Y
HN Search
Hacker News Search
new
|
comments
|
top
|
jobs
y00zzeek
searching PlanetScale…
1.
▲
2.
▲
3.
▲
4.
▲
5.
▲
6.
▲
3 ms
·
1.
▲
by
y00zzeek
9mo ago
The prompt was: RUST_LOG=debug RUSTFLAGS="-C target-cpu=native" cargo run --release --example keccak --no-default-features --features "std parallel blake3" The completion took 88s and 21.5GB RAM.
2.
▲
by
y00zzeek
9mo ago
Since the edit window is closed, I want to clarify the AIR structure for those asking about the "row" definition. In Hekate's Keccak AIR, the relationship is ~25 trace rows per 1 Keccak-f[1600] permutation. 2^24 Rows = The ra
3.
▲
by
y00zzeek
9mo ago
Fair point on the tone. I'll trade the rhetoric for physics. A "row" in this context is a single step of the Keccak-f[1600] permutation within the AIR (Algebraic Intermediate Representation) table. Most engines materialize th
4.
▲
by
y00zzeek
9mo ago
Exactly. If we can't prove 2^24 rows on a laptop, ZK will stay centralized forever. Hekate is my answer to the memory wall that forces teams into $2+/hour AWS instances. Proving should be a commodity, not a luxury.
5.
▲
by
y00zzeek
9mo ago
My motivation for building Hekate is simple: I am done watching well-funded teams with 50+ people and a busload of PhDs produce engineering trash. There is a massive, widening gap between academic brilliance and silicon-level implementation
6.
▲
Show HN: Hekate – A Zero-Copy ZK Engine Overcoming the Memory Wall
8 points
by
y00zzeek
9mo ago
|
10 comments
7.
▲
by
y00zzeek
9mo ago
Here is the the repo (currently closed source, but manifesto is up): https://github.com/oumuamua-corp/hekate
8.
▲
Show HN: Proving 67M ZK rows on a laptop in 28s (Winterfell OOMs)
1 points
by
y00zzeek
9mo ago
|
1 comments