Newform orbit 810.2.a.b
-expansion
Coefficient field
The coefficients generate . The eigenvalue of is a root of
AtkinβLehner signs
| 2 | |
| 3 | β () |
| 5 |
For every form in the orbit has .
Elliptic curve
By modularity this newform is the one attached to the isogeny class 810.b (Cremona's 810d). Its optimal curve 810d1 is
of rank 1 with torsion of order 3. Its , computed by Sagebrush's point-counting engine, equal the newform's for every prime .
Traces of Hecke eigenvalues
| 2 | 3 | 5 | 7 | 11 | 13 | 17 | 19 | 23 | 29 | 31 | 37 | 41 | 43 | 47 | 53 | 59 | 61 | 67 | 71 | 73 | 79 | 83 | 89 | 97 | |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| β1 | 0 | 1 | β1 | β6 | 2 | 0 | β4 | β9 | β3 | β4 | 8 | 3 | 8 | 3 | β6 | β6 | β13 | β13 | 6 | β4 | β10 | 9 | β9 | 2 |
for every is in the JSON.
SatoβTate
The normalized for the primes not dividing 810, against the semicircle :
Provenance
Stored in the atlas: computed by Sagebrush's modular symbols engine in 2.654 s (one thread) and proven: the dimensions are certified, and the characteristic polynomial of over is reconstructed by CRT past a rigorous coefficient bound. The label, dimension and for every equal LMFDB's.
The engine's checks
dim M^+ at levels [1, 2, 3, 5, 6, 9, 10, 15, 18, 27, 30, 45, 54, 81, 90, 135, 162, 270, 405, 810]: [0, 1, 1, 1, 3, 2, 3, 4, 5, 4, 10, 8, 11, 10, 22, 20, 29, 58, 56, 166] (certified at ell = 2147482801)T = [(7, 1), (11, 2), (13, 3), (17, 4), (19, 5), (23, 6), (29, 7), (31, 8), (37, 9), (41, 10), (43, 11), (47, 12), (53, 13), (59, 14), (61, 15), (67, 16)]; h of degree 16 = 1 x 16 (phi(ord eps) x dim S_k^new); 6 primes, 0 rejected
Recompute it yourself
In Sage mode at sagebrush.space (or npx sagebrush --sage):
newform_orbits(810, 2)[1] # 810.2.a.b
With the Rust engine (source), in engine/:
cargo run --release -p sagebrush-web --example atlas -- 810:2 1000