Reading the bench
Reading the bench…Reading the bench
Reading the bench…analysis
Every deadline in Shorwatch comes from one two-parameter model anchored on a published factoring estimate. Move the assumed machine and the whole queue re-orders, because the ordering is a consequence of the model rather than a stored opinion.
Gidney & Eakerå, Quantum 5, 433 (2019) · code distance d = 13
no key is within the assumed machine
x: RSA modulus bits · y: physical qubits (log scale, labelled) · lower labels: projected CRQ year · dashed line: assumed machine
Same ordering rule as the API, recomputed as you drag.
No measured keys yet. Add a watch on the workspace and measure it first.
Every constant here is also rendered on the standards route.
RSA comparable security
1024→80, 2048→112, 3072→128, 4096→152 bits
NIST SP 800-57 Part 1 Rev. 5, Table 2Approved curve security
P-224/P-256→128, P-384→192, P-521→256 bits
NIST FIPS 186-5 and SP 800-57 Part 1 Rev. 5RSA-2048 factoring anchor
4098 logical / ~20,000,000 physical qubits in 8 h
Gidney & Eakerå, Quantum 5, 433 (2019)CRQ projection basis
2035 for RSA-2048, 7.7 years per modulus doubling
Interpolation anchored on the published RSA-2048 estimate aboveAnchor reference
4,098 logical qubits and roughly 20,000,000 physical qubits at an error rate of 1e-3 for RSA-2048, in 8 hours. Logical qubits scale with n·log₂n; physical qubits scale with the square of the surface-code distance.