arnaucube 1375596a74 | 5 years ago | |
---|---|---|
bn128 | 5 years ago | |
circuitcompiler | 5 years ago | |
fields | 5 years ago | |
r1csqap | 5 years ago | |
r1csqapFloat | 6 years ago | |
.gitignore | 6 years ago | |
LICENSE | 6 years ago | |
README.md | 5 years ago | |
go.mod | 5 years ago | |
go.sum | 5 years ago | |
snark.go | 5 years ago | |
snark_test.go | 5 years ago |
zkSNARK library implementation in Go
Succinct Non-Interactive Zero Knowledge for a von Neumann Architecture
, Eli Ben-Sasson, Alessandro Chiesa, Eran Tromer, Madars Virza https://eprint.iacr.org/2013/879.pdfPinocchio: Nearly practical verifiable computation
, Bryan Parno, Craig Gentry, Jon Howell, Mariana Raykova https://eprint.iacr.org/2013/279.pdfImplementation from scratch in Go to understand the concepts. Do not use in production.
Not finished, implementing this in my free time to understand it better, so I don't have much time.
Current implementation status:
Example:
// compile circuit and get the R1CS
flatCode := `
func test(x):
aux = x*x
y = aux*x
z = x + y
out = z + 5
`
// parse the code
parser := circuitcompiler.NewParser(strings.NewReader(flatCode))
circuit, err := parser.Parse()
assert.Nil(t, err)
fmt.Println(circuit)
// witness
b3 := big.NewInt(int64(3))
inputs := []*big.Int{b3}
w := circuit.CalculateWitness(inputs)
fmt.Println("\nwitness", w)
/*
now we have the witness:
w = [1 3 35 9 27 30]
*/
// flat code to R1CS
fmt.Println("generating R1CS from flat code")
a, b, c := circuit.GenerateR1CS()
/*
now we have the R1CS from the circuit:
a == [[0 1 0 0 0 0] [0 0 0 1 0 0] [0 1 0 0 1 0] [5 0 0 0 0 1]]
b == [[0 1 0 0 0 0] [0 1 0 0 0 0] [1 0 0 0 0 0] [1 0 0 0 0 0]]
c == [[0 0 0 1 0 0] [0 0 0 0 1 0] [0 0 0 0 0 1] [0 0 1 0 0 0]]
*/
alphas, betas, gammas, zx := snark.Utils.PF.R1CSToQAP(a, b, c)
ax, bx, cx, px := snark.Utils.PF.CombinePolynomials(w, alphas, betas, gammas)
hx := snark.Utils.PF.DivisorPolinomial(px, zx)
// hx==px/zx so px==hx*zx
assert.Equal(t, px, snark.Utils.PF.Mul(hx, zx))
// p(x) = a(x) * b(x) - c(x) == h(x) * z(x)
abc := snark.Utils.PF.Sub(pf.Mul(ax, bx), cx)
assert.Equal(t, abc, px)
hz := snark.Utils.PF.Mul(hx, zx)
assert.Equal(t, abc, hz)
div, rem := snark.Utils.PF.Div(px, zx)
assert.Equal(t, hx, div)
assert.Equal(t, rem, r1csqap.ArrayOfBigZeros(4))
// calculate trusted setup
setup, err := snark.GenerateTrustedSetup(len(w), circuit, alphas, betas, gammas, zx)
assert.Nil(t, err)
fmt.Println("t", setup.Toxic.T)
// piA = g1 * A(t), piB = g2 * B(t), piC = g1 * C(t), piH = g1 * H(t)
proof, err := snark.GenerateProofs(circuit, setup, hx, w)
assert.Nil(t, err)
assert.True(t, snark.VerifyProof(circuit, setup, proof))
go test ./... -v
Thanks to @jbaylina, @bellesmarta, @adriamb for their explanations that helped to understand this a little bit. Also thanks to @vbuterin for all the published articles explaining the zkSNARKs.