This internship offers students the opportunity to gain insight into the world of high-performance computing on specialized hardware accelerators. The goal of the internship is to create benchmark tasks for the modern AMD VCK5000 and AMD Alveo U50LV acceleration cards. Typical tasks such as matrix multiplication, Fourier transforms, and neural network inference will be tested. The results will be compared with calculations on standard GPUs and CPUs, allowing for a practical assessment of the effectiveness and suitability of individual computing platforms.
What will the internship involve?
• Creating computational benchmarks on AMD FPGA cards
• Implementation of test tasks (e.g., matrix multiplication, FT calculations, neural network inference)
• Comparison of performance with alternatives running on GPU/CPU
• Processing of results and basic interpretation
• Working in a Linux environment, using the C++ language
What we need:
• Good knowledge of the C++ language
• Experience with the Linux environment (working in a terminal, scripting, etc.)
• Interest in high-performance computing, acceleration cards, and algorithm optimization
• Independence, diligence, and willingness to learn new technologies
What we offer:
• Access to high-performance computing hardware and modern technologies
• Opportunity to test various architectures in practice
• Flexible working hours and adaptation to your study schedule
• Professional guidance and mentoring from experienced experts
Are you interested in getting the most out of computing technology? Apply for an internship and find out which platform is truly the most powerful for typical computing tasks!