InfoQ Homepage GPU Content on InfoQ
Articles
RSS Feed-
TornadoVM: Accelerating Java with GPUs and FPGAs
The proliferation of heterogeneous hardware represents a problem for programming languages such as Java that target CPUs. TornadoVM extends the Graal JIT compiler to take advantage of GPUs & FPGAs and provides a flexible, high-level model whilst still enabling high performance and features such as live task migration.
-
Joint Forces: From Multithreaded Programming to GPU Computing
In this IEEE article, authors Frank Feinbube, Peter Troger and Andreas Polze discuss two major hardware trends in the desktop parallel programming space, multi-core CPU architectures and Graphics Processing Units (GPUs). They also talk about the best practices for GPU code optimization like algorithm design, memory transfer, control flow, instructions and precision.