讲解：CSCM98、Multiple Data、C++、c++Python| Statistic

CSCM98, Assignment #1:Accelerating fractals (SIMD+MT)Important readingBy submitting this coursework, electronically and/or hardcopy, you state that you fully understand and arecomplying with the universitys policy on Academic Integrity and Academic Misconduct. Note that course material can be used, but this is a personal piece of work and any extra material must bereferenced. The course page includes a zip file to run the original version of the code. Submission instructionswill be sent to you by email. The process will mainly be to reply to the email and attach yourBasicApplication.cpp file. Your code should be compilable with the given project.Some extra course material can be found at: http://cs.swansea.ac.uk/~csmora/CSCM98/A list of Single Instruction Multiple Data (SIMD) Advanced Vector Extensions (AVX) instructions can also befound at:https://software.intel.com/sites/landingpage/IntrinsicsGuide/DescriptionWe want to accelerate an algorithm that creates fractal images by combining SIMD programming andmultithreading.The algorithm creates a fractal image (see image above) by iterating for each pixel a fractal value. For eachpixel, the Iterate(pixel) function is called and returns more or less the number of iterations performed. While thecode uses complex numbers, the original C code to compute the iteration is already given and you just have toporCSCM98代做、代写Multiple Data、C++程序t the sequence of instructions to an SIMD device which shall process 8 pixels at once, as well asmultithreading the code. As usual, most of the code is already written, and you will have to complete and returnthe .cpp file. Marks breakdown 30%: Code is properly multithreaded and efficient. 25%: A basic use of SIMD instructions (AVX) is implemented and the code is significantly faster (SIMDaccelerated). 20%: The SIMD code is optimised heavily, making use of more advanced instructions than just add or mul.However, it is forbidden to use the _mm256_testz_ps function, which will have to be replaced by anotherone. 15%: Workload is balanced across all threads, and therefore the algorithm scales well. Please note here thatthere exists a very simple solution to this problem. 5%: Computation times are written as comments in the code, including the computer used to do so. 5%: Submission follows guidelines and does not involve extra work (e.g. correcting code).Notes: Any partial solution may return only partial marks. The exact marks will depend on the quality of theanswers and can only be evaluated upon reception of the coursework. This is a strictly personal work. Any external help must be referenced as comments in the code. Add your student ID and Name to the code. You will receive an email detailing how you should submit your work soon. 转自：http://www.6daixie.com/contents/13/4475.html