3 Replies Latest reply on Sep 15, 2015 8:56 AM by jtrudeau

    "OEM" of PC (AMD CPU, GPU) + OS: Windows 10 Pro. Cross-platform targets: Android/Linux-like, Windows

    code_review

      I built my own machine with AMD hardware CPU: FX-4350 Black, GPU: R9-2XX, 990 FX chipset and motherboard: ASROCK Extreme 9.  OS is Windows 10 Pro, or openSUSE 13.2 Linux on another SSD if necessary - Windows 10 still seems fragile and blue-screens.

       

      I'm beginning development to reach the mobile Android-and-other-Linux-like OS platforms and Windows 10, so I make heavy use of AMD-V in virtualizaton, (e.g. Android, Linux, Windows 10(phone) emulators).

      I was using VirtualBox 5.x until it spontaneously hosted 32-bit only environments. So I use Windows 10 Pro's Hyper-V for now.

       

      I prefer the Visual Studio (2015) IDE as of now.

       

      There is so much talk of "cross-platform" development within the Microsoft Visual Studio community. The most popular methods of packaging an app to a target platform's app store include using Google's Javascript V8 and Node.js projects, along with Apache Cordova. Secondly, there's a C/C++ way to write code that targets Android, Linux, Windows 10/phone/"Universal". I recently been programming in interpreted languages e.g. CPython 2.7/3.x, which has the advantage of being pre-installed in almost every Linux-like environment. I'm aware of the GIL, however, and it seems Python 3.x has ways to make the most out of  4 CPU cores, which is the upper-limit of compute core hardware on Android and most other machines atm, where if more cores are needed, the additional cores are provided by a GPU.  I know the Java language, but C# thoroughly. Downloading a run-time that doesn't already exist on the target platform would be an annoyance for some users. I know the C stdlib somewhat, but I'm very inexperienced with C++.

       

      I'm highly interested in using the AMD compute libraries to maximize performance on mobile and desktop platforms built with AMD CPU/GPU/HSA hardware.