Coursera Flash Sale
40% Off Coursera Plus for 3 Months!
Grab it
Explore the essential libraries and techniques for writing efficient CUDA C++ code without the complexity of custom kernel development in this 39-minute conference talk from ACCU 2025. Learn to evaluate when to use existing CUDA libraries versus writing custom kernels, and discover popular tools like Thrust, CUB, and libcu++ that simplify GPU programming with C++. Master the art of composing different CUDA libraries with your own custom code while building intuition about performance implications and heterogeneous memory management. Gain practical insights into algorithm design and synchronization for modern GPU programming, and develop the confidence to select the optimal tools for accelerating your C++ applications across diverse use cases.
