The HomomorphicEncryption.org FHE Benchmarking effort provides standardized, end-to-end workloads to objectively evaluate encrypted computation across open-source, closed-source, and hardware-accelerated backends. In this talk, we share the suite’s latest status and newly integrated workloads, analyze initial submission insights on hardware acceleration, memory bottlenecks, and deployment trade-offs, and discuss community collaboration.
A panel discussion on privacy at the Open Source AI Summit.
An invited talk on evaluating and comparing fully homomorphic encryption workloads.
An invited talk on HEIR’s hardware work and approaches to FHE benchmarking.
An invited introduction to fully homomorphic encryption and its applications.
Practical FHE applications across hardware platforms using the HEIR compiler.
A working group session focused on FHE use cases and meaningful benchmarking.
An introduction to HEIR’s unified compiler approach and extensible FHE backends.
A talk on FHE stack, FHE hierarchy of needs and navigating ciphertext representations on the path to real-world FHE.
An invited talk on compiler design and intermediate representations for FHE programs.
A keynote on transpilers and their role in making FHE development more practical.
A presentation on Google’s transpiler tooling for fully homomorphic encryption.
An overview of Google’s C++ transpiler for fully homomorphic encryption computations.