.NET — System.Numerics.Decimal128

IEEE 754-2019 compliant decimal128 high-performance software solution created by Miguel.

A pre-release review series tracking System.Numerics.Decimal128, the 128-bit decimal floating-point type Microsoft plans to introduce in .NET 11, targeted Nov 2026. One installment is published for each preview and release candidate through GA. Each edition run through the same standards-anchored harness so the type can be watched as it converges.

The reviews

.NET 11

Date Release Review Headline
2026-08-12 RC 1, build 26411.119 Read → A short edition: ten dailies on, no finding changed status, the 54,820-vector sweep reproduces exactly, and performance is stationary on two microarchitectures. The freeze is holding.
2026-08-02 RC 1, build 26401.101 Read → Quantum-preserving ToString lands and clause 5.12 closes; the sweep widens to 54,820 vectors with 0 failures; performance stops moving; rounding attributes and status flags remain absent, so the type ships numerically correct and not IEEE 754 conformant.
2026-07-28 preview 7, build 26376.106 Read → FMA arrives (truly fused); full IEEE 754 recommended-function surface; division up to 4× faster; new ln/log10 near-1 accuracy findings; RootN correct where the reference ports were not.
2026-07-14 preview 7, build 26366.102 Read → Numerically correct; no status flags; rounding is composed rather than fused (double-rounding); ToString needs work; missing TotalOrder; missing FMA; naive division.

Release candidate and GA editions will be added under this heading as they are published.