Run the published benchmarks yourself. Check the hash.
Each card runs a published problem in your browser with the same engine our registry records, compares our value with the cited reference within its stated tolerance, and stamps the result with a SHA-256 hash over the inputs and the verdict. Run it again and the hash is identical, so anyone can confirm that a run reproduces. The comparison shown on every card is our value against the published reference.
Verification suite: benchmark, our value, published value, difference, tolerance and status; the run button is on each row, and the source, inputs and hash open under it.
Problem
Ours
Published
Difference
Tolerance
Status
Details
Bearing capacity — closed formTerzaghi, Meyerhof, Hansen, Vesić and EN 1997-1 Annex D bearing capacity against worked examples from Das, Bowles, the EC7 Designers’ Guide, the JRC worked examples, IS 6403 and FHWA/GEO5 manuals.
the report evaluated Nq = e^(π tanφ)·tan²(45° + φ/2) with π = 3.14: that reproduces all four printed cells exactly (48.872, 74.804, 23.172, 27.716) and is the whole +0.09…+0.13 % offset (found in our internal review). 0.3 % still rejects every alternative Nγ form (nearest is 4.3 % away)
the report evaluated Nq = e^(π tanφ)·tan²(45° + φ/2) with π = 3.14: that reproduces all four printed cells exactly (48.872, 74.804, 23.172, 27.716) and is the whole +0.09…+0.13 % offset (found in our internal review). 0.3 % still rejects every alternative Nγ form (nearest is 4.3 % away)
Quantity
Nγ
EN 1997-1 Annex D Nq at φ′d = tan⁻¹(tan 38°/1.25) (M2)
the report evaluated Nq = e^(π tanφ)·tan²(45° + φ/2) with π = 3.14: that reproduces all four printed cells exactly (48.872, 74.804, 23.172, 27.716) and is the whole +0.09…+0.13 % offset (found in our internal review). 0.3 % still rejects every alternative Nγ form (nearest is 4.3 % away)
the report evaluated Nq = e^(π tanφ)·tan²(45° + φ/2) with π = 3.14: that reproduces all four printed cells exactly (48.872, 74.804, 23.172, 27.716) and is the whole +0.09…+0.13 % offset (found in our internal review). 0.3 % still rejects every alternative Nγ form (nearest is 4.3 % away)
Quantity
Nγ
Terzaghi, square footing — cohesion + surcharge terms (the Nγ term is a convention comparison, not reproduced)
BC-DAS-4.1 · recorded
968.329kPa
968.2kPa
+0.013 %
± 0.5 %
Ours968.329 kPa · Published968.2 kPa · +0.013 %
Not run
Source
Das, B.M., Principles of Foundation Engineering (Cengage), Ch. 4Example 4.1, p. 166
Inputs
c′ 20, φ′ 25°, γ 16.5, Df 1.5, B = L = 2 m, square; Das Table 4.1 Nc 25.13, Nq 12.72
Tolerance
Das prints q_u from factors rounded to 2–3 significant figures
Quantity
1.3c′Nc + qNq (printed 653.38 + 314.82)
General bearing equation, same footing
BC-DAS-4.3 · recorded
1,374kPa
1,373.2kPa
+0.058 %
± 0.5 %
Ours1,374 kPa · Published1,373.2 kPa · +0.058 %
Not run
Source
Das, B.M., Principles of Foundation Engineering (Cengage), Ch. 4Example 4.3
Inputs
as Ex 4.1, Vesić/Meyerhof general equation
Tolerance
Das prints q_u from factors rounded to 2–3 significant figures
Quantity
q_u
Inclined load β = 20°
BC-DAS-4.5 · recorded
273.801kPa
273.66kPa
+0.052 %
± 0.5 %
Ours273.801 kPa · Published273.66 kPa · +0.052 %
Not run
Source
Das, B.M., Principles of Foundation Engineering (Cengage), Ch. 4Example 4.5
Inputs
square 1.25 m, c′ 0, φ′ 30°, γ 18, Df 0.7, β 20°; Das Eq. 4.26 / Table 4.3 (Meyerhof 1963) inclination factors — vesic_inclination 'das_general' (ENG-3: the vesic default is now Vesić's own set)
Tolerance
Das prints q_u from factors rounded to 2–3 significant figures
Das, B.M., Principles of Foundation Engineering (Cengage), Ch. 4Example 4.7, p. PDF 217
Inputs
strip B 2 m, c′ 0, φ′ 40°, γ 16.5, Df 1.5, e 0.2 m
Tolerance
Das prints q_u from factors rounded to 2–3 significant figures
Quantity
q′_u
Eccentric + inclined strip
BC-DAS-4.12 · recorded
461.671kPa
461.98kPa
−0.067 %
± 0.5 %
Ours461.671 kPa · Published461.98 kPa · −0.067 %
Not run
Source
Das, B.M., Principles of Foundation Engineering (Cengage), Ch. 4Example 4.12
Inputs
strip B 1.5 m, φ′ 35°, γ 16, Df 1, e 0.15 m, β 20°; Das Eq. 4.26 / Table 4.3 (Meyerhof 1963) inclination factors — vesic_inclination 'das_general' (ENG-3: the vesic default is now Vesić's own set)
Tolerance
Das prints q_u from factors rounded to 2–3 significant figures
Quantity
q′_u
Case study — Bozozuk silo, φ = 0
BC-DAS-SILO · recorded
182.503kPa
181.8kPa
+0.387 %
± 0.5 %
Ours182.503 kPa · Published181.8 kPa · +0.387 %
Not run
Source
Das, B.M., Principles of Foundation Engineering (Cengage), Ch. 4Ch. 4 case study (Bozozuk)
Inputs
c_u 23.3, φ 0, γ 18, 7.2 m square, Df 1.52
Tolerance
Das prints q_u from factors rounded to 2–3 significant figures
φ 36°, γ 17.28, GWT 3.048 m, Df 1.219, B 2.134, L 30.48
Tolerance
NHI (2006) reads the depth factor off Table 8-6 as dq ≈ 1.15 (A.7-4). Since W1-D1 (13 Nov 2026) the engine applies AASHTO LRFD 10th (2024) Eq. 10.6.3.1.2a-10, dq = 1 + 2tanφ(1 − sinφ)²·arctan(Df/B) = 1.128 at Df/B = 0.571 (the linear Df/B form gave 1.141, −0.37 %), so the gap to NHI's table read is now −0.98 %: a code revision, not an engine error. Factors Nq 37.8 / Nγ 56.3 are also table-read
Quantity
q_ult
IS 6403 Table 1 Nc at φ = 45° — checked against the corrected value, NOT the printed cell
BC-IS-T1-NC45 · recorded
133.874
133.88
−0.005 %
± 0.1 %
Ours133.874 · Published133.88 · −0.005 %
Not run
Source
IS 6403:1981 (First Revision), Cl 5.1.1 Table 1Table 1, φ = 45° (printed Nc 138.88), p. 8 (PDF 15)
Inputs
strip, general shear, φ 45°
Tolerance
Table 1 printed to 2 decimals; the cell was cross-checked against Nc = (Nq − 1)cot φ
Quantity
Nc = (Nq − 1)cot φ with the printed Nq 134.88 (Martin exact 133.874)
IS 6403 Table 1 Nγ at φ = 30°
BC-IS-T1-NG30 · recorded
22.4025
22.4
+0.011 %
± 0.1 %
Ours22.4025 · Published22.4 · +0.011 %
Not run
Source
IS 6403:1981 (First Revision), Cl 5.1.1 Table 1Table 1, φ = 30°, p. 8 (PDF 15)
Inputs
strip, general shear, φ 30°
Tolerance
Table 1 printed to 2 decimals; the cell was cross-checked against Nc = (Nq − 1)cot φ
Fine, GEO5 Verification Manual No. 2 — Spread Footing (update 03/2016)case 1 (γG = 1.00), R_d after γR;v 1.40; factor set p. 3–4 ("A standard analysis proposed by J. Brinch Hansen is used"), p. 3–5
Inputs
1.80 × 2.20 m, Df 1.20, SM φ′ 31.5°, γ 17.5, N 910, Hy 120, Mx 200, terrain 7°; Hansen with hansen_factor_set 'geo5_reduced' (the manual's own s, d, i set)
Tolerance
Tolerance ± 1 % (our band for closed-form bearing checks). GEO5 prints a hand calculation (381.706 kPa, every factor rounded to 3 decimals) and its program's value (381.92 kPa), 0.06 % apart; the engine gives 381.919 kPa
as BC-GEO5-VM02 with self-weight γG 1.35: e 0.238 m, δ 6.574°; hansen_factor_set 'geo5_reduced'
Tolerance
Tolerance ± 1 % (our band for closed-form bearing checks). GEO5 prints a hand calculation (386.501 kPa) and its program's value (386.61 kPa), 0.03 % apart; the engine gives 386.612 kPa
Quantity
R_d / γR;v (hand calc; program 386.61)
Bearing capacity — finite elements / limit analysisPrandtl, Davis & Booker, Cox and Reissner/Martin collapse loads reproduced by displacement FE and FE limit analysis, against the PLAXIS and Optum G3 verification manuals. Heavy solves — run by the registry script.
Prandtl strip, Tresca, smooth
FE-PRANDTL-SMOOTH · recorded
5.13612
5.14159
−0.106 %
± 0.5 %
Ours5.13612 · Published5.14159 · −0.106 %
Recorded pass · not run here
Source
Potts, D.M. & Zdravković, L., Finite Element Analysis in Geotechnical Engineering Vol. 2 (Thomas Telford, 2001), §6.5.4.2; Martin table in Optum G3 strip-on-sand verificationPrandtl N_c = 2 + π, p. 220–221
Inputs
B 2 m, φ 0, c_u 1, weightless, E 1000, ν 0.49, domain 5 × 3 m, meshes n = 2 / 3 / 4.5
Tolerance
our acceptance limit for this benchmark: Prandtl 5.14·c_u within 0.5 %; exact solution
Quantity
N_c = q_u / c_u
Prandtl strip, Tresca, rough
FE-PRANDTL-ROUGH · recorded
5.1601
5.14159
+0.360 %
± 0.5 %
Ours5.1601 · Published5.14159 · +0.360 %
Recorded pass · not run here
Source
Potts, D.M. & Zdravković, L., Finite Element Analysis in Geotechnical Engineering Vol. 2 (Thomas Telford, 2001), §6.5.4.2; Martin table in Optum G3 strip-on-sand verificationPrandtl N_c = 2 + π, p. 220–221
Inputs
as smooth; footing horizontal DOFs fixed
Tolerance
exact for smooth and rough footings (Potts Vol. 2 p. 221); ≤ 0.5 %
Quantity
N_c = q_u / c_u
Davis & Booker (1973) strip, smooth, c = 1 + 2z kPa
FE-DAVIS-BOOKER-SMOOTH · recorded
7.81694kPa
7.79982kPa
+0.219 %
± 1 %
Ours7.81694 kPa · Published7.79982 kPa · +0.219 %
Recorded pass · not run here
Source
PLAXIS Validation KB0110019 (Bentley, 2018), after Davis & Booker (1973) Géotechniquestrip footing, smooth, p. 1–4
Inputs
B 2 m, φ 0, c = 1 + 2z kPa, weightless, E = 299 + 598z kPa, ν 0.495; p = β[(2+π)c_ref + B·c_inc/4], β 1.27. Reference = the analytic β·[(2+π)c_ref + B·c_inc/4] = 7.7998 kPa from the PDF's eq. 1 and its chart-read β; the PDF itself prints the collapse pressure as "7.8 kN/m²" (p. 4)
Tolerance
our acceptance limit: ≤ 1 % on three meshes, each converged; β is chart-read (Potts Vol. 2 p. 237 caveat)
Quantity
mean collapse pressure
Davis & Booker (1973) strip, rough
FE-DAVIS-BOOKER-ROUGH · recorded
9.15482kPa
9.08956kPa
+0.718 %
± 1 %
Ours9.15482 kPa · Published9.08956 kPa · +0.718 %
Recorded pass · not run here
Source
PLAXIS Validation KB0110019 (Bentley, 2018)strip footing, rough, p. 1–4
Inputs
as smooth, β 1.48. Reference = analytic 1.48·[(2+π) + 1] = 9.0896 kPa (eq. 1); the PDF prints "9.1 kN/m²" (p. 4)
Tolerance
gate: ≤ 1 %, three meshes, converged:true
Quantity
mean collapse pressure
Cox (1962) smooth rigid circle, axisymmetric
FE-COX-CIRCULAR · recorded
226.204kPa
225.6kPa
+0.268 %
± 1 %
Ours226.204 kPa · Published225.6 kPa · +0.268 %
Recorded pass · not run here
Source
PLAXIS Validation KB0109995 (Bentley, 2018), after Cox (1962)circular footing on Mohr–Coulomb, p. 1–3
Inputs
R 1 m, E′ 2400, ν′ 0.2, c′ 1.6, φ′ = ψ 30°, γ 16, K0 0.5, domain 4 × 4 m, meshes n = 0.667 / 1 / 1.5; γR/c = 10 → p = 141·c
B 10 m, φ = ψ 30°, c 0, γ 16, q0 0; recipe mesh (fan, n = 1, 1,040 elements); default N_γ domain
Tolerance
BRACKET ROW: pass when lower bound ≤ exact ≤ upper bound — the statement the bound theorems make, and the rule the finite-element registry already applies to the same engine (k5c-strip-ngamma rows, "LB ≤ exact ≤ UB"). ± 0.01 % is only the containment precision (Martin's N_γ is printed to 6 significant figures); Δ is not reported because the exact value need not sit at the mean of the bounds. The bracket width is recorded with every measurement and stated in the notes
Quantity
q_u = ½γB·N_γ (N_γ 7.65300)
N_γ strip on cohesionless sand, rough — FE limit analysis: do the lower and upper bounds contain the exact value?
BRACKET ROW: pass when lower bound ≤ exact ≤ upper bound — the statement the bound theorems make, and the rule the finite-element registry already applies to the same engine (k5c-strip-ngamma rows, "LB ≤ exact ≤ UB"). ± 0.01 % is only the containment precision (Martin's N_γ is printed to 6 significant figures); Δ is not reported because the exact value need not sit at the mean of the bounds. The bracket width is recorded with every measurement and stated in the notes
Quantity
q_u = ½γB·N_γ (N_γ 14.7543)
Pavement — layered elasticCritical pavement strains against the IRC:37-2018 IITPAVE reference output (Annex I).
IRC:37 Annex I Fig I.3 — εt at the bottom of the bituminous layer, dual-wheel standard axle
D1-IRC37-DUAL-WHEEL · recorded
128.481µε
128.3µε
+0.141 %
± 2 %
Ours128.481 µε · Published128.3 µε · +0.141 %
Not run
Source
IRC:37-2018 Guidelines for the Design of Flexible Pavements, Annex I, Fig. I.3 (IITPAVE output)Fig. I.3, p. 78 (PDF 88)
Inputs
3000 / 450 / 75 MPa, ν 0.35, h 140 / 500 mm; 2 × 20 kN at 0.56 MPa, 310 mm c/c; points R = 0 and 155 mm
Tolerance
53 D1 exit: strains within 2 %
Quantity
εt (IITPAVE)
IRC:37 Annex I Fig I.3 — εv on top of the subgrade, dual-wheel standard axle
D1-IRC37-DUAL-WHEEL-EV · recorded
205.59µε
205.3µε
+0.141 %
± 2 %
Ours205.59 µε · Published205.3 µε · +0.141 %
Not run
Source
IRC:37-2018 Guidelines for the Design of Flexible Pavements, Annex I, Fig. I.3 (IITPAVE output)Fig. I.3, p. 78 (PDF 88)
Inputs
as εt row
Tolerance
53 D1 exit: strains within 2 %
Quantity
εv (IITPAVE)
Rock slope — limit equilibriumPlanar rock-slide factor of safety against the Rocscience RocPlane verification manual (Sau Mau Ping).
Rock planar slide, no tension crack, saturated — Sau Mau Ping Road slope
D2-ROCKPLANAR-SAUMAUPING · recorded
0.818425
0.818425
0.000 %
± 0.1 %
Ours0.818425 · Published0.818425 · 0.000 %
Not run
Source
Rocscience RocPlane Verification Manual, Problem #1 (Hoek, Sau Mau Ping Road) — the 0.8184254 hand calculation is printed here; the equations (Fig. 6, eq. 4) are from Hoek, Practical Rock Engineering (Practical-Rock-Engineering-Full-Text.pdf), which does not print this valueVM #1, without tension crack, p. 8
Inputs
H 60 m, face 50°, plane 35°, c 100 kPa, φ 35°, γr 27, γw 10, α 0.08, no crack, Hw = H = 60 m
Tolerance
manual prints 7 s.f.
Quantity
factor of safety
Structural element design checksSteel strut buckling, geogrid long-term strength and ground-anchor bond against Eurocode 3 and FHWA GEC manuals.
Gardner, L. & Nethercot, D.A., Designers' Guide to Eurocode 3: EN 1993-1-1 (Thomas Telford), Ex. 6.7 (CHS 244.5×10 S355, pinned L 4 m) — the book is not in the library; the previous file pointer named DD ENV 1993-1-1:1992, which does not contain this example (corrected in our defect log)Example 6.7
Inputs
CHS 244.5×10 S355, A 7370 mm², I 50.73e6 mm⁴ (i = √(I/A)), pinned L_cr 4 m, curve a (α 0.21); γ_M1 = 1.00 (EN 1993-1-1 §6.1 NOTE 2B), E = 210 000 MPa (§3.2.6)
Tolerance
Designers' Guide prints N_b,Rd to 4 significant figures (2297 kN)
GEC-11 Table E1-7.3 prints T_al / T_r to 4 significant figures
Quantity
T_r = 0.90·T_al (T_al = 1085 printed)
Ground anchor grout–ground bond — GEC-4 Appendix A Example 1, allowable bond at L = 12 m
DC-ANCHOR-GEC4-EX1-BOND · recorded
600kN
600kN
0.000 %
± 0.5 %
Ours600 kN · Published600 kN · 0.000 %
Not run
Source
Sabatini, Pass & Bachus (1999) FHWA-IF-99-015 GEC-4 Ground Anchors and Anchored SystemsAppendix A Example 1 (soil, medium-dense silty sand), p. A-8/A-9
Inputs
soil anchor, load-transfer rate 100 kN/m, FS 2.0 (GEC-4 §5.3.6), fixed length 12 m; PTI DC35.1; tendon 32-mm Grade 150 bar (A 806 mm², f_pk 1035 MPa)
Tolerance
GEC-4 Example 1 prints the bond strength as a rounded hand value (600 kN)
Quantity
allowable bond = (100/2.0)·12
The content hash covers the inputs and the verdict at 10 significant figures, not the runtime or the timestamp, so it is identical on every machine running the same code. Heavy finite-element solves are shown read-only with their last recorded value; they are reproduced by app/scripts/verification-registry.mjs.
Run the same engines on your own design.
The analyses behind these benchmarks are the ones you run in a project. Start free and put your own numbers through them.