You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
👀 I’m interested in designing and developing distributed programming languages for the future world.
🌱 I’m currently learning artifical intelligence but I think AI is still not comparable with human's brain. Human's brain has 86 billion neurons all of which can work in parallel. Nevertheless AI algorithm can take advantage of multi-core GPUs and I acknowledge that a GPU core is much more powerful than a neuron in brain. However, all the GPU cores are homogeneous and have to do similiar job in a parallel computing task. This introduces complexity in data sharing/alignment and task scheduling and, as a result, designing algorithms to tackle complicated problems is extremely difficult. Moreover, the number of cores is limited, thousands of powerful cores still cannot compete with 86 billion neurons. To cope with this bottleneck, a possible solution is to redesign programming language to take advantage computing powers of distributed and different devices.
💞️ To this end, I’m currently developing a new programming language called MFP, which means Mathematic language For Parallel computing. This is a cross platform scripting language for both JVM (repo name is MFPLang4JVM) and Android (repo name is MFPAndrioLib). MFP is not implemented as a geek’s toy. Instead, it has a great target to distribute codes to heterogeneous Android and JVM devices and run the codes in parallel. Achieving this aim is extremely important to large-scale simulation, real-time remote control and sophisticated artificial intelligence (AI) tasks. Support and collaboration are very much welcomed. Feel free to send email to me if you have some great ideas about MFP.