NAME PROPERTY VALUE SOURCE mypool/mydata2 compression gzip-9 local Now you can copy the contents of /usr/lib into both your normal and compressing filesystem and observe the results. Don't forget the dot or period (".") in the find(1) command below: root@solaris:~# cd /usr/lib root@solaris:/usr/lib# find . -print | cpio -pdv /data1 root@solaris:/usr/lib# find . -print | cpio -pdv /data2 The copy into the compressing file system takes longer - as it has to perform the compression but the results show the effect: root@solaris:/usr/lib# zfs list NAME USED AVAIL REFER MOUNTPOINT mypool 1.35G 1.59G 31K /mypool mypool/mydata1 1.01G 1.59G 1.01G /data1 mypool/mydata2 341M 1.59G 341M /data2 Note that the available space in the pool is shared amongst the file systems. This behavior can be modified using quotas and reservations which are not covered in this lab but are covered extensively in the ZFS Administrators Guide. Back to top Exercise Z.4: ZFS Deduplication The deduplication property is used to remove redundant data from a ZFS file system. With the property enabled duplicate data blocks are removed synchronously. The result is that only unique data is stored and common componenents are shared. Task:See how to implement deduplication and its effects Lab: You will create a ZFS file system with deduplication turned on and see if it reduces the amount of physical storage needed when we again fill it with a copy of /usr/lib. root@solaris:/usr/lib# zfs destroy mypool/mydata2 root@solaris:/usr/lib# zfs set dedup=on mypool/mydata1 root@solaris:/usr/lib# rm -rf /data1/* root@solaris:/usr/lib# mkdir /data1/2nd-copy root@solaris:/usr/lib# zfs list NAME USED AVAIL REFER MOUNTPOINT mypool 1.02M 2.94G 31K /mypool mypool/mydata1 43K 2.94G 43K /data1 root@solaris:/usr/lib# find . -print | cpio -pd /data1 2142768 blocks root@solaris:/usr/lib# zfs list NAME USED AVAIL REFER MOUNTPOINT mypool 1.02G 1.99G 31K /mypool mypool/mydata1 1.01G 1.99G 1.01G /data1 root@solaris:/usr/lib# find . -print | cpio -pd /data1/2nd-copy 2142768 blocks root@solaris:/usr/lib#zfs list NAME USED AVAIL REFER MOUNTPOINT mypool 1.99G 1.96G 31K /mypool mypool/mydata1 1.98G 1.96G 1.98G /data1 You could go on creating copies for quite a while...but you get the idea. Note that deduplication and compression can be combined: the compression acts on metadata. Deduplication works across file systems in a pool and there is a zpool-wide property dedupratio: root@solaris:/usr/lib# zpool get dedupratio mypool NAME PROPERTY VALUE SOURCE mypool dedupratio 4.30x - Deduplication can also be checked using "zpool list": root@solaris:/usr/lib# zpool list NAME SIZE ALLOC FREE CAP DEDUP HEALTH ALTROOT mypool 2.98G 1001M 2.01G 32% 4.30x ONLINE - rpool 15.9G 6.66G 9.21G 41% 1.00x ONLINE - Before moving on to the next topic, destroy that dataset and free up some space: root@solaris:~# zfs destroy mypool/mydata1 Back to top Exercise Z.5: ZFS Encryption Task: Encrypt sensitive data. Lab: Explore basic ZFS encryption. This lab only covers the basics of ZFS Encryption. In particular it does not cover various aspects of key management. Please see the ZFS Adminastrion Manual and the zfs_encrypt(1M) manual page for more detail on this functionality. Back to top root@solaris:~# zfs create -o encryption=on mypool/data2 Enter passphrase for 'mypool/data2': ******** Enter again: ******** root@solaris:~# Creation of a descendent dataset shows that encryption is inherited from the parent: root@solaris:~# zfs create mypool/data2/data3 root@solaris:~# zfs get -r encryption,keysource,keystatus,checksum mypool/data2 NAME PROPERTY VALUE SOURCE mypool/data2 encryption on local mypool/data2 keysource passphrase,prompt local mypool/data2 keystatus available - mypool/data2 checksum sha256-mac local mypool/data2/data3 encryption on inherited from mypool/data2 mypool/data2/data3 keysource passphrase,prompt inherited from mypool/data2 mypool/data2/data3 keystatus available - mypool/data2/data3 checksum sha256-mac inherited from mypool/data2 You will find the online manual page zfs_encrypt(1M) contains examples. In particular, if time permits during this lab session you may wish to explore the changing of a key using "zfs key -c mypool/data2". Exercise Z.6: Shadow Migration Shadow Migration allows you to migrate data from an old file system to a new file system while simultaneously allowing access and modification to the new file system during the process. You can use Shadow Migration to migrate a local or remote UFS or ZFS file system to a local file system. Task: You wish to migrate data from one file system (UFS, ZFS, VxFS) to ZFS while mainaining access to it. Lab: Create the infrastructure for shadow migration and transfer one file system into another. First create the file system you want to migrate root@solaris:~# zpool create oldstuff c3t4d0 root@solaris:~# zfs create oldstuff/forgotten Then populate it with some files: root@solaris:~# cd /var/adm root@solaris:/var/adm# find . -print | cpio -pdv /oldstuff/forgotten You need the shadow-migration package installed: root@solaris:~# pkg install shadow-migration Packages to install: 1 Create boot environment: No Create backup boot environment: No Services to change: 1 DOWNLOAD PKGS FILES XFER (MB) Completed 1/1 14/14 0.2/0.2 PHASE ACTIONS Install Phase 39/39 PHASE ITEMS Package State Update Phase 1/1 Image State Update Phase 2/2 You then enable the shadowd service: root@solaris:~# svcadm enable shadowd root@solaris:~# svcs shadowd STATE STIME FMRI online 7:16:09 svc:/system/filesystem/shadowd:default Set the filesystem to be migrated to read-only root@solaris:~# zfs set readonly=on oldstuff/forgotten Create a new zfs file system with the shadow property set to the file system to be migrated: root@solaris:~# zfs create -o shadow=file:///oldstuff/forgotten mypool/remembered Use the shadowstat(1M) command to see the progress of the migration: root@solaris:~# shadowstat EST BYTES BYTES ELAPSED DATASET XFRD LEFT ERRORS TIME mypool/remembered 92.5M - - 00:00:59 mypool/remembered 99.1M 302M - 00:01:09 mypool/remembered 109M 260M - 00:01:19 mypool/remembered 133M 304M - 00:01:29 mypool/remembered 149M 339M - 00:01:39 mypool/remembered 156M 86.4M - 00:01:49 mypool/remembered 156M 8E 29 (completed) Note that if you had created /mypool/remembered as encrypted, this would be the preferred method of encrypting existing data. Similarly for compressing or deduplicating existing data. The procedure for migrating a file system over NFS is similar - see the ZFS Administration manual. That concludes this lab session.

  Read the article

• Why are SW engineering interviews disproportionately difficult?

  - by stackoverflowuser2010

  First, some background on me. I have a PhD in CS and have had jobs both as a software engineer and as an R&D research scientist, both at Very Large Corporations You Know Very Well. I recently changed jobs and interviewed for both types of jobs (as I have done in the past). My observation: SW engineer job interviews are way, way disproportionately more difficult than CS researcher job interviews, but the researcher job is higher paying, more competitive, more rewarding, more interesting, and has a higher upside. Here's a typical interview loop for researcher: Phone interview to see if my research is in alignment with the lab's researcher In-person, give presentation on my recent research for one hour (which represents maybe 9 month's worth of work), answer questions In-person one-on-one interviews with about 5 researchers, where they ask me very reasonable questions on my work/publications/patents, including: technical questions, where my work fits into related work, and how I can extend my work to new areas Here's a typical interview loop for SW engineer: Phone interview where I'm asked algorithm questions and maybe do some coding. Pretty standard. In-person interviews at the whiteboard where they drill the F*** out of you on esoteric C++ minutia (e.g. how does a polymorphic virtual function call work), algorithms (make all-pairs-shortest-path algorithm work for 1B vertices), system design (design a database load balancer), etc. This goes on for six or seven interviews. Ridiculous. Why would anyone be willing to put up with this? What is the point of asking about C++ trivia or writing code to prove yourself? Why not make the SE interview more like the researcher interview where you give a talk about what you've done? How are technical job interviews for other fields, like physics, chemistry, civil engineering, mechanical engineering?

  Read the article

• Distribute Sort Sample Service

  - by kaleidoscope

  How it works? Using the front-end of the service, a user can specify a size in MB for the input data set to sort. Algorithm CreateAndSplit The CreateAndSplit task generates the input data and stores them as 10 blobs in the utility storage. The URLs to these blobs are packaged as Separate work items and written to the queue. · Separate The Separate task reads the blobs with the random numbers created in the CreateAndSplit task and places the random numbers into buckets. The interval of the numbers that go into one bucket is chosen so that the expected amount of numbers (assuming a uniform distribution of the numbers in the original data set) is around 100 kB. Each bucket is represented as a blob container in utility storage. Whenever there are 10 blobs in one bucket (i.e., the placement in this bucket is complete because we had 10 original splits), the separate task will generate a new Sort task and write the task into the queue. · Sort The Sort task merges all blobs in a single bucket and sorts them using a standard sort algorithm. The result is stored as a blob in utility storage. · Concat The concat task merges the results of all Sort tasks into a single blob. This blob can be downloaded as a text file using this Web page. As the resulting file is presented in text format, the size of the file is likely to be larger than the specified input file. Anish

  Read the article

• Stereo images rectification and disparity: which algorithms?

  - by alessandro.francesconi

  I'm trying to figure out what are currently the two most efficent algorithms that permit, starting from a L/R pair of stereo images created using a traditional camera (so affected by some epipolar lines misalignment), to produce a pair of adjusted images plus their depth information by looking at their disparity. Actually I've found lots of papers about these two methods, like: "Computing Rectifying Homographies for Stereo Vision" (Zhang - seems one of the best for rectification only) "Three-step image recti?cation" (Monasse) "Rectification and Disparity" (slideshow by Navab) "A fast area-based stereo matching algorithm" (Di Stefano - seems a bit inaccurate) "Computing Visual Correspondence with Occlusions via Graph Cuts" (Kolmogorov - this one produces a very good disparity map, with also occlusion informations, but is it efficient?) "Dense Disparity Map Estimation Respecting Image Discontinuities" (Alvarez - toooo long for a first review) Anyone could please give me some advices for orienting into this wide topic? What kind of algorithm/method should I treat first, considering that I'll work on a very simple input: a pair of left and right images and nothing else, no more information (some papers are based on additional, pre-taken, calibration infos)? Speaking about working implementations, the only interesting results I've seen so far belongs to this piece of software, but only for automatic rectification, not disparity: http://stereo.jpn.org/eng/stphmkr/index.html I tried the "auto-adjustment" feature and seems really effective. Too bad there is no source code...

  Read the article

• Functional programming and stateful algorithms

  - by bigstones

  I'm learning functional programming with Haskell. In the meantime I'm studying Automata theory and as the two seem to fit well together I'm writing a small library to play with automata. Here's the problem that made me ask the question. While studying a way to evaluate a state's reachability I got the idea that a simple recursive algorithm would be quite inefficient, because some paths might share some states and I might end up evaluating them more than once. For example, here, evaluating reachability of g from a, I'd have to exclude f both while checking the path through d and c: So my idea is that an algorithm working in parallel on many paths and updating a shared record of excluded states might be great, but that's too much for me. I've seen that in some simple recursion cases one can pass state as an argument, and that's what I have to do here, because I pass forward the list of states I've gone through to avoid loops. But is there a way to pass that list also backwards, like returning it in a tuple together with the boolean result of my canReach function? (although this feels a bit forced) Besides the validity of my example case, what other techniques are available to solve this kind of problems? I feel like these must be common enough that there have to be solutions like what happens with fold* or map. So far, reading learnyouahaskell.com I didn't find any, but consider I haven't touched monads yet. (if interested, I posted my code on codereview)

  Read the article

• How do graphics programmers deal with rendering vertices that don't change the image?

  - by canisrufus

  So, the title is a little awkward. I'll give some background, and then ask my question. Background: I work as a web GIS application developer, but in my spare time I've been playing with map rendering and improving data interchange formats. I work only in 2D space. One interesting issue I've encountered is that when you're rendering a polygon at a small scale (zoomed way out), many of the vertices are redundant. An extreme case would be that you have a polygon with 500,000 vertices that only takes up a single pixel. If you're sending this data to the browser, it would make sense to omit ~499,999 of those vertices. One way we achieve that is by rendering an image on a server and and sending it as a PNG: voila, it's a point. Sometimes, though, we want data sent to the browser where it can be rendered with SVG (or canvas, or webgl) so that it can be interactive. The problem: It turns out that, using modern geographic data sets, it's very easy to overload SVG's rendering abilities. In an effort to cope with those limitations, I'm trying to figure out how to visually losslessly reduce a data set for a given scale and map extent (and, if necessary, for a known map pixel width and height). I got a great reduction in data size just using the Douglas-Peucker algorithm, and I believe I was able to get it to keep the polygons true to within one pixel. Unfortunately, Douglas-Peucker doesn't preserve topology, so it changed how borders between polygons got rendered. I couldn't readily find other algorithms to try out and adapt to the purpose, but I don't have much CS/algorithm background and might not recognize them if I saw them.

  Read the article

• Which algorithms/data structures should I "recognize" and know by name?

  - by Earlz

  I'd like to consider myself a fairly experienced programmer. I've been programming for over 5 years now. My weak point though is terminology. I'm self-taught, so while I know how to program, I don't know some of the more formal aspects of computer science. So, what are practical algorithms/data structures that I could recognize and know by name? Note, I'm not asking for a book recommendation about implementing algorithms. I don't care about implementing them, I just want to be able to recognize when an algorithm/data structure would be a good solution to a problem. I'm asking more for a list of algorithms/data structures that I should "recognize". For instance, I know the solution to a problem like this: You manage a set of lockers labeled 0-999. People come to you to rent the locker and then come back to return the locker key. How would you build a piece of software to manage knowing which lockers are free and which are in used? The solution, would be a queue or stack. What I'm looking for are things like "in what situation should a B-Tree be used -- What search algorithm should be used here" etc. And maybe a quick introduction of how the more complex(but commonly used) data structures/algorithms work. I tried looking at Wikipedia's list of data structures and algorithms but I think that's a bit overkill. So I'm looking more for what are the essential things I should recognize?

  Read the article

• How to improve Algorithmic Programming Solving skill? [closed]

  - by gaurav

  Possible Duplicate: How can I improve my problem-solving ability? How do you improve your problem solving skills? Should I learn design patterns or algorithms to improve my logical thinking skills? What to do when you're faced with a problem that you can't solve quickly? Are there non-programming related activities akin to solving programming problems? I am a computer engineering graduate. I have studied programming since three years. I am good in coding and programming. I have been trying to compete in algorithmic competitions on sites such as topcoder,spoj since one and a half year, but I am still unable to solve problems other than too easy problems. I have learned from people that it takes practice to solve such problems. I try to solve those problems but sometimes I am unable to understand and even if I do understand I am unable to think of a good algorithm for solving it. Even if I solve I get Wrong answer and I am unable to figure out what is the problem with my code as it works on samples given on the sites but fails on test cases which they do not provide. I really want to solve those problems and become good in algorithms. I have read books for learning algorithms like Introduction to algorithms by CLRS,practicing programming questions. I have gone through some questions but they don't answer this question. I have seen the questions which are said duplicates but those questions focus on overall programming, but I am asking for algorithm related programming, basically for competing in programming which involve solving a problem statement then online judge will automatically evaluate it, such type of programming is quite different from the type of programming these questions discuss.

  Read the article

• Why is Quicksort called "Quicksort"?

  - by Darrel Hoffman

  The point of this question is not to debate the merits of this over any other sorting algorithm - certainly there are many other questions that do this. This question is about the name. Why is Quicksort called "Quicksort"? Sure, it's "quick", most of the time, but not always. The possibility of degenerating to O(N^2) is well known. There are various modifications to Quicksort that mitigate this problem, but the ones which bring the worst case down to a guaranteed O(n log n) aren't generally called Quicksort anymore. (e.g. Introsort). I just wonder why of all the well-known sorting algorithms, this is the only one deserving of the name "quick", which describes not how the algorithm works, but how fast it (usually) is. Mergesort is called that because it merges the data. Heapsort is called that because it uses a heap. Introsort gets its name from "Introspective", since it monitors its own performance to decide when to switch from Quicksort to Heapsort. Similarly for all the slower ones - Bubblesort, Insertion sort, Selection sort, etc. They're all named for how they work. The only other exception I can think of is "Bogosort", which is really just a joke that nobody ever actually uses in practice. Why isn't Quicksort called something more descriptive, like "Partition sort" or "Pivot sort", which describe what it actually does? It's not even a case of "got here first". Mergesort was developed 15 years before Quicksort. (1945 and 1960 respectively according to Wikipedia) I guess this is really more of a history question than a programming one. I'm just curious how it got the name - was it just good marketing?

  Read the article

• What is a simple deformer in which vertices deform linearly with control points?

  - by sebf

  In my project I want to deform a complex mesh, using a simpler 'proxy' mesh. In effect, each vertex of the proxy/collision mesh will be a control point/bone, which should deform the vertices of the main mesh attached to it depending on weight, but where the weight is not dependant on the absolute distance from the control point but rather distance relative to the other affecting control points. The point of this is to preserve complex three dimensional features of the main mesh while using physics implementations which expect something far simpler, low resolution, single surface, etc. Therefore, the vertices must deform linearly with their respective weighted control points (i.e. no falloff fields or all the mesh features will end up collapsed) - as if each vertex was linked to a point on the plane created by the attached control points and deformed with it. I have tried implementing the weight computation algorithm in this paper (page 4) but it is not working as expected and I am wondering if it is really the best way to do what I want. What is the simplest way to 'skin'* an arbitrary mesh, to another arbitrary mesh? *By skin I mean I need an algorithm to determine the best control points for a vertex, and their weights.

  Read the article

• How can I find the shortest path between two subgraphs of a larger graph?

  - by Pops

  I'm working with a weighted, undirected multigraph (loops not permitted; most node connections have multiplicity 1; a few node connections have multiplicity 2). I need to find the shortest path between two subgraphs of this graph that do not overlap with each other. There are no other restrictions on which nodes should be used as start/end points. Edges can be selectively removed from the graph at certain times (as explained in my previous question) so it's possible that for two given subgraphs, there might not be any way to connect them. I'm pretty sure I've heard of an algorithm for this before, but I can't remember what it's called, and my Google searches for strings like "shortest path between subgraphs" haven't helped. Can someone suggest a more efficient way to do this than comparing shortest paths between all nodes in one subgraph with all nodes in the other subgraph? Or at least tell me the name of the algorithm so I can look it up myself? For example, if I have the graph below, the nodes circled in red might be one subgraph and the nodes circled in blue might be another. The edges would all have positive integer weights, although they're not shown in the image. I'd want to find whatever path has the shortest total cost as long as it starts at a red node and ends at a blue node. I believe this means the specific node positions and edge weights cannot be ignored. (This is just an example graph I grabbed off Wikimedia and drew on, not my actual problem.)

  Read the article

• Sorting versus hashing

  - by Paul Siegel

  My problem is as follows. I have an array of n strings with m < n of them distinct. I want to create a one-to-one function which assigns each of the m distinct strings to the numbers 0 ... m-1. For example, if my strings are: Bob, Amy, Bob, Charlie, Amy then the function: Bob -> 0, Amy -> 1, Charlie -> 2 would meet my needs. I have thought of three possible approaches: Sort the list of strings, remove duplicates, and construct the function using a search algorithm. Create a hash table and check each string to see if it is already in the table before inserting it. Sort the list of strings, remove duplicates, and put the resulting list into a hash table. My code will be written in Java, and I will likely use standard Java algorithms: merge sort for sorting, binary search for searching, and whatever the standard Java hash table algorithm is. Question: Assume that after creating the function I will have to evaluate it on each of the n original strings. Which of the three approaches is fastest? Is there a better way? Part of the problem is that I don't really know what's going on "under the hood" in standard hashing algorithms. Any help would be appreciated.

  Read the article

• Greiner-Hormann clipping problem

  - by Belgin

  I have a set of planar polygons in 3D space defined by their vertices in counterclockwise order. Let's define the 'positive face' as being the face of the 3D polygon such as when observed, the vertices appear in counterclockwise order, and the 'negative face', the face which when observed, the vertices appear in clockwise order. I'm doing perspective projection of the set of polygons onto a projection polygon defined by the points in this order: (0, h, 0), (0, 0, 0), (w, 0, 0), and (w, h, 0), where w and h are strictly positive integers. The positive face of this projection polygon is oriented towards positive Z, and the camera point is somewhere at (0, 0, d), where d is a strictly negative number. In order to 'clip' the projected polygons into the projection polygon, I'm applying the Greiner-Hormann (PDF) clipping algorithm, which requires that the clipper and the to-be-clipped polygons be in the same order (i.e. clockwise or counterclockwise). My question is the following: How can I determine whether the projected face of the 3D polygon is the negative or the positive one? Meaning, how do I find out if I have to work with the vertices in normal or inverted order for the algorithm to work? I noticed that only if the 3D polygon is facing the projection polygon with its negative face, both of them are in the same order (counterclockwise), otherwise, a modification needs to be done. Here is a picture (PNG) that illustrates this. Note that the planes described by the polygon from the set and the projection polygon may not always be parallel.

  Read the article

• What is the best way to "carve" a terrain created from a heightmap?

  - by tigrou

  I have a 3d landscape created from a heightmap. I'd like to "carve" some holes in that terrain. That will allow me to create bridges, caverns and tunnels inside it. That operation will be done in the game editor so it doesn't need to be realtime. In the end, rendering is done using traditional polygons. What would be the best/easiest way to do that ? I already think about several solutions : Solution 1 1) Create voxels from the heightmap (very easy). In other words, fill a 3D array like this : voxels[32][32][32] from the heightmap values. 2) Carve holes in the voxels as i want (easy too). 3) Convert voxels to polygons using some iso-surface extraction technique (like marching cubes). 4) Reduce (decimate) polygons created in 3). This technique seems to be the most promising for giving good results (untested). However the problem with marching cubes is that they tends to produce lots of polygons thus reducing them is mandatory. Implementing 4) also seems not trivial, i have read several papers on the web and it seems pretty complex. I was also unable to find an example, code snippet or something to start writing an algorithm for triangle mesh decimation. Maybe there is a special decimation algorithm (simpler) for meshes created from marching cubes ? Solution 2 1) Create some triangle mesh from the heighmap (easy). 2) Apply severals 3D boolean operation (eg: subtraction with a sphere) to carve the mesh. 3) apply some procedure to reduce polygons (optional). Operation 2) seems to be very complex and to be honest i have no idea how to do that. Also applying many boolean operation seems to be slow and will maybe degrade the triangle mesh every time a boolean operation is applied.

  Read the article

< Previous Page | 146 147 148 149 150 151 152 153 154 155 156 157  | Next Page >