Curtain demo runs on AMD through LLVM
Minotaurus/LLVM is a compiler plugin for LLVM which is capable to produce GPU code in OpenCL segments from standard C++. Although OpenCL is the same language and API for both NVidia and AMD, there are differences between the driver capabilities, for example, AMD GPUs do not support goto statements. It is easier to generate source code using goto statements than discovering structured code, that's why we were happy at our first steps on Fusion which was a GCC based solution of the problem.
Now the same feature, structured code generation, is done in the LLVM variant of Minotaurus with some extras like fall-through switch support. We tested the Curtain Demo on AMD HD 5770.
User login
Latest posts
-
Structured code generation for AMD GPUs are now supported in Minotaurus/LLVM
-
Julia demo can be compiled through LLVM now. We tested the new LLVM - Minotaurus interface today successfully.
-
Julia demo runs on AMD Fusion now. The speedup is around 4x compared to CPU. With the arrival of the Fusion platform we decided to tune the code generation of Minotaurus to support AMD GPUs as well.
-
Please find here our new demo video of Minotaurus. As an actual state of our main project, the system can compile the same C++ code parts to both CPU and GPU. Any comments are welcome! ...
This project has received funding from the European Union’s Horizon 2020 research and innovation programme under grant agreement No 767912.