Innovating ZK Cloud Computing with Leo-ICICLE Integration
Since our debut with our first blog post, Introducing Ingonyama, we have set forth our mission: to lower the barrier of entry to Zero Knowledge Proofs (ZKP). Our strategy is straightforward: we seek out and utilize cutting-edge technologies to build products that accelerate and democratize access to ZKP technology.
This approach demands that our focus be on developing a robust infrastructure, spanning software, hardware, and algorithm optimization. In this pursuit, ICICLE has emerged as our flagship project. Originally developed as a CUDA library, ICICLE stands as a testament to our commitment to an uncompromising development experience (DevEx). Designed with simplicity in mind, ICICLE offers APIs in Rust, Go, and C++ that strike a fine balance between ease of use and powerful functionality, catering to both engineers and researchers alike.
The release of ICICLE v2 marked a significant milestone for us, introducing capabilities for polynomial arithmetic that greatly enhance the library’s versatility. This update enables comprehensive end-to-end prover implementations directly on GPUs.
Announcing Plans for ICICLE v3 in Collaboration with Accseal
Today, we are excited to unveil the next evolution of ICICLE. With the upcoming Version 3, we are expanding support beyond Nvidia GPUs to include Accseal’s Leo ZK-ASIC. This integration marks a pivotal development, introducing specialized ZK hardware that could revolutionize the way developers engage with ZK technologies.
We have been working closely with the Accseal team, testing the Leo chip, and collaboratively designing new algorithms to leverage its capabilities. Notably, the Leo chip enhances operations such as Multi-Scalar Multiplications (MSM) and Number Theoretic Transform (NTT), which are critical for optimizing the performance of ZK provers.
A Collaborative Experience and Looking Ahead
Working with the Accseal team has been a real joy thus far. As we integrate the Leo chip into ICICLE v3, we anticipate new opportunities not only for enhanced computational performance but also for significant energy optimization. Plans are underway to incorporate Leo into our ZaKi deployment architecture, which could introduce more cost-effective instances.
We are humbled by the widespread adoption of ICICLE and grateful for the opportunity to actively advance ZK performance during this unique period. Transitioning ICICLE from being solely GPU-based to also supporting native ZK ASIC technology marks an exciting step for us. We are eager to explore together the new possibilities this change will unlock.
Follow Ingonyama
Twitter: https://twitter.com/Ingo_zk
YouTube: https://www.youtube.com/@ingo_zk
GitHub: https://github.com/ingonyama-zk
LinkedIn: https://www.linkedin.com/company/ingonyama
Join us: https://www.ingonyama.com/career