Chris أُعيد تغريده
Chris
2.3K posts
Modular zk multi-prover std-Rust with a functional language hidden inside? Oh yes!
powdr labs@powdr_labs
𝗔𝗻𝗻𝗼𝘂𝗻𝗰𝗶𝗻𝗴 𝗽𝗼𝘄𝗱𝗿𝗩𝗠 𝟬.𝟭! powdrVM combines multi-prover support, std Rust, modularity, and high performance, setting the stage for future-proof ZK applications. Read the details on our blog or follow this thread for a tl;dr 👇 powdr.org/blog/powdrvm
English
Chris أُعيد تغريده
I am so proud of these folks, they have done an amazing job. Really cool tool. If you wanna do ZK, you gotta check powdr out. It's another level. NOT just for Ethereum or crypto -- you can do ANY program in it!
powdr labs@powdr_labs
We are excited to share that 𝑝𝑜𝑤𝑑𝑟 labs is now an independent organization focused on the next generation of ZK tooling, security, and applied research. Read the full announcement on our blog or follow this thread for a tl;dr 👇 powdr.org/blog/powdrlabs
English
Chris أُعيد تغريده
We are excited to share that 𝑝𝑜𝑤𝑑𝑟 labs is now an independent organization focused on the next generation of ZK tooling, security, and applied research.
Read the full announcement on our blog or follow this thread for a tl;dr 👇
powdr.org/blog/powdrlabs
English
I'm thrilled to work on powdr, especially with this capable team!
My hope is that it will amplify the already fast progress in the zk space by simplifying collaborations between isolated projects and making it easy to implement new research ideas.
powdr labs@powdr_labs
Introducing 𝑝𝑜𝑤𝑑𝑟, a modular stack for zkVMs! Earlier this year, @ethchris, @vanderkriek & @leonardoalt started the 𝑝𝑜𝑤𝑑𝑟 project to improve zkVM tooling. With the continued support from the @ethereum Foundation, the team will soon become 𝑝𝑜𝑤𝑑𝑟 𝑙𝑎𝑏𝑠. More ↓
English
@harpaljadeja You reveal what the signature to the message looks like. This is much more than just proving that you know the private key.
English
Weren't we using ZK all along?
Everytime I sign a message I am proving that I own the private key without revealing the private key 🤔
English
.@CPerezz19 just alpha-leaked POWDR: twitter.com/CPerezz19/stat…
It is a zkVM toolkit that relies on user-extensible languages and multi-stage compilation and features fully automated witness generation. It can generate proofs for Rust code via riscv.
github.com/chriseth/powdr
CPerezz.eth ⟠ 🦀@CPerezz19
Today, another awesome session on the ZK Vietnam Residency by @PrivacyScaling and @0xPARC . First @adria0 on Plaf(Plonkish Arithmetisation Format) + Powdr and Halo2
English
@CPerezz19 The future just became the present: github.com/chriseth/powdr…
(although it does not cover all riscv opcodes yet)
English
Finally, Adria spoke about the future of these tools. And how we might also get a RISC-V frontend soon.
Also shared some vision on opcode folding techniques!!
Also, ⚠️⚠️WE ALREADY HAVE A POWDR VERSION WORKING WITH HALO2!!!⚠️⚠️
English
Chris أُعيد تغريده
Today, another awesome session on the ZK Vietnam Residency by @PrivacyScaling and @0xPARC .
First @adria0 on Plaf(Plonkish Arithmetisation Format) + Powdr and Halo2
English
Chris أُعيد تغريده
I wanted to dive into the deepest details of zkRollups, so I'm writing a simple one from scratch and it's been extremely fun! It's quite modular, easy to change, and works end-to-end. I'm not in Denver, but if anyone wants to hack on it I'm happy to help!
github.com/leonardoalt/fu…
English
@mveehkim @mgrczyk It needs to do the shift in case something else is in the same storage slot, if you have `bool flag; address impl;`, for example. The old optimizer does not deal well with optimizing stack operations across sloads. Try the new one as @leonardoalt suggests.
English
@DeFinnTheFarmer The problem is that it actually IS easy to implement, but it's very likely to be used incorrectly with terrible consequences.
English
Btw even EIP 1553 is not a consensus. It's NOT easy to implement and relevant players like @ethchris and Daniel Kirchner have relevant concerns about it. We should at least hear all sides equally.
Belo Horizonte, Brazil 🇧🇷 English
It's so unfair to call users for a debate on EIP 1553 x EOF!
In one side we have "hey look we're trying to push a gas reduction update in your behalf".
On the other side we have a complex EVM update that is so hard to explain and so far from the end user.
It's not fair.
Belo Horizonte, Brazil 🇧🇷 English
@sendmoodz @big_tech_sux I'm not at all concerned about any compiler. I'm concerned about people writing buggy code or contracts not being composable anymore.
English
in general IF one is concerned about an evm opcode because of the complexity it introduces to the compiler, consider refactoring compiler. vyper added support in very few LOC cc @big_tech_sux
English
still in disagreement here. the concept of transient storage already exists in the evm. the only new class of bug created by transient storage opcodes is that you forget to clear slots in a call and it clears automatically. in case of reentrancy locks, that's a low severity bug
Chris@ethchris
Still of the same opinion on EIP-1153 (transient storage) as in my comments on the eth magicians thread in ethereum-magicians.org/t/eip-1153-tra… Summary: It's a very complicated feature that breaks invariants if implemented at the transaction level. It does more harm than good.
English
@z0age A call to a smart contract using transient storage can have totally different semantics depending on whether or not another call to the same smart contract has happened in the same transaction. If you split the calls to two different transactions, it again behaves differently.
English
I see it sort of like the argument for EXTSLOAD: there’s already a way (albeit an unnecessarily complicated one) to do this… so why not simplify things and lean into more efficient ways of doing what’s already possible?
Make the capabilities of the EVM explicit and transparent!
English
What’s this invariant at the tx level that’s being violated here?
Doesn’t everyone realize that devs are currently doing things like using the gas cost of warm vs cold access, or even selfdestruct semantics, to get around this supposed invariant?
Chris@ethchris
Still of the same opinion on EIP-1153 (transient storage) as in my comments on the eth magicians thread in ethereum-magicians.org/t/eip-1153-tra… Summary: It's a very complicated feature that breaks invariants if implemented at the transaction level. It does more harm than good.
English
Still of the same opinion on EIP-1153 (transient storage) as in my comments on the eth magicians thread in ethereum-magicians.org/t/eip-1153-tra… Summary: It's a very complicated feature that breaks invariants if implemented at the transaction level. It does more harm than good.
English