Coursera Flash Sale
40% Off Coursera Plus for 3 Months!
Grab it
Explore the evolution of compiler technology in this conference talk that traces the journey from early assembly translators to the cutting-edge Compiler 4.0 paradigm. Learn how traditional optimizing compilers like GCC and LLVM automated code transformation for CPUs, and discover how ML compilers such as XLA, TVM, and TorchInductor revolutionized computational graph compilation for modern accelerators. Understand the scalability challenges facing current Compiler 3.0 systems when dealing with diverse model architectures and hardware targets, and examine how the combinatorial optimization problem exceeds the capabilities of static heuristics and bounded kernel search. Discover yasp.compile's innovative agentic ML compiler approach that explicitly reasons about hardware constraints to generate tailored low-level implementations for specific model-hardware pairs. Gain insights into how hardware-aware graph optimization combined with learned code generation and cost modeling can reduce manual kernel engineering while improving adaptability across heterogeneous edge accelerators, representing a significant advancement in compiler technology for edge AI applications.
