Size species arrays from the generated mechanism on AMD

[sbryngelson]

Summary

Follow-up to #1915 and #1852. Generic (non-case-optimized) amdflang builds declared species arrays with AMD_NUM_SPECIES_MAX = 60 and sys_size arrays with AMD_SYS_SIZE_MAX = 70: 26 uses in 11 files, each an #:if not MFC_CASE_OPTIMIZATION and USING_AMD branch. Now that MFC generates m_thermochem, the species count is known at build time.

• The toolchain writes build/include/<target>/thermochem.fpp next to case.fpp: #:set NUM_SPECIES = <m_thermochem's num_species> and #:set CHEMISTRY = <bool>.
• Species arrays are declared dimension(${NUM_SPECIES}$) on every backend. Every species USING_AMD branch collapses to one declaration, and the per-thread arrays carry the mechanism's exact count (10 for h2o2, 11 for the Add heterogeneous reacting surface boundary conditions #1821 carbon mechanism) instead of 60. Raise the amdflang species bound so mechanisms above ten species build #1852-style "raise the bound" changes are no longer needed.
• AMD_SYS_SIZE_MAX = 10 + NUM_SPECIES for chemistry builds (chemistry pins num_fluids to 1, leaving at most 10 flow variables), so HLLC star states and CBC L vectors shrink from 70 to 20 for h2o2. Non-chemistry builds keep 70, so no non-chemistry case loses headroom. The sys_size check stays; the species-count check goes with the bound.
• Fypp resolves #:include in the including file's directory before -I paths, so the default thermochem.fpp (for Fypp runs outside the toolchain) lives in src/common/include/defaults/, which cmake/Fypp.cmake searches last.

A literal extent rather than dimension(num_species): my first attempt used the num_species parameter directly and failed C4EB58A8 on amdflang GPU (large tolerance mismatch). That run overlapped another build in the same checkout, so I do not claim it proves an amdflang bug, but the literal is what the former fallback used, is identical on every compiler, and passes cleanly.

Performance

MI250X (one GCD), amdflang 23.2.1, generic --gpu mp --no-mpi build, examples/2D_premixed_flame_vortex (1024x512, h2o2, HLLC, reactions + diffusion), 500 steps, interleaved runs:

	master fef3c276	this PR
s/step (3 runs)	0.05117, 0.05184, 0.05162	0.05004, 0.04960, 0.04977
mean	0.05154	0.04980 (-3.4%)

Every run of this PR is faster than every master run. All 17 output fields are bit-identical to master at step 500.

Testing

• amdflang 23.2.1, OpenMP offload, MI250X (OMP_TARGET_OFFLOAD=MANDATORY): ./mfc.sh test --gpu mp --only Chemistry - 16 passed, 0 failed. This is the configuration the guards applied to.
• gfortran 12.2 CPU: --only Chemistry - 16 passed, 0 failed.
• Checked that the generated include takes precedence (NUM_SPECIES=10, AMD_SYS_SIZE_MAX=20 for an h2o2 chemistry build) and that the default is used only when no generated file exists.
• ./mfc.sh precheck passes.

Interaction with #1821

#1821 adds new ${AMD_NUM_SPECIES_MAX}$ declarations in m_ibm.fpp. Whichever of the two merges second needs them switched to ${NUM_SPECIES}$; that is mechanical and I will do it.

Prepared with Claude Code.

Contribution Policy

We do not accept pull requests generated primarily by AI without genuine understanding or real-world usage context.

All contributions are expected to demonstrate:

• A clear understanding of the codebase
• Alignment with product direction
• Thoughtful reasoning behind changes
• Evidence of real-world usage or hands-on experience with the problem

If these expectations are not met, we would prefer to implement the changes ourselves rather than spend time reviewing low-effort submissions.

---

Acknowledgement

• I confirm this PR meets the above expectations and reflects my own understanding and real-world context.

---

PR template credit: junegunn

[Copilot]

Copilot was unable to review this pull request because the user who requested the review has reached their quota limit.

[github-actions]

Lines of Code

File	Lines	Diff
src/common/m_chemistry.fpp	404	-16
src/simulation/m_riemann_state.fpp	1185	-8
src/simulation/m_compute_cbc.fpp	274	-7
src/simulation/m_riemann_solver_hllc.fpp	1315	-5
src/common/include/defaults/thermochem.fpp	2	+2
src/common/m_checker_common.fpp	37	-2
src/common/m_variables_conversion.fpp	960	-2
src/simulation/m_riemann_solver_hll.fpp	614	-2
src/simulation/m_riemann_solver_lf.fpp	515	-2
src/common/include/shared_parallel_macros.fpp	161	-1
src/simulation/m_cbc.fpp	1117	-1
src/simulation/m_ibm.fpp	1334	-1

Directory	Lines	Diff
common	10402	-19
simulation	27822	-26
total	46635	-45

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 61.48%. Comparing base (fef3c27) to head (7958e6f).

Additional details and impacted files

@@           Coverage Diff           @@
##           master    #1918   +/-   ##
=======================================
  Coverage   61.48%   61.48%           
=======================================
  Files          86       86           
  Lines       22485    22485           
  Branches     3298     3298           
=======================================
  Hits        13826    13826           
  Misses       6206     6206           
  Partials     2453     2453           

☔ View full report in Codecov by Harness.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.