Topic: Factorization cuda
Kind time of days. Faced necessity of implementation of factorization on GPU. Now implemented simple search in search of a zero remainder of division, provided that a generated denominator - a simple natural number. According to it, it would be desirable to learn, whether there are variants more the optimized code capable of effective multisequencing. Tried p-1 Pollarda (multisequencing for the account of a choice of a random number individually for each flow) - as that did not please.... The gain of naturally linear is not present. And here Fermat a good variant, but not parallelizable (at least I did not see such possibility) And could not find the information yet, whether there are implementations for CUDA, allowing to operate with great numbers? Thankful in advance.