Lv schrieb: > My question is along the lines of this paper. > "http://citeseer.ist.psu.edu/555028.html"; > > Is there anybody who developed a unit or did any work for using GPU's > for numerical computation with FPK/Lazarus?
Only in C. > > Also, if anyone knows, what are the increase in speed expected over the > best Intel or AMD has to offer compared to a current best GPU? Hard to say, depends on your application. A modern graphics card has a peak performance of ~500 GFLOPS, a modern core CPU ~25 GFLOPS, so if you can explore the performance of the GPU, you can get a speed up of 20 in theory. > > > Developing a floating point (if such a GPU exists in double or extended > precision) unit on a GPU should not be so difficult as the Unit only > need to do a few basic elementary functions from which all the remainder > of functions and special functions can be obtained. > > Here is an overview of GPU issues. > "http://64.233.169.104/search?q=cache:SastpUrSdOMJ:www.gpgpu.org/sc2006/workshop/Apple_GPUintrinsics.pdf+gpu+numerical+floating&hl=en&ct=clnk&cd=1&gl=us"; > > > Anyone have any info? > > Thanks > > _________________________________________________________________ > To unsubscribe: mail [EMAIL PROTECTED] with > "unsubscribe" as the Subject > archives at http://www.lazarus.freepascal.org/mailarchives > _________________________________________________________________ To unsubscribe: mail [EMAIL PROTECTED] with "unsubscribe" as the Subject archives at http://www.lazarus.freepascal.org/mailarchives
