SIX MILLION-DOLLAR PROBLEMS

The Clay Institute offers $1,000,000 for each of six unsolved problems.
Every one has a finite mirror in the ring Z/2310Z.
Not solutions — structural reflections. All computed. All verifiable.

Click each problem. Run the computation yourself.

1318 lines of C → JavaScript. gcc -O2 -lm. Zero dependencies.

M1
Riemann Hypothesis
Eigenvalue spacing statistics · Poisson vs GUE
Do the eigenvalue spacings of the CRT Laplacian look like a random (Poisson) or quantum (GUE) system? CRT channels are independent — the spacings converge to Poisson, not GUE. This is the integrable side: no level repulsion.
LIVE SPACING HISTOGRAM
Ring:
M2
Yang-Mills Mass Gap
Spectral gap formula · Confinement scaling
The spectral gap is exactly 4 sin²(π/p_max). The physical mass gap = N × gap diverges as N grows — this IS confinement. The scaling gap × ln(N)² converges to 4π² = 39.478.
GAP SCALING TABLE
M3
P vs NP
CRT search speedup · Live race
CRT decomposes search into independent channels. Find a target in O(Σpi) instead of O(N). The speedup grows unboundedly. Watch the race live.
BRUTE FORCE vs CRT SEARCH
Ring:
M4
Navier-Stokes Regularity
CRT Laplacian decomposition · 5 independent channels
The Laplacian decomposes into 5 independent channel Laplacians via CRT. Each channel is too small for singularities. Cross-channel correlation < 0.007. Smoothness is preserved.
LAPLACIAN CHANNEL DECOMPOSITION
M5
Hodge Conjecture
Idempotent projection map · 30 unreachable classes
32 idempotents project onto eigenvalue classes. But 30 of 288 classes are unreachable by any projection — the unit-only classes. Finite Hodge fails. Click an idempotent to see its image.
IDEMPOTENT COVERAGE MAP
M6
Birch and Swinnerton-Dyer
Elliptic curve point counts · L-value product
Count points on elliptic curves over each Fp. The L-value factors into 5 independent CRT components. Hasse bound holds. Enter curve parameters to see the computation live.
ELLIPTIC CURVE OVER CRT CHANNELS
y² = x³ + x +

Every Problem Breaks at K = 3

Six of humanity's hardest problems. All have finite mirrors in the ring. All break at the same threshold. 1318 lines of C. Zero dependencies.

CRT ANALYSISSTANDARD STATUS
RiemannPoisson (KS=0.027). CRT = integrable.Unsolved (164 years)
Yang-Millsgap = 4sin²(π/p). Exact formula.Unsolved (70 years)
P vs NP82.5x CRT speedup on Z/2310ZUnsolved (53 years)
Navier-Stokes5 independent channels. Regularity.Unsolved (180 years)
HodgeFAILS. 30 unreachable classes.Unsolved (84 years)
BSDL-value = product of 5 CRT factorsUnsolved (59 years)