Performance: optimize module HS matrix output

[goodchong]

Reminder

• I have read AGENTS.md and docs/developers_guide/agent_governance.md.
• I have linked an issue or explained why this PR does not need one.
• I have added adequate unit tests and/or case tests, or explained why not.
• I have listed the exact verification commands run and their results.
• I have described user-visible behavior changes, including INPUT parameter changes.
• I have explained core-module impact for ESolver, HSolver, ElecState, Hamilt, Operator, Psi, or other source/ changes.
• I have requested any needed governance exception below.

Linked Issue

Fix #

Unit Tests and/or Case Tests for my changes

• Commands run:
• Result summary:
• Checks not run, with reason:

What's changed?

• This PR improves the performance of source_io/module_hs without changing the user-facing matrix formats.

• Replaces row-by-row dense MPI reductions for sparse R-space matrices with sparse entry gathering. Communication
  changes from approximately O(N²) to O(nnz) per R block.

• Builds CSR values, column indices, and row pointers directly in memory, eliminating temporary index files.

• Merges duplicate distributed entries before applying the sparse threshold.

• Iterates only over locally owned rows and columns when calculating r(R), avoiding global nlocal² traversal on every
  MPI rank.

• Batches dense H(k)/S(k)/Vxc matrix reductions in groups of 64 rows and removes the redundant per-row MPI barrier.

• Pre-indexes HContainer atom-pair R blocks, avoiding repeated scans of every atom pair for each lattice vector while
  keeping memory bounded to one materialized R block.

• Adds an optional Google Benchmark for sparse single-R CSR output at different matrix sizes.

• Preserves existing text and native binary output ordering and formats.

Governance Notes

• INPUT/docs changes:
• Core module impact:
• Exceptions requested:

[mohanchen]

could you avoid using PARAM?

[mohanchen]

Could you explain more about what this PR does?