Flaky: PLU decomposition ComplexDouble returns wrong result on Linux

### Multik version

0.3.0

### Module

multik-openblas

### Platform

JVM

### Operating System

Linux

### Description

`NativeLinAlgEx.pluC` for `ComplexDouble` matrices produces incorrect PLU decomposition intermittently on Linux (Ubuntu 22.04.5 LTS, gcc-12 12.3.0, JDK 21 Temurin). The test generates a 4x4 random ComplexDouble matrix (seeded with `Random(42)`), decomposes it into P, L, U, and verifies `P*L*U = A` with epsilon `1e-8`. Failed in 2 out of 4 CI runs.

### Reproducer

```kotlin
val a = data.getComplexDoubleM(4) // deterministic 4x4 ComplexDouble matrix from Random(42)
val (p, l, u) = NativeLinAlgEx.pluC(a)

val lu = NativeLinAlgEx.dotMMComplex(l, u)
val plu = NativeLinAlgEx.dotMMComplex(p, lu)
assertComplexFloatingNDArray(a, plu, epsilon = 1e-8) // Fails intermittently
```

### Additional information

```text
Flaky — failed in 2/4 CI runs.
AssertionFailedError at NativeLinAlgTest.kt:415.
CI runs: 22511673326, 22511741037.
Environment: Ubuntu 22.04.5 LTS, gcc-12 (12.3.0), g++-12, JDK 21 Temurin, x86_64.
```

Start with NativeLinAlgTest.kt at line 415 and reproduce the seeded 4x4 ComplexDouble case through NativeLinAlgEx.pluC, dotMMComplex, and assertComplexFloatingNDArray. Compare successful and failed runs in the listed Linux CI environment, then verify that the decomposition consistently satisfies P*L*U = A within epsilon 1e-8.

Written by the indexing model from the issue text.