Add MREEF multirate extrapolated explicit Euler solver with tests

[singhharsh1708]

Checklist:
Appropriate tests were added
Any code changes were done in a way that does not break public API
All documentation related to code changes were updated
The new code follows the
contributor guidelines
, in particular the SciML Style Guide
andCOLPRAC
Any new documentation only uses public API
Additional context:
This PR adds MREEF (Multirate Extrapolated Explicit Euler), based on Constantinescu & Sandu (2010).
The method targets SplitODEProblem systems and combines multirate Euler base steps with Richardson extrapolation to achieve higher-order accuracy.

Basic convergence tests and compatibility checks (SplitODEProblem, ForwardDiff, StaticArrays, complex states) were added under test/multirate/mreef_tests.jl. All existing OrdinaryDiffEqLowOrderRK tests pass locally with no regressions.

This implementation was mainly developed to better understand the solver architecture in OrdinaryDiffEq and to explore multirate methods.

[singhharsh1708]

It looks like GPU CI is failing with a runtime error. I'm investigating the issue and will check compatibility with JLArrays / GPU backends.
Locally, all CPU tests pass including ForwardDiff, StaticArrays, and complex states. I'll update shortly once I identify the exact failure.

[singhharsh1708]

All CI checks are now passing.
the FBDF GPU test was previously marked as @test_broken but is now passing, so it has been updated to @test. i reverted this change and project.toml change both

[singhharsh1708]

rebased onto latest upstream/master and cleaned up the changes CI is passing now

[ChrisRackauckas]

Rebase onto the new testing infrastructure.

[singhharsh1708]

@ChrisRackauckas Rebased onto latest upstream/master and applied runic formatting fixes. All checks should now pass.

[singhharsh1708]

@ChrisRackauckas can you run test on this?

[singhharsh1708]

@ChrisRackauckas is this correct now?

[singhharsh1708]

@ChrisRackauckas
Addressed all review comments:
Moved MREEF to new OrdinaryDiffEqMultirate subpackage
Added docstring via generic_solver_docstring
Fixed per-step allocation: T pre-allocated in cache
Fixed alias: using @.. in-place assignment
Removed integrator.k pointer reassignments
Convergence tests now use DiffEqDevTools.test_convergence with analytic solution
Stats test: nf ≈ m * nf2 atol=m (offset from initialize! contributing 1 to each counter)
FSAL + Hermite interpolation flagged for follow-up PR
19/19 tests passing, Runic formatted. CI running.