Projects tagged ‘benchmark’ and ‘c’

NObjective bridge to Objective-C provides the necessary classes to develop and run .NET applications which interact with Objective-C frameworks and libraries under Mac OS X with high performance and reliability.

John Walker's FBENCH in C#. Made to benchmark .NET vs Mono, but I am sure it has other uses also... Compiles with .NET 2.0 and Mono 1.9

The Crackling Fire benchmark brute-forces all the valid states of a 27-segment wooden folding puzzle (finding the two solutions in the process). At the time of this writing, a sequential C ... [More] implementation is provided along with two ports: concurrent Java with Fork/Join, and concurrent Erlang. An illustration of the unfolded puzzle: [Less]

The NObjective bridge to Objective-C provides the necessary classes to develop and run .NET applications that interact with Objective-C frameworks and libraries under Mac OS X with high performance ... [More] and reliablity. Description: High-performance solution with smallest cpu/memory overheads compared to similar bridges ( monobjc, mobjc, cocoa# ) Huge library of automatically generated proxies around objc classes More than 2300 classes and 23000 methods can be used directly from your .NET code Proxies are genereated using objc-introspection assisted by .h files parsing to achieve more natural method signatures Uses NObjectiveAST library for header parsing Generates essential enums and structs Also proxies can be generated without C header files Transparent cross-runtime exception rethrowing from managed to unmanaged world from unmanaged to managed world Easy access to major Mac OS X frameworks and libraries Ability to generate objc proxies for any framework Ability to build your own types that can be used inside Cocoa Combines flexibility of managed and power of unmanaged code to achieve highest performance ships with x86/ppc UB library Used special coding standart that developed with high perception in mind not habit Client applications can be written in any .NET language ( C#, VB.NET, etc ) Currently works only on MacOS Leopard 10.5 ( x86/ppc ) Open Source, Free Software Notes: Library targeted for mature Objective-C developers so if something not work.. try to do it on Objective-C first. Library doesn't provide any tools to simplify build process on Mac. But you can use build script located at "example" directory to start from. If users base will be enough I can further improve performance up to three times. Also with same condition I can add support for MacOS Tiger 10.4 Feedback and remarks are welcomed on http://groups.google.ru/group/nobjective Results of performance benchmark ( 19.11.2008 ): Running benchmark on my Mac Pro ( Intel Xeon ): native objc results: Invocation: 0.855ms ImpInvocation: 0.741ms Allocation: 2.303ms Vararg marshaling: 115.754ms try/catch overhead: 1.116ms Exception rethrow: 35.797ms TotalMemory: 324 KiB NObjective results: Invocation: 9.713ms Allocation: 32.133ms Vararg marshaling: 568.244ms try/catch overhead: 111.074ms Exception rethrow: 1806.227ms TotalMemory: 736 KiB monobjc results: Invocation: 34.632ms Allocation: 151.963ms Vararg marshaling: 928.468ms TotalMemory: 8400 KiB mobjc results: Invocation: 162.901ms Allocation: 290.397ms try/catch overhead: 419.917ms Exception rethrow: 13066.63ms TotalMemory: 1368 KiB Cocoa# results: Invocation: 348.862ms Allocation: 1826.093ms Vararg marshaling: 8268.366ms TotalMemory: 14716 KiB Note: vararg support not supported by mobjc Note: cross-runtime exception rethrowing not supported by monobjc/cocoa# Running benchmark on my macmini ( PowerPC G4 ) native objc results: Invocation: 4.357ms Allocation: 12.821ms Vararg marshaling: 429.634ms try/catch overhead: 8.817ms Exception rethrow: 137.837ms TotalMemory: 316 KiB NObjective results: Invocation: 48.434ms Allocation: 191.182ms Vararg marshaling: 2253.413ms try/catch overhead: 633.873ms Exception rethrow: 6135.595ms TotalMemory: 756 KiB monobjc results: Invocation: 173.06ms Allocation: 745.805ms Vararg marshaling: 4117.254ms TotalMemory: 8356 KiB mobjc results: Invocation: 1669.981ms Allocation: 4265.934ms try/catch overhead: 3622.483ms Exception rethrow: 94040.912ms TotalMemory: 34132 KiB Cocoa# results: Invocation: 1483.704ms Allocation: 6298.719ms Vararg marshaling: 32280.169ms TotalMemory: 14668 KiBContacts: Email: eugeny.grishul(a)gmail.com Facebook: http://www.facebook.com/profile.php?id=1038466393 VKontakte: http://vkontakte.ru/id181869 WBR, Eugeny Grishul [Less]

The goal of this project is to develop a set of algorithms implemented in many different languages including compiled and interpreted. Thoose test will allow us to know how much will an application ... [More] be improved if implemented into another language. The tests are compared with a crude standard C. And the algorithm is initially implemented in Python and ported to all the others languages. Please, if you run the tests, send me the results at jacobidiego(at)gmail.com Any contributions are welcomed. The next results are promediated from all runs and may vary on the algorithm used and the parameters passed. Results for Quine-MacCLusky Algorithm: "A procedural minterms reduction algorithm for logic ecuations, similar in function to Karnaugh" Python is 34 times slower than C. C++ is 3 times slower than .C Note that, Python efficiency varies too much depending on the number of variables to manipulate. Plain Standard C is a pain to develop, most developers uses APIs like GLib to make this task done. Another project aiming to bench languages is at: http://shootout.alioth.debian.org/ [Less]

CompBenchmarks automates installation, configuration (using specific compilers or arbitrary options), and evaluation of some well-known C/C++ benchmarks and applications. Results are given in XML ... [More] files that can be sent to the CompBenchmarks Web site. [Less]

io testing tool