1/ Introducing Cloud-ZK: A Toolkit for Developing Zero Knowledge Proof Acceleration in the Cloud 🥳
TL;DR
Today we #opensource a new way to accelerate ZK. Using #FPGA. In #AWS cloud. Easy. Accessible. Cheap.
Read on, or check out the blog
Thread 🧵👇
github.com/ingonyama-zk/c…
TL;DR
Today we #opensource a new way to accelerate ZK. Using #FPGA. In #AWS cloud. Easy. Accessible. Cheap.
Read on, or check out the blog
Thread 🧵👇
github.com/ingonyama-zk/c…
2/
Zero Knowledge (ZK) technology is a fundamental building block for decentralized computing.
Its two main applications are privacy-preserving computation and verifiable computation.
Blogpost: medium.com/@ingonyama/clo…
Zero Knowledge (ZK) technology is a fundamental building block for decentralized computing.
Its two main applications are privacy-preserving computation and verifiable computation.
Blogpost: medium.com/@ingonyama/clo…
3/
For specific types of ZK such as SNARK and STARK-based systems, additional properties include public verifiability, smaller proof sizes, and fast verification.
This makes these kinds of ZK perfect for use in blockchains for scalability and privacy purposes.
For specific types of ZK such as SNARK and STARK-based systems, additional properties include public verifiability, smaller proof sizes, and fast verification.
This makes these kinds of ZK perfect for use in blockchains for scalability and privacy purposes.
4/
At present, ZK tech is being developed and used by leading blockchains (e.g. @Filecoin, @AleoHQ), L2s (e.g. Starknet, @zksync) and decentralized applications (e.g. @darkforest_eth, ZKAttestor)
At present, ZK tech is being developed and used by leading blockchains (e.g. @Filecoin, @AleoHQ), L2s (e.g. Starknet, @zksync) and decentralized applications (e.g. @darkforest_eth, ZKAttestor)
5/
BUT.. nothing great ever comes easy 🤷♂️
The Prover must run a computationally intensive algorithm with significant data blowup during the computation. Recent estimates suggest up to 10M in prover overhead while producing the proof, compared to directly running the computation
BUT.. nothing great ever comes easy 🤷♂️
The Prover must run a computationally intensive algorithm with significant data blowup during the computation. Recent estimates suggest up to 10M in prover overhead while producing the proof, compared to directly running the computation
6/
Today, prover overhead is considered the main computational bottleneck for applied ZK 🚧
Without exception, every project built on ZK technology is facing or will face this bottleneck, which manifests adversely in either latency, throughput, memory, power consumption, or cost
Today, prover overhead is considered the main computational bottleneck for applied ZK 🚧
Without exception, every project built on ZK technology is facing or will face this bottleneck, which manifests adversely in either latency, throughput, memory, power consumption, or cost
7/
ZK computation requires modular arithmetic on large field sizes
Trials on CPUs lead 2 the conclusion that modern CPU architecture is just not built to handle this form of computation efficiently
As a result, the need for specialized hardware ZK prover acceleration is clear
ZK computation requires modular arithmetic on large field sizes
Trials on CPUs lead 2 the conclusion that modern CPU architecture is just not built to handle this form of computation efficiently
As a result, the need for specialized hardware ZK prover acceleration is clear
8/
Until now, the majority of hardware experimentation for accelerating ZK has been done with GPUs. In our recent paper, PipeMSM, we explored bringing functioning and operational ZK for the first time to FPGAs
eprint.iacr.org/2022/999.pdf
Until now, the majority of hardware experimentation for accelerating ZK has been done with GPUs. In our recent paper, PipeMSM, we explored bringing functioning and operational ZK for the first time to FPGAs
eprint.iacr.org/2022/999.pdf
9/
We believe this to be a superior approach for accelerating ZK computation, on the path to developing ZK for Application Specific Integrated Circuits (ASICs)
It is our belief that the energy efficiency in FPGAs is more suitable for ZK due to their function-specific design
We believe this to be a superior approach for accelerating ZK computation, on the path to developing ZK for Application Specific Integrated Circuits (ASICs)
It is our belief that the energy efficiency in FPGAs is more suitable for ZK due to their function-specific design
10/
With FPGAs at the base, we take a holistic approach to ZK optimization with designs based on a novel algorithmic approach and hardware-specific optimizations
With FPGAs at the base, we take a holistic approach to ZK optimization with designs based on a novel algorithmic approach and hardware-specific optimizations
11/
We implemented and tested our design on FPGA, and highlight the promise of optimized hardware over state-of-the-art GPU-based MSM solvers, in terms of both speed and energy expenditure.
We implemented and tested our design on FPGA, and highlight the promise of optimized hardware over state-of-the-art GPU-based MSM solvers, in terms of both speed and energy expenditure.
12/
A disadvantage of FPGAs compared to GPUs is their lack of accessibility, while GPUs have effectively become commodity hardware.
This means there is a high barrier for ZK users simply getting an FPGA into their hands.
A disadvantage of FPGAs compared to GPUs is their lack of accessibility, while GPUs have effectively become commodity hardware.
This means there is a high barrier for ZK users simply getting an FPGA into their hands.
13/
One promising approach to overcoming this barrier is by utilizing cloud computing
Clouds offering FPGA resources such as @awscloud, @alibaba_cloud have made FPGAs more accessible with “on-tap” accessibility.
This is a GAME-CHANGER in the world of hardware acceleration.
One promising approach to overcoming this barrier is by utilizing cloud computing
Clouds offering FPGA resources such as @awscloud, @alibaba_cloud have made FPGAs more accessible with “on-tap” accessibility.
This is a GAME-CHANGER in the world of hardware acceleration.
14/
Introducing Cloud-ZK
With the Cloud-ZK toolkit we have made #opensource, the benefits of ZK hardware acceleration on FPGAs are accessible to anyone, anywhere, anytime and with costs comparable to any other standard CPU instance on an hourly basis (link in Tweet 1)
Introducing Cloud-ZK
With the Cloud-ZK toolkit we have made #opensource, the benefits of ZK hardware acceleration on FPGAs are accessible to anyone, anywhere, anytime and with costs comparable to any other standard CPU instance on an hourly basis (link in Tweet 1)
15/
Our FPGA code achieves 4x the baseline of Zprize FPGA MSM competition, where max prize criteria is 2x :)
zprize.io/prizes/acceler…
Our FPGA code achieves 4x the baseline of Zprize FPGA MSM competition, where max prize criteria is 2x :)
zprize.io/prizes/acceler…
16/
As an example application, our AWS F1 instance can be used to accelerate Aleo prover.
In terms of proofs per second, the accelerator can generate up to ~5x proofs per second compared to running on CPU alone
As an example application, our AWS F1 instance can be used to accelerate Aleo prover.
In terms of proofs per second, the accelerator can generate up to ~5x proofs per second compared to running on CPU alone
17/
Follow for more Zero Knowledge hardware acceleration news as we work to make ZKP inexpensive, accessible, and fast
FIN
Github: github.com/ingonyama-zk
Follow for more Zero Knowledge hardware acceleration news as we work to make ZKP inexpensive, accessible, and fast
FIN
Github: github.com/ingonyama-zk
• • •
Missing some Tweet in this thread? You can try to
force a refresh
