Add the 1844th performance column, and what two runs of one commit say about the timings (#1099)

The column for 2.4.0. Seventy-five commits from the 1769th, carrying the rule-set
exchange, Hensel lifting along an evaluation homomorphism, and four wrong answers found by
writing rules out as data.

**The Kernel Benchmark was run twice on this same commit, and the two runs are 30-57%
apart on every one of eighteen benchmarks.** Allocation over the same pair agrees to
0.03% -- `SimplifyEasy` 128,098 B against 128,096 B. That is the thousandfold difference
`PerformanceGate` already argues from, measured here on the workflow this file's columns
come from.

So a timing difference between two of these columns is evidence only above about 50%. The
1769th's write-up reads -7%, -10%, -15% as improvements and concludes "every row is faster
or flat"; none of those is distinguishable from the runner. Only its `SolveEasy` row
survives, and it survives because allocation moved 2.3x alongside the 2.5x -- which is the
rule that write-up itself states.

The table takes run 2 and run 1 is recorded in the prose, so the choice is visible.

**And the exchange cost 13% of `SimplifyEasy`, which this file cannot measure but a
desktop can.** 82,676 ns at the 1769th, 148,223 after the conversions, 94,904 after the
node-type index, 93,732 at the 1843rd -- one machine, both arms, standard deviation under
half a per cent. Every conversion was measured against the commit in front of it and each
came back free or better; nothing was measured against the start.

Counted rather than guessed: 4,914 rule-set invocations on this input at both commits, so
the work shape is identical and this is per-operation overhead -- about 1.3 ns per
invocation, 14 ns per rule attempt. `InequalityEquality` is 26% of it, `InvertNegativePowers`
15% from a single rule. Three ways of making the dispatch cheaper were measured and none
moved it; about 40% is unattributed. The lever is the invocation count, not the dispatch.


Claude-Session: https://claude.ai/code/session_01Bjumi5K7fg8yx6UK1mZTQd

Co-authored-by: Claude Opus 5 <noreply@anthropic.com>

Author: Rafael-SOWNet