Brevis releases Pico Prism, enabling real-time Ethereum proofs on consumer-grade hardware

Chainfeeds Guide:

Pico Prism (zkVM) achieves a 3.4x improvement in performance efficiency on the RTX 5090 GPU.

Source:

Author:

Brevis

Opinion:

Brevis: On October 15, 2023, Brevis announced that its Pico Prism achieved 99.6% real-time proof coverage (within 12 seconds) and 96.8% coverage within 10 seconds for current Ethereum mainnet blocks (45M Gas limit) on consumer-grade hardware. This breakthrough demonstrates that real-time proofs can be achieved using the commonly available RTX 5090 GPU, without the need for expensive dedicated data center equipment. This achievement marks a high level of economic viability for broader deployment of the technology, while supporting Ethereum's decentralization goals. Under the same 36M Gas limit as previous benchmarks, Pico Prism exhibited performance far superior to previous solutions. In terms of coverage, Pico Prism achieved 98.9% coverage within 10 seconds, compared to only 40.9% for previous solutions, an increase of 141%; in terms of hardware cost, Pico Prism costs $128K, a 50% reduction compared to the previous $256K; for average proof time, Pico Prism achieved 6.04 seconds, 71% faster than the previous 10.3 seconds; and GPU requirements were significantly reduced, from 160 RTX 4090s to 64 RTX 5090s, a 60% decrease. Overall, Pico Prism's performance efficiency improved by 3.4 times, demonstrating its exceptional cost-effectiveness. For the 45M Gas limit, Pico Prism achieved 99.6% coverage (within 12 seconds) and 96.8% coverage (within 10 seconds), with an average proof time of 6.9 seconds. This outstanding performance shows that Pico Prism can not only handle the current production volume of Ethereum blocks but also meet Ethereum's decentralization requirements. Brevis co-founder and CEO Michael pointed out that Pico Prism's excellent performance on consumer-grade hardware perfectly aligns with Ethereum's decentralization goals, enabling more users to participate without the need for expensive hardware. Pico Prism's outstanding performance stems from its complete reconstruction from a single-machine proof architecture to a distributed multi-GPU cluster. Unlike traditional single-GPU computing architectures, Pico Prism distributes compute-intensive tasks to multiple GPUs for parallel processing through a highly parallelized pipeline, while the CPU is responsible for coordinating and setting up tasks. Through this distributed architecture, Pico Prism achieves maximum efficiency utilization among multiple GPUs, significantly enhancing hardware performance. In addition, the Pico Prism system leverages Pico's modular design, allowing computing resources to be flexibly configured as needed. Each module can independently complete part of the task, and the system coordinates to achieve the final result. In this way, Pico Prism achieves higher computational efficiency with lower hardware costs and fewer GPUs, which is one of the key advantages over solutions requiring expensive dedicated data center equipment. In terms of technical architecture, Pico Prism also supports nearly linear scalability. As more GPUs are added, the increase in computing power is almost linear, meaning the architecture can effectively scale to handle more tasks without large-scale reconstruction. Through distributed computing and a highly parallelized pipeline, Pico Prism achieves performance on consumer-grade hardware that previously required high-end dedicated hardware, demonstrating the tremendous potential of zero-knowledge proof technology. Pico Prism's technological advancements have far-reaching significance for Ethereum scaling. Traditionally, Ethereum transaction verification requires more than 800,000 validators worldwide to independently recalculate, but Pico Prism, through cryptographic proofs, can centralize the verification process, with a single prover generating a mathematical proof and the remaining validators completing verification in milliseconds. This approach not only greatly reduces the cost of verification but also increases the speed of verification, further advancing Ethereum's scaling process.