angle-cliff
At its defaults it draws error against the angle between the flow and the grid, at pe = 3.13. Three curves against the flow angle on a logarithmic vertical axis. The tuned scheme is exact at zero — 2.4·10⁻¹⁷ — and is 0.00794 five degrees later. Central differencing and upwinding start at 0.517 and 0.136 and fall as the angle grows, because the exact solution's own size in the interior falls with it.
angle-cliff is one function in lib/figures/supg2d.js —
the tuned diffusion in two dimensions — exact along one axis, and the direction it does not go. Everything below came out of it during this build, at
arguments taken from the essays rather than invented for this page. A figure here is the
figure a reader meets in an essay, and if the generator changes, this page changes with it.
At its defaults
Drawn even though every essay passes arguments — which on this site is every essay, at 100% of placements since the standard pass. A default nothing exercises is a trap for the next essay to call this with none, and this is the page where a default that has drifted from the figures around it becomes visible.
Three curves against the flow angle on a logarithmic vertical axis. The tuned scheme is exact at zero — 2.4·10⁻¹⁷ — and is 0.00794 five degrees later. Central differencing and upwinding start at 0.517 and 0.136 and fall as the angle grows, because the exact solution's own size in the interior falls with it.
eps: 0.015
The arguments are the ones A direction the smoother cannot see passes. A value drawn at the generator's defaults instead would be a picture no essay asked for and no assertion has been run against.
Three curves against the flow angle on a logarithmic vertical axis. The tuned scheme is exact at zero — 2.6·10⁻¹⁷ — and is 0.00474 five degrees later. Central differencing and upwinding start at 0.367 and 0.178 and fall as the angle grows, because the exact solution's own size in the interior falls with it.
eps: 0.01
The arguments are the ones Aggregating what the matrix calls strong passes. A value drawn at the generator's defaults instead would be a picture no essay asked for and no assertion has been run against.
Three curves against the flow angle on a logarithmic vertical axis. The tuned scheme is exact at zero — 2.4·10⁻¹⁷ — and is 0.00794 five degrees later. Central differencing and upwinding start at 0.517 and 0.136 and fall as the angle grows, because the exact solution's own size in the interior falls with it.
eps: 0.02
The arguments are the ones Coarsening in one direction only passes. A value drawn at the generator's defaults instead would be a picture no essay asked for and no assertion has been run against.
Three curves against the flow angle on a logarithmic vertical axis. The tuned scheme is exact at zero — 3.5·10⁻¹⁷ — and is 0.00303 five degrees later. Central differencing and upwinding start at 0.263 and 0.198 and fall as the angle grows, because the exact solution's own size in the interior falls with it.
eps: 0.05
The arguments are the ones Smoothing a whole line at once passes. A value drawn at the generator's defaults instead would be a picture no essay asked for and no assertion has been run against.
Three curves against the flow angle on a logarithmic vertical axis. The tuned scheme is exact at zero — 1.7·10⁻¹⁶ — and is 5.86·10⁻⁴ five degrees later. Central differencing and upwinding start at 0.0556 and 0.157 and fall as the angle grows, because the exact solution's own size in the interior falls with it.
eps: 0.03
The arguments are the ones The error smoothing cannot reach passes. A value drawn at the generator's defaults instead would be a picture no essay asked for and no assertion has been run against.
Three curves against the flow angle on a logarithmic vertical axis. The tuned scheme is exact at zero — 3.1·10⁻¹⁶ — and is 0.00156 five degrees later. Central differencing and upwinding start at 0.145 and 0.2 and fall as the angle grows, because the exact solution's own size in the interior falls with it.
What it checked while drawing
Every figure above asserted its own claims on the way to being drawn, and a claim that failed
would have failed the build rather than drawn a wrong picture. Those assertions used to leave
no trace at all: a passing one returned true and the only evidence the figure had
checked anything was that nothing crashed. The list below is what they actually said, collected
by running this generator with an observer installed — not a description of
what it is believed to check.
8 distinct claims across 6 sets of arguments, grouped below by shape — because most of them are one sentence with a different number in it, and how many separate times that sentence was put to the test is the informative part.
a diffusion small enough for a layer and large enough to draw
a grid the dense solve is affordable at
a sweep that starts on the axis and leaves it
and at this Péclet number is still the best of the three at the largest angle
and is the worst of the three at the largest angle drawn
and not five degrees off it
LU is for square matrices
the tuned scheme is exact on the axis
Against the rule
It draws a decomposition and prints its residual. It calls
solve2d,
and every figure above carries the badge — which residualcheck verifies by looking
for it in the emitted SVG rather than by finding the call that builds one. A badge that is
constructed and then left out of the body is the failure that check exists for.
Across the library: the rule bites on 70
of 151 generators —
55 print a residual and
15 are exempt with a published reason;
81 factorise nothing.
Read from lib/residual-rule.js, which is the same body the gate enforces from,
and the gate's last check fails the build if this page and it disagree about any generator.
Where it is called
Changing this generator changes every figure on this list. That is what makes the list worth publishing rather than keeping in a check script.
A direction the smoother cannot see
Give the Laplacian a strong direction and multigrid stops working — from 0.2016 a cycle to 0.9565 — with every component unchanged and the condition number identical to twelve digits. The problem did not get harder. The link between the method's two halves broke.
Iterating, instead of factorisingAggregating what the matrix calls strong
The depth phase measured every method it had on the 45°-rotated anisotropic operator — 0.784, 0.883, 0.844 — and diagnosed the failure as being in the discretisation rather than in the hierarchy. Smoothed aggregation is the standard answer to anisotropy. It returns 0.789.
Iterating, instead of factorisingCoarsening in one direction only
Leave the smoother alone and halve only the strong direction, and the smoothing factor is 0.3340 — identical to line relaxation's, at every anisotropy and every weight, to twelve digits. The convergence factors are then a factor of three apart, and at 45° both repairs fail outright.
Iterating, instead of factorisingExact along one axis
The tuned diffusion makes the answer exact at every node, and in two dimensions it holds at exactly one flow angle. Five degrees off the grid the relative error goes from 1.2·10⁻¹⁴ to 6.9, and by twenty degrees the scheme is worse than the upwinding it was built to improve on.
Iterating, instead of factorisingSmoothing a whole line at once
Solve every grid line in the strong direction exactly rather than sweeping over it, and the smoothing factor goes from 0.9993 back to 0.3340 — which is the one-dimensional answer, on a problem that is not one-dimensional. The repair replaces one ε in the closed form by a one.
Iterating, instead of factorisingThe diffusion that makes the answer exact
Upwinding adds h/2 of artificial diffusion. Central differencing adds none. Add ε·ξ·Pe with ξ = coth(Pe) − 1/Pe and the computed solution is the exact one at every grid point, to 2.4·10⁻¹⁷ — at every Péclet number, on the problem it was derived from and on no other.
Iterating, instead of factorisingThe direction the diffusion does not go
Streamline diffusion adds τbbᵀ, a rank-one tensor that annihilates every direction across the flow. That is the design. The price is 18, 50 and 48 nodes where the computed solution leaves the interval the equation guarantees — and half a coefficient of crosswind diffusion halves the error at every angle while costing exactly nothing where the scheme was exact.
Iterating, instead of factorisingThe error smoothing cannot reach
One weighted Jacobi sweep multiplies every mode of the error by a number, and the number is a sine. Half the modes are cut by three or better, and the other half come back at 0.999 — which is not a failure of the method but the fact the whole of multigrid is built on.